QuanterLab · Research

Dalio's Holy Grail, Measured and Then Walked at Five Breadths: Three, Five, Seven, Ten and Fifteen Funds

Dalio's chart says fifteen uncorrelated streams cut risk by four fifths. Run his own construction on the fifteen most liquid ETFs and the count comes back 3.3, with 2.1 in 2022 and nothing above 3 since.

By Serhat Girgin

Universe · Holy Grail menus (fixed lists) (membership resolution not recorded)

Method · Comparative: Trio (SPY, TLT, GLD), equal risk contribution vs The full fifteen-fund shelf, equal risk contribution

Manipulated variable ·

The manipulated variable is MENU BREADTH on one weighting rule: arm A holds the trio (SPY, TLT, GLD) and arm B holds the full fifteen-fund shelf (five equity funds, five bond and credit funds, gold, silver, broad commodities, listed property, the euro), both weighted by equal risk contribution re-estimated point-in-time from the trailing 252 trading days. Everything else is registered identical: quarterly rebalance with re-estimation, one-year windows, total return with dividends from the payment record, ten basis points per one-way traded dollar, the 60/40 SPY/AGG as the benchmark line. Every menu is RULE-DERIVED and frozen before any walk ran - the trio, the published All Seasons sleeves, one fund per asset class, the seven plus the three broadest missing classes, and the full shelf - and this walk is one of a DECLARED FAMILY OF SIX rules across five registered walks (the five menus under equal risk contribution, plus the fifteen under inverse volatility as the correlation-blind twin), with a 60/40 carried as the author benchmark, not a searched rule. The fifteen funds are among the most liquid in existence: any pricing refusal is a data defect and halts the walk. No result of any arm was known when this text was registered.

Step size · 1 year per forward window

Out-of-sample span · 2008-01-02 → 2025-12-31

Compiled · August 30, 2026

Search family · the Holy Grail family: six book-and-weighting rules on one fixed ETF shelf across five registered walks - the three-fund trio and the full fifteen-fund shelf under equal risk contribution as the two-arm comparative, the five-fund All Seasons sleeves, the curated seven (one fund per asset class) and the ten-fund book under equal risk contribution as single-arm walks, and the fifteen-fund shelf under inverse volatility as the correlation-blind twin - with a 60/40 SPY/AGG carried as the study benchmark, outside the family (N = 6, every member reported)

Abstract

The most liquid fund shelf a person can buy holds 3.3 independent bets. Dalio's chart needs fifteen.

The Holy Grail of investing, in his words: find fifteen good, uncorrelated return streams and risk falls by roughly eighty percent while return stands still. At correlation zero, fifteen equal streams carry a quarter of one stream's volatility. Whether the shelf a person can actually buy supplies those streams is what this study measures, in his own units. Standardize every fund to ten percent volatility, weight them equally, and the composite's volatility inverts to the average pairwise correlation, which converts to a count of effective bets. Fifteen funds, 2008 through 2025: correlation 0.26, 3.3 bets. The ceiling across every breadth we cut is about 3.5; five funds reach it and nothing wider gets past it. In 2022 the count fell to 2.1 and it has not been above 3 since.

Then we walked it. Five rule-derived menus of the same shelf, three funds to fifteen, run under identical equal-risk-contribution machinery: quarterly, point-in-time, eighteen one-year windows from January 2008, total return net of costs, against a 60/40 SPY/AGG. The three-fund book returned 8.00 percent a year and the fifteen-fund book 3.78, half the rate for 28 percent less volatility. The trio led in sixteen of eighteen windows and 100.0 percent of paired bootstrap paths. Return floors at one fund per class: seven, ten and fifteen funds land within 0.13 points a year of each other. Breadth won two windows in eighteen years, the 2009 rebound and 2022, where the shelf lost 4.3 points less than the trio.

Every walk is a registered study on this platform; the circuits, per-window books and frozen records are linked below, and every line on Figure 1 can be switched on and off.

Dalio’s own exhibit, measured on the shelf. Top: every fund standardized to ten percent volatility, equally weighted; the curves are Dalio's chart at correlation 0, 0.10 and 0.25, and the dots are the five menus as walked - all five sit on the 0.25 line, and none follows the promised line down. Bottom: the effective number of bets the fifteen tickets held each calendar year - 3.3 on the full record, 2.1 in 2022, under 3 ever since.

1 Methodology

The menus and their rules. Five fixed lists, frozen in the platform's code and selectable by name only, each breadth derived from a stated rule rather than picked by hand. TRIO: the three-asset book, SPY, TLT, GLD, equities, duration, metal. ALL SEASONS FIVE: the sleeves of Dalio's published retail recipe (SPY, TLT, IEF, GLD, DBC). CURATED SEVEN: exactly one fund per asset class the liquid shelf offers, equities SPY, duration TLT, inflation-linked TIP, metal GLD, commodity DBC, property VNQ, currency FXE. TEN: the seven plus the three broadest classes still missing, small-cap IWM, developed international EFA, investment-grade credit LQD. FIFTEEN: the full shelf, redundancies included, the ten plus Japan EWJ, emerging VWO, intermediate duration IEF, aggregate bonds AGG, silver SLV. The youngest fund, SLV, first trades in April 2006, and with a 252-day estimation window it dates the walk: eighteen one-year windows anchored each January 2008 through 2025.

The engine and the clock are shared. Every book runs identical machinery: equal risk contribution weights, re-estimated quarterly from the trailing year of data available that day, one-year windows, total return with dividends credited from the payment record, ten basis points per one-way traded dollar. The only sealed difference between the host's two arms is the menu name, and the only difference down the ladder is the menu name. Equal risk contribution solves for the weights at which every fund contributes the same share of the book's variance, so a fund that diversifies the others is given more money. It uses the full correlation matrix, which is why it is the harness for a claim about correlation. The sixth book re-runs the full shelf under inverse volatility, which sees each fund's own volatility and nothing else: a correlation-blind comparison, run to measure what the matrix is worth at fifteen funds. The eighteen seals of each walk sit close together on one August day because a supervisor registered each window's design and only then scored it, in order; the timestamps record that ordering.

The measurement, in the chart's own units. Dalio's exhibit plots portfolio volatility against the number of streams held, each stream at ten percent volatility, one curve per correlation level. We build that composite literally: each fund's daily return scaled to ten percent annualized volatility over the period being measured, equally weighted. The composite's volatility then inverts algebraically to the exact average pairwise correlation, and correlation converts to an effective number of bets, N over one plus N minus one times rho, the number of genuinely independent streams the menu behaves as. No estimation, no model; arithmetic on the same daily returns the walks traded.

The family and the rulers. Six book-and-weighting rules across five registered walks, declared as one search family of six before compilation; the deflated statistics in Table 1 use that count. The public ruler is a 60/40 SPY/AGG rebalanced monthly on the same total-return basis, the purple line. The platform's usual SPY reference is off on this page: every menu here holds SPY as one sleeve, the question is breadth against breadth, and a single equity index is not a ruler for that. The All Weather paper made the same call. The fifteen funds are among the most liquid in existence; every window priced every fund, and the census is empty. Sharpe, here and throughout, is each book's own daily series, the basis Table 1 prints.

2 Results

2.1 Headline

Trio (SPY, TLT, GLD), equal risk contribution, Sharpe

0.88

own daily series, Table 1 basis

The full fifteen-fund shelf, equal risk contribution, Sharpe

0.60

own daily series, Table 1 basis

The statistic this paper stands on

Fifteen funds weighted for equal risk returned 3.78 percent a year across eighteen registered windows; three funds returned 8.00 and led in sixteen windows of eighteen, 100.0 percent of paired bootstrap paths. Measured the way Dalio draws it, the full shelf held 3.3 effective bets against the fifteen the chart promises - and 2.1 in 2022.

60/40 SPY/AGG · study benchmark, Sharpe

0.69

7.84%/yr · worst drawdown -34.8% · computed

show or hide a line: Trio (SPY, TLT, GLD), equal risk contribution The full fifteen-fund shelf, equal risk contribution 60/40 SPY/AGG, rebalanced monthly (computed) All Seasons five (ERC) Curated seven, one per class (ERC) Ten funds (ERC) Fifteen funds, inverse volatility

Figure 1. Both arms stitched through the identical windows, Trio (SPY, TLT, GLD), equal risk contribution (+299.5%), The full fifteen-fund shelf, equal risk contribution (+94.9%), 60/40 SPY/AGG, rebalanced monthly (computed) (+288.8%), All Seasons five (ERC) (+144.2%), Curated seven, one per class (ERC) (+97.1%), Ten funds (ERC) (+99.4%), Fifteen funds, inverse volatility (+104.7%). The benchmark of this study is the 60/40 SPY/AGG line, the unit its own literature measures itself in. Dotted verticals mark the step boundaries; the dashed horizontal is break-even. These figures compound each arm's own stitched daily series; the pooled statistics in the text inner-join both arms' trading days, one session apart, both are printed from the frozen record.

**Table 1.** Both arms are the same shelf at different breadths: equal risk contribution, quarterly point-in-time re-estimation, one-year windows, total return net of costs. Sharpe is each book’s own daily series.
Book	Total return	Per year	Volatility	Sharpe	Worst drawdown	Mean window
Trio (SPY, TLT, GLD), equal risk contribution	+299.5%	8.00%	9.2%	0.88	-22.9%	+8.60%
60/40 SPY/AGG, rebalanced monthly (computed) (companion run)	+288.8%	7.84%	11.9%	0.69	-34.8%	+8.48%
All Seasons five (ERC) (companion run)	+144.2%	5.08%	7.1%	0.73	-18.4%	+5.34%
Fifteen funds, inverse volatility (companion run)	+104.7%	4.06%	6.9%	0.61	-20.0%	+4.31%
Ten funds (ERC) (companion run)	+99.4%	3.91%	7.1%	0.57	-19.8%	+4.14%
Curated seven, one per class (ERC) (companion run)	+97.1%	3.84%	6.9%	0.58	-18.9%	+4.07%
The full fifteen-fund shelf, equal risk contribution	+94.9%	3.78%	6.6%	0.60	-19.1%	+4.10%

Volatility, Sharpe and worst drawdown are computed on each book's own stitched daily series over the identical trading days that Figure 1 draws, so the panel and the figure are the same arithmetic. Sharpe carries no cash hurdle. Mean window is the arithmetic average of the one-year window returns and does not compound to the total beside it; the difference is volatility drag.

Figure 2. The measurement ladder: Trio (SPY, TLT, GLD), equal risk contribution's whole walk, chained three ways. price only (+196.1%), with dividends (+308.2%), net of costs (+299.5%). The distance between the two green pairs is the dividends collected; the sliver between the last two greens is the cost bill. Every other figure on this page uses the deepest rung, net of costs.

Figure 3. The same walk, re-based to 1.00x at the first window starting in 2010, 16 of the 18 windows above. Trio (SPY, TLT, GLD), equal risk contribution (+291.4%), The full fifteen-fund shelf, equal risk contribution (+90.9%), 60/40 SPY/AGG, rebalanced monthly (computed) (+323.2%). This is a subset of Figure 1, not a correction to it. The study is anchored before the 2007–09 crisis on purpose: a method that only works in calm markets should be caught doing it. But one crisis window and its equally singular recovery set the vertical scale for the whole of Figure 1, and everything after 2010 is compressed into the bottom of it. This figure shows the same windows, same method, same data, with that period outside the frame, so the post-crisis years can be read at their own scale. Neither figure stands on its own; the full record is what the study claims, and the era rows below put a number on how much of the gap came from which period.

**Figure 4.** Every menu, window by window. The bars shrink as the menu widens: the same green years and the same 2022 loss in every row, at roughly half the amplitude by the bottom of the ladder. The wide books out-earn the trio in 2009 and lose less in 2022.

**Figure 5.** Drawdown from peak. Every menu lives in the same 18-to-23 percent band against the 60/40's 34.8; what breadth changes is which crisis sets the worst drawdown: the bond-heavy narrow menus bottom in October 2022, the property-and-credit-carrying wide ones in 2008. The deepest mark in the family is the trio's 22.9, set in the year the shelf held 2.1 bets.

**Table 2.** Eighteen one-year windows, January 2008-2025, total return net of costs, chained on the charged basis Table 1 uses. DSR prob. is the probability the book survives deflation by the declared six-rule family.
Book	Total	CAGR % a year	Worst DD	DSR prob.
Trio: SPY, TLT, GLD (ERC)	+299.5%	8.00	-22.9 (2022-10)	0.99
All Seasons five (ERC)	+144.2%	5.08	-18.4 (2022-10)	0.97
Curated seven, one per class (ERC)	+97.1%	3.84	-18.9 (2008-11)	0.89
Ten funds (ERC)	+99.4%	3.91	-19.8 (2008-10)	0.88
Fifteen funds (ERC)	+94.9%	3.78	-19.1 (2022-10)	0.90
Fifteen funds, inverse volatility	+104.7%	4.06	-20.0 (2008-10)	0.91
60/40 SPY/AGG	+288.8%	7.84	-34.8 (2009-03)	ruler

**Table 3.** Each menu passed through Dalio’s own construction: every fund standardized to ten percent volatility, equally weighted; the composite’s volatility inverts to the exact mean pairwise correlation, and effective bets are N over one plus N minus one times rho. Correlations print to three decimals so the count recomputes from the row. Full record, 2008-2025.
Menu	Tickets	Mean pairwise corr	Effective bets	Realized book vol %	CAGR % a year
Trio	3	-0.032	3.2	9.2	8.00
All Seasons five	5	0.108	3.5	7.2	5.08
Curated seven	7	0.169	3.5	6.9	3.84
Ten	10	0.242	3.1	7.1	3.91
Fifteen	15	0.257	3.3	6.6	3.78
The chart’s promise	15	0.000	15.0	never happens

2.2 Per-step results

**Table 4.** One row per step, raw out-of-sample results. A short window can pair a negative return with a positive annualised Sharpe: at high daily volatility the arithmetic mean of daily returns sits above the compounded window return, and the Sharpe reads the former. Volatility drag, printed rather than smoothed.
#	Out-of-sample window	Trio (SPY, TLT, GLD), equal risk contribution SR	The full fifteen-fund shelf, equal risk contribution SR	All Seasons five (ERC) SR	Curated seven, one per class (ERC) SR	Ten funds (ERC) SR	Fifteen funds, inverse volatility SR
1	2008-01-02 → 2008-12-31	0.26	-0.20	0.24	-0.13	-0.30	-0.43
2	2009-01-02 → 2009-12-31	0.04	0.62	0.03	0.45	0.56	0.95
3	2010-01-04 → 2010-12-31	1.81	1.69	1.85	1.29	1.34	1.65
4	2011-01-03 → 2011-12-30	1.90	1.21	1.99	1.72	1.40	0.93
5	2012-01-03 → 2012-12-31	1.38	1.77	1.39	1.39	1.71	1.77
6	2013-01-02 → 2013-12-31	-0.04	-0.21	-0.37	-0.49	-0.11	-0.21
7	2014-01-02 → 2014-12-31	2.49	0.51	0.87	0.34	0.38	0.69
8	2015-01-02 → 2015-12-31	-0.38	-0.74	-0.84	-1.07	-0.96	-0.68
9	2016-01-04 → 2016-12-30	1.12	1.01	1.25	1.02	1.06	1.15
10	2017-01-03 → 2017-12-29	3.11	2.65	2.49	2.55	2.73	2.53
11	2018-01-02 → 2018-12-31	-0.24	-0.76	-0.28	-0.66	-0.84	-0.85
12	2019-01-02 → 2019-12-31	2.73	3.06	2.77	2.74	3.13	3.11
13	2020-01-02 → 2020-12-31	1.25	0.93	1.10	0.93	0.84	1.14
14	2021-01-04 → 2021-12-31	0.60	0.44	1.06	1.20	0.93	0.61
15	2022-01-03 → 2022-12-30	-1.27	-1.20	-1.01	-1.11	-1.17	-1.22
16	2023-01-03 → 2023-12-29	1.22	0.97	0.84	0.92	1.04	1.02
17	2024-01-02 → 2024-12-31	1.71	0.83	1.31	0.94	0.83	0.83
18	2025-01-02 → 2025-12-31	2.13	2.33	2.01	2.07	1.97	2.35

Figure 6. Trio (SPY, TLT, GLD), equal risk contribution: every step's out-of-sample curve overlaid, each rebased to 1× at its own start. Read alongside Table 1: consistent shape across steps is the walk-forward's evidence; a single lucky leg is not.

Figure 7. The full fifteen-fund shelf, equal risk contribution: the same windows, the other arm. Compare shape-for-shape with the previous figure: the two arms trade the identical out-of-sample legs.

2.3 Search accounting

This paper's search is a declared family: the Holy Grail family: six book-and-weighting rules on one fixed ETF shelf across five registered walks - the three-fund trio and the full fifteen-fund shelf under equal risk contribution as the two-arm comparative, the five-fund All Seasons sleeves, the curated seven (one fund per asset class) and the ten-fund book under equal risk contribution as single-arm walks, and the fifteen-fund shelf under inverse volatility as the correlation-blind twin - with a 60/40 SPY/AGG carried as the study benchmark, outside the family, counted at N = 6 evaluated books. Every member is either a registered walk with its own sealed hypothesis and frozen record, or a derived average computed from those frozen records; every member is reported, in the family matrix table and the robustness figure, and none was selected away. The count is declared by the author rather than derived from one project's ledger, because the members are sibling registered studies; the declaration names them and is frozen in this artifact. What the source strategy's author searched before publishing is not knowable from here and is not counted. The registered per-step record below still guarantees each window's hypothesis was hashed and registered before that window was scored.

2.4 The comparison

Both arms trade the same registered windows, so their returns can be PAIRED: inside each window the two return series are inner-joined date by date and the difference r_{Trio (SPY, TLT, GLD), equal risk contribution} − r_{The full fifteen-fund shelf, equal risk contribution} is the object under test. Because this is ONE pre-declared contrast, frozen at registration before any window was scored, the paired statistic needs no multiple-testing deflation; the arm-level records carry the declared family count of §2.3 as their search accounting, and this contrast, sealed per window before scoring, is not multiplied by it.

In the table: Arm A = Trio (SPY, TLT, GLD), equal risk contribution · Arm B = The full fifteen-fund shelf, equal risk contribution.

**Table 5.** Window-by-window paired comparison. Δ is the growth gap (Arm A − Arm B) over the window's paired dates. The All Seasons five (ERC) and Curated seven, one per class (ERC) and Ten funds (ERC) and Fifteen funds, inverse volatility columns carry companion books' growth over the full window; the Δ and Leader columns compare only this study's two registered arms.
#	Window	Paired bars	Arm A	Arm B	Δ	Leader	All Seasons five (ERC)	Curated seven, one per class (ERC)	Ten funds (ERC)	Fifteen funds, inverse volatility
1	2008-01-03 → 2008-12-31	252	+2.5%	-2.5%	+5.0 pp	Arm A	+1.8%	-2.0%	-4.0%	-5.6%
2	2009-01-05 → 2009-12-31	251	-0.2%	+4.9%	-5.1 pp	Arm B	-0.2%	+3.9%	+4.9%	+8.3%
3	2010-01-05 → 2010-12-31	251	+15.5%	+9.6%	+6.0 pp	Arm A	+12.6%	+8.3%	+8.5%	+10.2%
4	2011-01-04 → 2011-12-30	251	+17.6%	+7.3%	+10.2 pp	Arm A	+13.7%	+12.1%	+9.6%	+6.1%
5	2012-01-04 → 2012-12-31	249	+8.5%	+6.5%	+2.0 pp	Arm A	+6.4%	+6.2%	+7.4%	+6.9%
6	2013-01-03 → 2013-12-31	251	-0.6%	-1.4%	+0.7 pp	Arm A	-2.5%	-3.4%	-1.0%	-1.5%
7	2014-01-03 → 2014-12-31	251	+14.6%	+1.9%	+12.7 pp	Arm A	+3.8%	+1.2%	+1.4%	+2.4%
8	2015-01-05 → 2015-12-31	251	-3.2%	-4.1%	+0.9 pp	Arm A	-5.4%	-6.7%	-5.8%	-3.8%
9	2016-01-05 → 2016-12-30	251	+8.1%	+5.0%	+3.2 pp	Arm A	+7.0%	+5.6%	+5.7%	+5.6%
10	2017-01-04 → 2017-12-29	250	+15.2%	+9.9%	+5.4 pp	Arm A	+10.3%	+10.3%	+10.6%	+9.1%
11	2018-01-03 → 2018-12-31	250	-2.0%	-3.6%	+1.6 pp	Arm A	-1.8%	-3.6%	-4.4%	-3.9%
12	2019-01-03 → 2019-12-31	251	+20.1%	+12.5%	+7.6 pp	Arm A	+14.6%	+12.4%	+13.8%	+12.7%
13	2020-01-03 → 2020-12-31	252	+16.9%	+8.7%	+8.2 pp	Arm A	+9.8%	+8.3%	+8.5%	+11.7%
14	2021-01-05 → 2021-12-31	251	+4.8%	+2.0%	+2.7 pp	Arm A	+6.2%	+6.3%	+4.9%	+2.8%
15	2022-01-04 → 2022-12-30	250	-16.5%	-12.2%	-4.3 pp	Arm B	-11.6%	-11.2%	-12.5%	-12.6%
16	2023-01-04 → 2023-12-29	249	+12.5%	+7.5%	+5.0 pp	Arm A	+6.9%	+6.8%	+8.0%	+7.8%
17	2024-01-03 → 2024-12-31	251	+16.9%	+5.7%	+11.2 pp	Arm A	+9.5%	+5.7%	+5.4%	+5.5%
18	2025-01-03 → 2025-12-31	249	+24.2%	+16.0%	+8.2 pp	Arm A	+15.1%	+13.1%	+13.6%	+15.8%

Paired Sharpe of the difference track: 0.97 · block bootstrap (2000 paths, block 10, seed 1234): P(Trio (SPY, TLT, GLD), equal risk contribution beats The full fifteen-fund shelf, equal risk contribution) = 100.0%.

Window win-rate. Trio (SPY, TLT, GLD), equal risk contribution led 16 of 18 windows (88.9%), The full fifteen-fund shelf, equal risk contribution led 2, and the mean window gap of +4.51 pp points the same way. Widest single window: 2014 at +12.7 pp.

**Table 6.** The same comparison split at 2010. Pooling the whole walk into one row hides which side of the split the difference came from.
Period	Windows	Trio (SPY, TLT, GLD), equal risk contribution	The full fifteen-fund shelf, equal risk contribution	60/40 SPY/AGG	Mean gap	Trio (SPY, TLT, GLD), equal risk contribution led
All windows	18	+8.61%	+4.09%	+8.46%	+4.51 pp	16/18
Before 2010	2	+1.15%	+1.20%	-2.80%	-0.05 pp	1/2
2010 onward	16	+9.54%	+4.46%	+9.87%	+5.08 pp	15/16

Figure 8. Mean window return per period. Trio (SPY, TLT, GLD), equal risk contribution above, The full fifteen-fund shelf, equal risk contribution below, with the gap at right. The pooled bar and the post-2010 bar are the same comparison over different periods.

The two eras disagree by 5.13 pp. The pooled figure is therefore not a standing property of either method, it is dominated by the later period. Read the two rows, not the average.

3 The circuit

The strategy is a circuit of platform primitives, frozen when the study is registered. Below is the circuit as wired on the canvas, the objective it encodes and how the search runs through it, followed by the mathematics each primitive actually computes, the same formulas the execution engine runs. The complete parameterisation is preserved in the study ledger (Appendix A).

The hypothesis under test

A COMPARATIVE study: The trio (SPY, TLT, GLD), equal risk contribution vs The full fifteen-fund shelf, equal risk contribution, walked on the same registered out-of-sample windows. The trio (SPY, TLT, GLD), equal risk contribution: A configured universe, rebalanced quarterly across the selected basket, and run out-of-sample from the anchor: anything the design estimates from history, where it estimates at all, is re-estimated at each anchor from pre-anchor data only, and the walk advances through registered out-of-sample windows; its disposition is the realized forward path versus the benchmark. The full fifteen-fund shelf, equal risk contribution: A configured universe, rebalanced quarterly across the selected basket, and run out-of-sample from the anchor: anything the design estimates from history, where it estimates at all, is re-estimated at each anchor from pre-anchor data only, and the walk advances through registered out-of-sample windows; its disposition is the realized forward path versus the benchmark. The arms differ in: Universe, hg_book: trio → fifteen. The contrast under test: whether The trio (SPY, TLT, GLD), equal risk contribution generates better risk-adjusted returns than The full fifteen-fund shelf, equal risk contribution over the identical out-of-sample windows.

Figure 9. The frozen circuit, every node a primitive, every wire a typed data-flow; the two arms are colour-coded (Trio (SPY, TLT, GLD), equal risk contribution green, The full fifteen-fund shelf, equal risk contribution blue, shared feeds neutral). Each box is one step of the strategy; data flows along the wires left to right, and no box can see data dated later than the box feeding it. The whole diagram was frozen when the hypothesis was registered. Click any node to open what that step ran with and what it produced.

Envelopes show counts, ratios, dates, and the parameters the author chose. Full price and per-name data series are not republished: the underlying market data is licensed to QuanterLab. Point figures quoted in the prose, a named holding's return over a stated span, are summary facts derived from public market prices, not redistributed series.

What each part does

Universe, The starting set of tickers, the Holy Grail menus: rule-derived fixed lists of exchange-traded funds at five breadths of the liquid shelf, selectable by menu name only; no membership reconstruction applies.

Price Loader, Bulk OHLCV fetch for the whole universe, point-in-time, no future bars.

Portfolio Optimizer, MVO · HRP · IVOL, three rigorous ways to weight a pool.

Transaction Cost, Charge for trading, slippage + commission on every turn.

Portfolio Backtest, Replay the portfolio forward, rebalanced, point-in-time, with costs.

The objective and the search

The trio (SPY, TLT, GLD), equal risk contribution

Universe	holy_grail, named fixed instrument lists, no membership reconstruction applies: trio {SPY, TLT, GLD}.
Validation & out-of-sample	portfolio forward test (buy-and-hold book) (1y horizon from the anchor, quarterly rebalance); overlays: Transaction Cost.
Other components	Combine: Portfolio Optimizer.

The full fifteen-fund shelf, equal risk contribution

Universe	holy_grail, named fixed instrument lists, no membership reconstruction applies: fifteen {SPY, IWM, EFA, EWJ, VWO, TLT, IEF, TIP, LQD, AGG, GLD, SLV, DBC, VNQ, FXE}.

Every other specification row is identical to The trio (SPY, TLT, GLD), equal risk contribution's table above.

What differs between the arms, one difference; the comparison is clean:

paramUniverse, hg_book: trio → fifteen

Companion book. All Seasons five (ERC) comes from a companion run of the identical configuration on its own sealed anchor schedule. Its results run through this paper's tables under its own name.

Companion book. Curated seven, one per class (ERC) comes from a companion run of the identical configuration on its own sealed anchor schedule. Its results run through this paper's tables under its own name.

Companion book. Ten funds (ERC) comes from a companion run of the identical configuration on its own sealed anchor schedule. Its results run through this paper's tables under its own name.

Companion book. Fifteen funds, inverse volatility comes from a companion run of the identical configuration on its own sealed anchor schedule. Its results run through this paper's tables under its own name.

Everything else is held identical, so an out-of-sample gap between the arms is attributable to this one change.

Cost elements are wired into the circuit, the realised drag is reported per step in Appendix B.

Show the mathematics, 5 primitives, formulas and parity notes

3.1 Universe

The starting set of tickers, resolved point-in-time from the index change-log, so names delisted or removed later still compete on the dates they traded.

Before any math, you need a list of stocks. An index preset (S&P 500, Nasdaq-100, Dow 30) is reconstructed as it stood ON your anchor date by replaying the historical add/drop change-log backwards, so a 2018 backtest sees the 2018 membership, not today's winners.

Point-in-time membership

Start from today's constituents and un-apply every membership change after the anchor t:

\mathcal{U}(t) = \mathcal{U}_{\text{now}} \;\ominus\; \{\text{adds after } t\} \;\oplus\; \{\text{drops after } t\}

Constituents resolved from the index change-log; the same point-in-time set the factor + screening modules use.

3.2 Price Loader

Bulk OHLCV fetch for the whole universe, point-in-time, no future bars.

Momentum, volatility, trend, every price-based metric needs history. This loads open/high/low/close/volume for all names in parallel, clipped so nothing after the anchor can leak in. The lookback window is derived automatically from the deepest metric you wired.

The window is derived, not guessed

It loads exactly enough history for the hungriest downstream metric plus a warm-up buffer:

W = \max_k(\text{lookback}_k) + \text{buffer}, \qquad \text{bars} \le \text{anchor } t

3.3 Portfolio Optimizer

MVO · HRP · IVOL, three rigorous ways to weight a pool.

Beyond simple rules, three portfolio-theory optimizers. MVO maximizes return per unit of risk on the efficient frontier; HRP spreads risk across correlation clusters with no return forecast; IVOL is the simplest risk-balanced rule. All re-solve point-in-time at every rebalance.

MVO, maximize the Sharpe ratio

\max_{\mathbf w}\ \frac{\mathbf w^{\top}\boldsymbol\mu - r_f}{\sqrt{\mathbf w^{\top}\boldsymbol\Sigma\,\mathbf w}} \quad \text{s.t.}\ \ \mathbf 1^{\top}\mathbf w = 1,\ \ 0 \le w_i \le w_{\max}

Σ via Ledoit–Wolf shrinkage; long-only; negative-expected-return names excluded (SLSQP).

HRP, cluster, then split risk

d_{ij} = \sqrt{\tfrac12\,(1-\rho_{ij})} \ \to\ \text{hierarchical clusters}\ \to\ \text{recursive inverse-variance bisection}

No return forecast, robust to estimation error.

IVOL, inverse volatility

w_i = \frac{1/\sigma_i}{\sum_j 1/\sigma_j}

1% volatility floor. Fallback chain (IVOL → per-name → equal) is disclosed if a method degenerates.

Byte-identical to FM101FBKT optimizer engines (shared_libs/factor_core).

3.4 Portfolio Backtest

Replay the portfolio forward, rebalanced, point-in-time, with costs.

Holds the basket and rebalances on schedule, re-selecting and re-optimizing point-in-time at each rebalance (so it only ever uses information available then), and reports the equity curve, Sharpe, drawdown and trade stats, optionally net of cost and risk overlays.

Compounded equity

E_t = E_{t-1}\big(1 + \mathbf w_{t}^{\top}\mathbf r_t - \text{costs}_t\big)

Drawdown

\text{DD}_t = \frac{E_t}{\max_{\tau\le t}E_\tau} - 1, \qquad \text{MaxDD} = \min_t \text{DD}_t

Financing a levered book

\text{charge}_t \;=\; \text{loan}_t \cdot \frac{\text{spread}}{252}, \qquad \text{loan}_t = \begin{cases}\max(0,\,-\text{cash}_t) & \text{institutional (netted)}\\ \max(0,\,\text{long MV}_t - E_t) & \text{retail (no netting)}\end{cases}

A levered long/short book (the β-neutral Long/Short Select) borrows its excess notional. WHO you are decides the loan: a prime broker nets short-sale proceeds against the margin loan, a fully-netted BAB book carries almost none, while a retail margin account cannot net, so the same book borrows the long leg’s excess over equity. Profiles: institutional = 50 bps spread + 25 bps GC short borrow; retail = 350 bps + 150 bps (the BEST retail tier, so any verdict is conservative); custom = your own knobs, spread on negative cash. Charged daily, reported as financing_drag_pct, never silent.

3.5 Transaction Cost

Charge for trading, slippage + commission on every turn.

Real trading isn't free. This deducts a cost proportional to how much you trade (turnover), in basis points, so the backtest reflects net, not gross, performance.

Cost per rebalance

\text{cost}_t = \frac{\text{bps}}{10{,}000}\;\times\;\text{turnover}_t, \qquad \text{turnover}_t = \tfrac12\sum_i \lvert w_{i,t}-w_{i,t^-}\rvert

4 Projection calibration, pooled across the walk

Every rebalance carried a Monte Carlo cone and a 95% VaR estimated before the segment it is scored against. Two questions, pooled over the whole study: did realized outcomes land inside the band as often as the band claims, and were VaR breaches as frequent as 5%?

This section is produced by the forward tester itself: every portfolio backtest fits the cone and the VaR estimate at each rebalance and scores them against the segment that followed. It does not require, and this circuit does not contain, a Monte Carlo primitive; that primitive is a separate, standalone analysis.

Figure 10. Projected range versus what occurred, at each of 144 scored rebalance segments, pooled across both arms. The final rebalance of each step has no following segment to score, the ledger marks those rows “no segment follows this rebalance”, which is why this count sits below the raw rebalance totals in the table beneath. Each vertical bar is that rebalance's P5–P95 Monte Carlo cone with the median ticked; the dot is the realized return of the segment that followed. Filled green = the outcome landed inside its own cone; red = it did not. The strip beneath repeats that as one mark per rebalance, so a run of misses in one period is visible as a run. Every cone was fitted only on data prior to the segment it is scored against.

Arm	Steps	Rebalances	In band	Coverage	Expected	VaR days	Breach rate	Expected
Trio (SPY, TLT, GLD), equal risk contribution	18	85	62 / 72	86.1% ±4.08	90.0%	4457	5.79% ±0.35	5.0%
The full fifteen-fund shelf, equal risk contribution	18	85	61 / 72	84.7% ±4.24	90.0%	4457	5.65% ±0.346	5.0%

Note. The platform grades its own risk model in the appendix: across the six books the 90 percent Monte Carlo cones contained 84.7 to 86.1 percent of realized rebalance segments, and the one-day 95 percent VaR was breached on 5.30 to 5.81 percent of days against the 5 expected. The same appendix explains the occasional window that pairs a negative return with a positive annualised Sharpe: volatility drag, stated once.

± values are binomial standard errors on the estimate. A coverage figure below the expected band means the projection was over-confident; a breach rate above 5% means the same of the risk model. Both forecasts used only data prior to the segment scored.

5 Discussion

5.1 Findings

The bets ledger explains every result that follows. The trio's three assets are independent: average pairwise correlation minus 0.03, effective bets 3.2 from 3 tickets. Equities, duration and gold do not move together. Widen to the All Seasons five and the count rises to just 3.5, and that is the ceiling: seven funds hold 3.5, ten hold 3.1, all fifteen hold 3.3. Every fund added past the trio duplicates an exposure the book already holds. The year-by-year strip is sharper: the count fell to 2.1 in 2022 and has sat under 3 since. It is lowest in the years when diversification is needed most.

The walked returns follow that ledger, and Table 1 carries the full run. The return falls in two steps, three points of annual rate between three funds and five, another point between five and seven, and then nothing: seven, ten and fifteen funds all sit at roughly 3.8 to 3.9 a year with Sharpe between 0.57 and 0.60. Volatility does fall: 9.2 percent for the trio down to 6.6 for the shelf, a 28 percent reduction, which is what a correlation of 0.26 permits and nowhere near the promised three quarters. What the shelf delivered was half the compound rate for a quarter off the volatility, Sharpe 0.60 against the trio's 0.88.

The host comparative is one-sided. The trio led the fifteen-fund shelf in sixteen windows of eighteen, by 4.2 points of compound rate, in 100.0 percent of two thousand paired block-bootstrap paths. The two windows breadth won are the two the correlation ledger predicts. 2009: the junk rebound, plus 4.9 against minus 0.2, carried by emerging markets, credit and silver, none of which the trio holds. 2022: minus 12.3 against minus 16.6, four points of cushion in the year stocks and long bonds fell together and everything the trio held fell with them.

The 2022 cushion came from the inflation lines, not from breadth. The two smallest losses that year were the seven-fund book's 11.2 and the five-fund book's 11.6: both carry the commodity sleeve, the seven adds inflation-linked Treasuries, and the five's second, shorter Treasury line diluted the long-bond loss. Ten and fifteen funds lost 12.5 and 12.3, because small caps, international equity and credit fell with SPY and diluted the cushion. In the crash year the wider books lost more.

Which funds are added matters more than how many. The five-fund book beat the curated seven by 1.2 points a year. The seven swapped the second Treasury line for property and a currency pair, and paid for it twice: VNQ carries equity beta and dragged the seven's worst drawdown back to November 2008, and FXE is a driftless pair that consumed a seventh of the risk budget.

Reading the correlation matrix added nothing at fifteen funds. The inverse-volatility fifteen, blind to every off-diagonal, returned 4.06 a year at Sharpe 0.61, against equal risk contribution's 3.78 and 0.60 on the identical menu: inside noise, slight edge to the blind book. Our All Weather study found the same tie at five funds. On a long-only fund shelf the off-diagonals carry little information the diagonals do not, and scaling by own volatility does the work. I expected the matrix to matter at fifteen names, where there is more of it to read. It did not, and this is the second study that says so.

Every menu's worst drawdown lands in the same 18-to-23 band against the 60/40's 34.8, so the diversification claim holds at that level regardless of breadth; what breadth changes is which crisis sets the mark. The bond-heavy narrow menus bottomed in October 2022, the property-and-credit-carrying wide ones in 2008. The deepest of the family is the trio's 22.9, set in the year the shelf held 2.1 bets, which is the caveat that belongs beside the trio's 8.00 a year.

5.2 Interpretation

The chart is true. Nothing in this record contradicts the arithmetic: at correlation zero the risk of fifteen equal streams is a quarter of one stream's, and if the correlation on offer were zero, the fifteen-fund composite would have printed 2.6 percent volatility instead of 5.5. The 5.5 is the standardized composite of Dalio's construction; the walked book, un-standardized, ran 6.6. The chart assumes a correlation the shelf does not supply. Measured, it is 0.26. Bridgewater's own version of the grail is built from institutional return streams, alphas across asset classes, geographies and horizons, engineered to be near-independent. The ETF shelf sells repackagings of three exposures: equity beta, duration, and the dollar. Pass the fifteen most liquid funds through Dalio's own construction and the count comes back 3.3.

The bets count prices the rest of the record. A shelf that sells 3.5 independent bets makes every menu wider than five a dilution: the floor at 3.8 to 3.9 a year for seven, ten and fifteen funds follows from holding the same three exposures at the lower carry of the funds added. The ledger was computed before a single window was walked, and the walked ladder came out in its order.

The trio's 8.00 a year carries a caveat. Three-fund concentration won this era because this era paid duration and gold, the same regime that carried the All Seasons recipe past its optimized cousins in our All Weather study. The one year the era argument inverted, 2022, the trio printed the family's deepest loss while the widest books cushioned it. A reader who takes the trio from this paper and expects 8 percent forever is ignoring the strip under the exhibit: the count has printed between 2.1 and 2.9 since 2022, and at that level the trio's 3.2 is most of what is available.

The correlation-blind book is the third result. Equal risk contribution uses the full matrix and inverse volatility uses only the diagonal, and on the same fifteen funds they finish inside noise of each other, as they did at five funds in our previous study. On shelves where every fund is one of three exposures, the off-diagonals carry almost no information the diagonals did not already have. What would change the result is a genuinely new stream, and the liquid shelf does not offer one.

The strip under the exhibit is the part a reader can use today. The count swung between 2.7 and 4.5 through 2021, printed 2.1 in 2022, then 2.4, 2.4 and 2.9. Stock-bond correlation turned positive in 2022 and the count has stayed under 3 in the three years since: the shelf's diversification budget sits at the low of the record, and a book built from this shelf today starts from 2 to 3 independent bets, not fifteen. The appendix grades the platform's own risk model on this record and the grade is mixed, cones a touch too narrow and the daily loss floor broken slightly more often than designed; the numbers are printed there, not smoothed here.

Run your own fifteen. The construction is arithmetic on daily returns, the menus are named lists, and the circuit on this page runs unchanged on any list of tickers: swap the menu, walk the same eighteen windows, read your own ledger.

This study is one member of a declared search family: the same design walked at several sealed settings across sibling registered projects, every member either a registered walk with its own frozen record or a derived average of those records, and every member reported. The family size is declared by the author and named in the lineage; it is the search-accounting count for this paper. What was searched before the source strategy was published is not knowable from here and is not counted.

5.3 Limitations

The record covers one era, and the era favoured the trio's assets. Fourteen of eighteen windows sit in the long bond bull; the trio's margin is concentrated there, and 2022 is the one window of the opposite regime. Two windows sit before 2010 and they split one each; everything in the pooled gap comes from the sixteen windows after. The return ladder's ordering is era-shaped and presented as description. The bets ledger is the more durable result, the shelf's correlation structure held between 2.7 and 4.5 effective bets through every regime in the record, but the 8.00-versus-3.78 gap is one era's record.

The menus follow stated rules, and the author chose the rules. Five ways of slicing one shelf, plus one weighting variation, declared as a family of six before compilation; the deflated probabilities in Table 1 carry that count, and every book survives it at 0.88 or better. No other breadths, weightings or fund substitutions were run. The fifteen-fund shelf itself is one opinion of what the liquid shelf is; a different fifteen would move the constants.

The correlation is measured daily, standardized within the period being measured, and the effective-bet count is the average-correlation formula, not an eigenvalue decomposition. Daily correlation understates crisis co-movement at longer horizons; measured monthly, the 2022 reading would come in below 2.1.

Costs are ten basis points per one-way traded dollar across the board, which treats FXE and SLV spreads the same as SPY's and flatters the wide menus slightly; the direction runs against the paper's conclusion. Taxes and financing are absent, and nothing here is levered.

References

QuanterLab reference architecture

Gelman, A., & Loken, E. (2013). The garden of forking paths: Why multiple comparisons can be a problem, even when there is no “fishing expedition.” Working paper, Columbia University.
Harvey, C. R., Liu, Y., & Zhu, H. (2016). … and the Cross-Section of Expected Returns. Review of Financial Studies, 29(1), 5–68. doi:10.1093/rfs/hhv059
Lo, A. W. (2002). The Statistics of Sharpe Ratios. Financial Analysts Journal, 58(4), 36–52. doi:10.2469/faj.v58.n4.2453

Author’s references?

Dalio, R. (2017). Principles: Life and Work. Simon and Schuster. (The Holy Grail of investing: fifteen good, uncorrelated return streams; the risk-versus-streams exhibit.)
Bridgewater Associates. The All Weather Story. Bridgewater research notes. (Return streams engineered for independence as the institutional form of the same claim.)
Meucci, A. (2009). Managing Diversification. Risk 22(5), 74-79. (The effective number of bets as a diversification measure; this paper uses the average-correlation form, not Meucci’s eigenvalue form.)
Maillard, S., Roncalli, T., Teiletche, J. (2010). The Properties of Equally Weighted Risk Contribution Portfolios. Journal of Portfolio Management 36(4), 60-70.
QuanterLab Research (2026). Dalio's All Weather, constructed and then tested against alternatives. quanterlab.com/research/dalio-s-all-weather-constructed-and-then-tested-against-inverse-volatility-equal (Part one of this pair: the published recipe held up against its optimized cousins, and equal risk contribution tied inverse volatility at five funds.)

Appendix A Reproducibility in QuanterLab

Each step is backed by a frozen run report. The study is re-derivable from the ledger below.

#	Commit	Report	Anchor	OOS window
1	6579bb1f9c29	2086	2008-01-01	2008-01-02 → 2008-12-31
2	12df1f47e3c6	2087	2009-01-01	2009-01-02 → 2009-12-31
3	07a5b0f4f6e5	2088	2010-01-01	2010-01-04 → 2010-12-31
4	715f905d74e9	2089	2011-01-01	2011-01-03 → 2011-12-30
5	b7448bd6cd37	2090	2012-01-01	2012-01-03 → 2012-12-31
6	9d7e1cd603a4	2091	2013-01-01	2013-01-02 → 2013-12-31
7	29d8a115e054	2092	2014-01-01	2014-01-02 → 2014-12-31
8	1041906f8d63	2093	2015-01-01	2015-01-02 → 2015-12-31
9	f2bf4d3b0957	2094	2016-01-01	2016-01-04 → 2016-12-30
10	ef2e11702d2d	2095	2017-01-01	2017-01-03 → 2017-12-29
11	12f4461f4339	2096	2018-01-01	2018-01-02 → 2018-12-31
12	6a8a86aba926	2097	2019-01-01	2019-01-02 → 2019-12-31
13	9932c07302b8	2098	2020-01-01	2020-01-02 → 2020-12-31
14	bebf0c9d8acd	2099	2021-01-01	2021-01-04 → 2021-12-31
15	d1a98f4b3aa2	2100	2022-01-01	2022-01-03 → 2022-12-30
16	a23bac1781cc	2101	2023-01-01	2023-01-03 → 2023-12-29
17	142f113a8b4b	2102	2024-01-01	2024-01-02 → 2024-12-31
18	b0715cb820e2	2103	2025-01-01	2025-01-02 → 2025-12-31

Appendix A2 Registration record

What this record does and does not establish. Every window in this study is historical: the data existed before the study began, so this is sequential sealing on past windows, not pre-registration in the clinical-trial sense, and no procedure could make it so. What the platform does enforce is order, each step's specification was frozen and hashed before that step was scored, and the walk cannot advance past a step that was never run or close one with a result registered for a different window. The two timestamp columns below are the evidence: read them together and each seal precedes its own run, and each run precedes the next seal. A study whose seals all post-date its runs would show it here. Wall-clock spacing between seals varies with the author's schedule and queue latency; the ordering, not the tempo, is the claim.

“A COMPARATIVE study: The trio (SPY, TLT, GLD), equal risk contribution vs The full fifteen-fund shelf, equal risk contribution, walked on the same registered out-of-sample windows. The trio (SPY, TLT, GLD), equal risk contribution: A configured universe, rebalanced quarterly across the selected basket, and run out-of-sample from the anchor: anything the design estimates from history, where it estimates at all, is re-estimated at each anchor from pre-anchor data only, and the walk advances through registered out-of-sample windows; its disposition is the realized forward path versus the benchmark. The full fifteen-fund shelf, equal risk contribution: A configured universe, rebalanced quarterly across the selected basket, and run out-of-sample from the anchor: anything the design estimates from history, where it estimates at all, is re-estimated at each anchor from pre-anchor data only, and the walk advances through registered out-of-sample windows; its disposition is the realized forward path versus the benchmark. The arms differ in: Universe, hg_book: trio → fifteen. The contrast under test: whether The trio (SPY, TLT, GLD), equal risk contribution generates better risk-adjusted returns than The full fifteen-fund shelf, equal risk contribution over the identical out-of-sample windows.”

The same hypothesis was registered independently at every step, hashed before each step's out-of-sample window was scored:

**Table 7.** Registration audit, one row per registered step, with the time each specification was frozen and the time its window was scored. The hypothesis is identical on every row by design: it was registered once and re-registered unchanged at each anchor. Rows that differ would mean the specification moved mid-walk, which is the thing this record exists to rule out. The timestamps are the separate claim: each seal precedes its own run, and each run precedes the next seal.
#	Anchor	Registered at (UTC)	Run completed (UTC)
1	2008-01-01	2026-08-30 16:13:54	2026-08-30 16:14:12
2	2009-01-01	2026-08-30 16:14:15	2026-08-30 16:14:30
3	2010-01-01	2026-08-30 16:14:33	2026-08-30 16:14:48
4	2011-01-01	2026-08-30 16:14:51	2026-08-30 16:15:06
5	2012-01-01	2026-08-30 16:15:09	2026-08-30 16:15:24
6	2013-01-01	2026-08-30 16:15:28	2026-08-30 16:15:43
7	2014-01-01	2026-08-30 16:15:46	2026-08-30 16:16:01
8	2015-01-01	2026-08-30 16:16:04	2026-08-30 16:16:19
9	2016-01-01	2026-08-30 16:16:22	2026-08-30 16:16:37
10	2017-01-01	2026-08-30 16:16:40	2026-08-30 16:16:55
11	2018-01-01	2026-08-30 16:16:58	2026-08-30 16:17:13
12	2019-01-01	2026-08-30 16:17:17	2026-08-30 16:17:32
13	2020-01-01	2026-08-30 16:17:35	2026-08-30 16:17:50
14	2021-01-01	2026-08-30 16:17:53	2026-08-30 16:18:08
15	2022-01-01	2026-08-30 16:18:11	2026-08-30 16:18:26
16	2023-01-01	2026-08-30 16:18:29	2026-08-30 16:18:44
17	2024-01-01	2026-08-30 16:18:47	2026-08-30 16:19:02
18	2025-01-01	2026-08-30 16:19:06	2026-08-30 16:19:21

Appendix B Per-step diagnostics

Realized in the projection tables below is the risk engine scoring its own forecast: the buy-and-hold return of the segment that followed each rebalance, on the same gross basis the cone was projected on. It is deliberately not the charged, calendar-window total return the study’s tables print, so the two will not reconcile line by line; the cone and its outcome share one basis, which is what a calibration test requires. Each row names its segment’s span so a boundary session is visible.

What each step's run actually did beyond its return: capital allocation across lanes and regimes, the portfolio book's rebalancing and cost drag, and how positions were sized. Harvested from the frozen run reports, present where the circuit produced them.

Open the full per-step grid (18 steps: every rebalance, capital routing and sizing, per window)

Step 1 · 2008-01-02 → 2008-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 100.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 14.06% of 249 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2008-01-01 → 2008-04-01	-3.8179%	2.2123%	8.0181%	-0.1098%	yes	0.6463%	7 / 60
2008-04-01 → 2008-07-01	-4.689%	1.8037%	8.1107%	-1.413%	yes	0.7651%	4 / 63
2008-07-01 → 2008-10-01	-4.7416%	1.5279%	7.6054%	-2.3743%	yes	0.7188%	5 / 63
2008-10-01 → window end	-4.7961%	1.2%	6.9978%	4.8647%	yes	0.7072%	19 / 63

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.4× · cost drag 0.14%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 75.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 14.86% of 249 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2008-01-01 → 2008-04-01	-2.9987%	0.9908%	4.7575%	1.8393%	yes	0.4559%	5 / 60
2008-04-01 → 2008-07-01	-3.1159%	1.4006%	5.7035%	-0.6519%	yes	0.5322%	5 / 63
2008-07-01 → 2008-10-01	-3.406%	1.1388%	5.4703%	-6.0267%	no	0.533%	8 / 63
2008-10-01 → window end	-3.8635%	0.6606%	4.9725%	0.6148%	yes	0.5333%	19 / 63

Step 2 · 2009-01-02 → 2009-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 75.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 7.66% of 248 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2009-01-01 → 2009-04-01	-6.1784%	1.9239%	9.8901%	-7.0387%	no	0.9894%	10 / 60
2009-04-01 → 2009-07-01	-7.79%	0.8958%	10.4576%	-2.852%	yes	1.1159%	6 / 62
2009-07-01 → 2009-10-01	-9.1758%	0.6322%	10.4989%	7.8694%	yes	1.1996%	0 / 63
2009-10-01 → 2010-01-01	-8.4156%	1.8259%	12.1627%	-0.4622%	yes	1.1838%	3 / 63
2010-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2009-01-01 → 2009-04-01	-5.5089%	0.7894%	6.8748%	-3.9656%	yes	0.6558%	12 / 60
2009-04-01 → 2009-07-01	-6.4925%	0.1362%	7.2776%	1.272%	yes	0.7528%	6 / 62
2009-07-01 → 2009-10-01	-7.1168%	0.3%	7.5801%	5.585%	yes	0.8255%	0 / 63
2009-10-01 → 2010-01-01	-6.664%	1.012%	8.5625%	-0.3398%	yes	0.7628%	2 / 63
2010-01-01	no segment follows this rebalance, not scored

Step 3 · 2010-01-04 → 2010-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2010-01-01 → 2010-04-01	-10.1643%	0.4719%	11.223%	0.7462%	yes	1.2461%	2 / 60
2010-04-01 → 2010-07-01	-9.8438%	0.1421%	11.302%	4.7319%	yes	1.2487%	0 / 62
2010-07-01 → 2010-10-01	-9.5857%	0.9684%	11.6656%	7.697%	yes	1.2198%	0 / 63
2010-10-01 → 2011-01-01	-8.6477%	2.116%	13.0359%	0.2287%	yes	1.2204%	3 / 63
2011-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2010-01-01 → 2010-04-01	-7.9033%	0.1244%	8.0229%	0.1397%	yes	0.8118%	2 / 60
2010-04-01 → 2010-07-01	-7.8678%	-0.1597%	8.2437%	0.833%	yes	0.8555%	0 / 62
2010-07-01 → 2010-10-01	-7.712%	0.1752%	7.9556%	5.9573%	yes	0.8407%	0 / 63
2010-10-01 → 2011-01-01	-6.7509%	1.1747%	8.9896%	-0.1474%	yes	0.8323%	4 / 63
2011-01-01	no segment follows this rebalance, not scored

Step 4 · 2011-01-03 → 2011-12-30

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.3× · cost drag 0.13%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2011-01-01 → 2011-04-01	-5.841%	3.1371%	13.9167%	1.4805%	yes	1.1445%	0 / 61
2011-04-01 → 2011-07-01	-4.9464%	2.9172%	11.4828%	1.8192%	yes	1.031%	0 / 62
2011-07-01 → 2011-10-01	-4.002%	3.3815%	10.6103%	6.0909%	yes	0.846%	5 / 63
2011-10-01 → 2012-01-01	-3.5853%	3.2827%	10.6842%	3.7518%	yes	0.8288%	4 / 62
2012-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2011-01-01 → 2011-04-01	-4.1412%	2.4253%	10.1072%	1.4665%	yes	0.7674%	1 / 61
2011-04-01 → 2011-07-01	-3.5826%	2.2608%	8.4954%	1.3751%	yes	0.7199%	1 / 62
2011-07-01 → 2011-10-01	-3.0783%	2.4122%	7.6909%	-0.4292%	yes	0.629%	10 / 63
2011-10-01 → 2012-01-01	-3.2703%	1.6173%	6.7825%	2.1437%	yes	0.5963%	5 / 62
2012-01-01	no segment follows this rebalance, not scored

Step 5 · 2012-01-03 → 2012-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2012-01-01 → 2012-04-01	-3.2508%	3.0601%	10.4193%	2.1564%	yes	0.8138%	2 / 61
2012-04-01 → 2012-07-01	-3.2617%	3.326%	10.4046%	2.1133%	yes	0.8025%	2 / 62
2012-07-01 → 2012-10-01	-3.5203%	2.9739%	9.9468%	3.6647%	yes	0.7933%	1 / 62
2012-10-01 → window end	-2.7226%	3.2243%	10.1312%	-2.6177%	yes	0.7392%	1 / 61

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2012-01-01 → 2012-04-01	-2.7706%	1.4027%	6.1627%	1.5889%	yes	0.5551%	2 / 61
2012-04-01 → 2012-07-01	-2.6699%	1.6337%	6.1542%	0.3199%	yes	0.5361%	2 / 62
2012-07-01 → 2012-10-01	-2.8103%	1.4267%	5.8747%	2.3009%	yes	0.542%	0 / 62
2012-10-01 → window end	-2.4324%	1.4174%	5.7932%	-0.5246%	yes	0.5142%	0 / 61

Step 6 · 2013-01-02 → 2013-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2013-01-01 → 2013-04-01	-3.7106%	2.589%	8.3193%	1.3346%	yes	0.7336%	2 / 59
2013-04-01 → 2013-07-01	-3.2603%	2.9015%	8.8632%	-6.6047%	no	0.6846%	9 / 63
2013-07-01 → 2013-10-01	-4.9631%	1.5631%	7.9057%	1.2692%	yes	0.7463%	5 / 63
2013-10-01 → 2014-01-01	-5.282%	1.5389%	8.1877%	1.1234%	yes	0.7847%	3 / 63
2014-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.3× · cost drag 0.13%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 75.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 5.24% of 248 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2013-01-01 → 2013-04-01	-3.2204%	1.1174%	4.9925%	0.2668%	yes	0.5603%	1 / 59
2013-04-01 → 2013-07-01	-2.8096%	1.3164%	5.2324%	-4.5845%	no	0.4567%	7 / 63
2013-07-01 → 2013-10-01	-4.375%	0.2853%	4.7338%	1.4503%	yes	0.6238%	4 / 63
2013-10-01 → 2014-01-01	-5.1055%	0.3646%	5.6285%	-0.0026%	yes	0.6723%	1 / 63
2014-01-01	no segment follows this rebalance, not scored

Step 7 · 2014-01-02 → 2014-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.1× · cost drag 0.11%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2014-01-01 → 2014-04-01	-5.5972%	0.7074%	6.7997%	4.4798%	yes	0.7311%	1 / 60
2014-04-01 → 2014-07-01	-4.9579%	0.7736%	6.887%	4.5012%	yes	0.6858%	0 / 62
2014-07-01 → 2014-10-01	-4.7892%	1.297%	7.1869%	-0.6806%	yes	0.6487%	3 / 63
2014-10-01 → 2015-01-01	-5.2309%	0.8618%	6.7599%	4.2409%	yes	0.6704%	4 / 63
2015-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2014-01-01 → 2014-04-01	-4.7216%	0.2081%	4.9077%	2.2439%	yes	0.5581%	1 / 60
2014-04-01 → 2014-07-01	-4.1663%	0.2077%	4.8086%	2.766%	yes	0.5074%	0 / 62
2014-07-01 → 2014-10-01	-4.0331%	0.6687%	5.1578%	-3.8702%	yes	0.4911%	4 / 63
2014-10-01 → 2015-01-01	-4.1466%	0.2259%	4.3878%	-0.7554%	yes	0.5005%	3 / 63
2015-01-01	no segment follows this rebalance, not scored

Step 8 · 2015-01-02 → 2015-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2015-01-01 → 2015-04-01	-4.9052%	1.0791%	6.8419%	1.3223%	yes	0.6853%	4 / 60
2015-04-01 → 2015-07-01	-4.9086%	1.2044%	7.7495%	-4.387%	yes	0.6971%	5 / 62
2015-07-01 → 2015-10-01	-5.7837%	0.6502%	6.9009%	-2.2066%	yes	0.735%	7 / 63
2015-10-01 → 2016-01-01	-5.3737%	0.854%	6.8911%	0.2086%	yes	0.7552%	5 / 63
2016-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.1× · cost drag 0.11%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2015-01-01 → 2015-04-01	-4.251%	-0.0509%	3.9245%	0.6539%	yes	0.4929%	4 / 60
2015-04-01 → 2015-07-01	-4.45%	-0.0494%	4.5812%	-2.4614%	yes	0.5307%	2 / 62
2015-07-01 → 2015-10-01	-5.0767%	-0.35%	4.1663%	-3.0284%	yes	0.5352%	4 / 63
2015-10-01 → 2016-01-01	-4.4926%	-0.2293%	3.8251%	-1.0757%	yes	0.5141%	3 / 63
2016-01-01	no segment follows this rebalance, not scored

Step 9 · 2016-01-04 → 2016-12-30

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 50.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 4.44% of 248 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2016-01-01 → 2016-04-01	-5.6145%	0.4402%	6.2772%	7.7613%	no	0.7851%	0 / 60
2016-04-01 → 2016-07-01	-4.5803%	1.6218%	7.6297%	5.1855%	yes	0.7136%	1 / 63
2016-07-01 → 2016-10-01	-4.3954%	1.8136%	7.8279%	-0.2085%	yes	0.6921%	4 / 63
2016-10-01 → window end	-4.5901%	1.3745%	7.7501%	-6.4047%	no	0.7195%	6 / 62

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2016-01-01 → 2016-04-01	-4.5559%	-0.6183%	3.0998%	4.2143%	no	0.5011%	0 / 60
2016-04-01 → 2016-07-01	-4.0018%	-0.029%	3.7378%	3.4118%	yes	0.4567%	2 / 63
2016-07-01 → 2016-10-01	-3.7313%	0.2327%	3.9904%	0.3986%	yes	0.446%	3 / 63
2016-10-01 → window end	-3.8354%	0.0689%	4.1558%	-4.3838%	no	0.4662%	5 / 62

Step 10 · 2017-01-03 → 2017-12-29

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2017-01-01 → 2017-04-01	-5.1883%	0.4923%	7.082%	4.2149%	yes	0.7474%	0 / 61
2017-04-01 → 2017-07-01	-5.5548%	0.406%	6.7811%	1.869%	yes	0.7092%	1 / 62
2017-07-01 → 2017-10-01	-4.7608%	0.9514%	7.0426%	3.1153%	yes	0.6342%	1 / 62
2017-10-01 → 2018-01-01	-4.119%	1.5377%	7.564%	4.0363%	yes	0.5958%	1 / 62
2018-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.3× · cost drag 0.13%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2017-01-01 → 2017-04-01	-4.1758%	-0.3078%	4.0926%	2.1284%	yes	0.4595%	3 / 61
2017-04-01 → 2017-07-01	-4.4534%	-0.1144%	4.4486%	1.2549%	yes	0.4855%	0 / 62
2017-07-01 → 2017-10-01	-4.1585%	0.1431%	4.6644%	2.4074%	yes	0.485%	0 / 62
2017-10-01 → 2018-01-01	-3.7125%	0.7542%	5.4558%	2.3371%	yes	0.4978%	0 / 62
2018-01-01	no segment follows this rebalance, not scored

Step 11 · 2018-01-02 → 2018-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.5× · cost drag 0.15%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2018-01-01 → 2018-04-01	-3.4054%	1.916%	7.0045%	-1.5258%	yes	0.5625%	14 / 60
2018-04-01 → 2018-07-01	-4.0766%	1.8063%	7.4863%	-0.2464%	yes	0.644%	3 / 63
2018-07-01 → 2018-10-01	-4.7858%	0.6376%	6.4042%	-0.3189%	yes	0.6707%	0 / 62
2018-10-01 → 2019-01-01	-5.4987%	-0.2837%	5.2519%	-0.2465%	yes	0.6752%	3 / 62
2019-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.3× · cost drag 0.13%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2018-01-01 → 2018-04-01	-3.1791%	1.0766%	5.1049%	-1.325%	yes	0.4478%	12 / 60
2018-04-01 → 2018-07-01	-2.7395%	1.1874%	4.9073%	-0.2156%	yes	0.4043%	2 / 63
2018-07-01 → 2018-10-01	-3.2208%	0.3013%	3.9733%	-0.1799%	yes	0.4188%	3 / 62
2018-10-01 → 2019-01-01	-3.508%	-0.1752%	3.2931%	-2.6363%	yes	0.4226%	5 / 62
2019-01-01	no segment follows this rebalance, not scored

Step 12 · 2019-01-02 → 2019-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 75.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 6.85% of 248 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2019-01-01 → 2019-04-01	-5.4172%	0.0766%	5.344%	4.8083%	yes	0.6891%	4 / 60
2019-04-01 → 2019-07-01	-3.827%	1.0854%	6.2794%	6.5536%	no	0.6444%	1 / 62
2019-07-01 → 2019-10-01	-3.6283%	1.5896%	6.5949%	5.5623%	yes	0.6275%	8 / 63
2019-10-01 → 2020-01-01	-3.6219%	1.9696%	7.352%	1.6676%	yes	0.6537%	4 / 63
2020-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 75.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 3.63% of 248 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2019-01-01 → 2019-04-01	-3.7114%	-0.3277%	2.8494%	4.1921%	no	0.4214%	1 / 60
2019-04-01 → 2019-07-01	-2.8374%	0.3094%	3.5773%	2.96%	yes	0.4045%	0 / 62
2019-07-01 → 2019-10-01	-2.7663%	0.5835%	3.7395%	1.9287%	yes	0.3973%	6 / 63
2019-10-01 → 2020-01-01	-2.9827%	0.5091%	3.8034%	1.5203%	yes	0.416%	2 / 63
2020-01-01	no segment follows this rebalance, not scored

Step 13 · 2020-01-02 → 2020-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.3× · cost drag 0.13%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 75.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 8.84% of 249 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2020-01-01 → 2020-04-01	-3.3738%	1.8282%	7.8265%	-1.0702%	yes	0.6731%	8 / 61
2020-04-01 → 2020-07-01	-5.6365%	1.6966%	9.6473%	10.5838%	no	0.7709%	3 / 62
2020-07-01 → 2020-10-01	-4.6541%	3.5812%	11.7114%	4.3922%	yes	0.7368%	8 / 63
2020-10-01 → window end	-4.4229%	4.1712%	12.682%	1.5752%	yes	0.8216%	3 / 63

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.5× · cost drag 0.15%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 50.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 9.24% of 249 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2020-01-01 → 2020-04-01	-2.903%	0.4199%	4.1767%	-5.9714%	no	0.4478%	10 / 61
2020-04-01 → 2020-07-01	-5.1869%	-0.388%	4.6838%	7.5349%	no	0.4844%	6 / 62
2020-07-01 → 2020-10-01	-4.2636%	0.9961%	6.0451%	3.5188%	yes	0.4898%	6 / 63
2020-10-01 → window end	-3.7374%	1.5936%	6.7132%	3.4882%	yes	0.503%	1 / 63

Step 14 · 2021-01-04 → 2021-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 75.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 4.84% of 248 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2021-01-01 → 2021-04-01	-4.3055%	4.3091%	12.8062%	-7.8867%	no	0.8846%	7 / 60
2021-04-01 → 2021-07-01	-5.4248%	2.9975%	12.2254%	4.8978%	yes	0.9839%	1 / 62
2021-07-01 → 2021-10-01	-5.5431%	3.6693%	12.8534%	-0.2431%	yes	0.9599%	3 / 63
2021-10-01 → 2022-01-01	-6.4067%	3.0136%	12.4304%	5.2192%	yes	0.9751%	1 / 63
2022-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.5× · cost drag 0.15%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2021-01-01 → 2021-04-01	-3.1003%	2.1378%	7.1419%	-3.2971%	no	0.4775%	6 / 60
2021-04-01 → 2021-07-01	-3.6788%	1.5223%	7.037%	2.7015%	yes	0.5199%	2 / 62
2021-07-01 → 2021-10-01	-4.2856%	1.9086%	7.9074%	-0.5572%	yes	0.5488%	2 / 63
2021-10-01 → 2022-01-01	-4.7758%	1.6848%	7.9591%	0.7459%	yes	0.587%	2 / 63
2022-01-01	no segment follows this rebalance, not scored

Step 15 · 2022-01-03 → 2022-12-30

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.1× · cost drag 0.11%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 50.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 10.93% of 247 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2022-01-01 → 2022-04-01	-5.6558%	2.8215%	12.9464%	-2.6663%	yes	0.9602%	7 / 61
2022-04-01 → 2022-07-01	-6.8573%	1.8753%	12.3441%	-12.0062%	no	1.0918%	8 / 61
2022-07-01 → 2022-10-01	-8.0634%	0.3067%	8.6038%	-8.4312%	no	1.0874%	6 / 63
2022-10-01 → 2023-01-01	-10.3717%	-2.2302%	6.7047%	2.9746%	yes	1.159%	6 / 62
2023-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2022-01-01 → 2022-04-01	-4.6407%	1.2461%	8.0873%	-3.4588%	yes	0.6418%	7 / 61
2022-04-01 → 2022-07-01	-5.5269%	0.8226%	8.2412%	-8.5834%	no	0.7185%	10 / 61
2022-07-01 → 2022-10-01	-5.3573%	0.5388%	6.2364%	-7.8148%	no	0.7536%	12 / 63
2022-10-01 → 2023-01-01	-7.6179%	-1.6033%	4.8412%	3.7073%	yes	0.8274%	5 / 62
2023-01-01	no segment follows this rebalance, not scored

Step 16 · 2023-01-03 → 2023-12-29

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.1× · cost drag 0.11%

Projection accuracy, realized outcome fell inside the P5–P95 cone in 75.0% of 4 scored rebalances (a well-calibrated 90% band contains ~90.0%) · 95% VaR breached on 2.44% of 246 days (expected ~5.0%)

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2023-01-01 → 2023-04-01	-9.6006%	-1.4193%	8.3582%	6.7169%	yes	1.1739%	2 / 61
2023-04-01 → 2023-07-01	-9.156%	-0.788%	9.2281%	-0.1418%	yes	1.171%	1 / 61
2023-07-01 → 2023-10-01	-9.3653%	-0.5327%	9.2196%	-6.7221%	yes	1.1873%	2 / 62
2023-10-01 → 2024-01-01	-10.6211%	-1.7816%	7.9912%	12.6962%	no	1.1704%	1 / 62
2024-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.1× · cost drag 0.11%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2023-01-01 → 2023-04-01	-7.1658%	-0.8603%	6.5122%	3.5407%	yes	0.8501%	4 / 61
2023-04-01 → 2023-07-01	-7.6516%	-1.1417%	6.489%	-0.9601%	yes	0.8919%	1 / 61
2023-07-01 → 2023-10-01	-8.2325%	-1.2698%	6.2664%	-3.8466%	yes	0.907%	3 / 62
2023-10-01 → 2024-01-01	-8.9591%	-1.8906%	5.7728%	7.3533%	no	0.9172%	1 / 62
2024-01-01	no segment follows this rebalance, not scored

Step 17 · 2024-01-02 → 2024-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2024-01-01 → 2024-04-01	-9.9938%	-0.2846%	9.4415%	6.1316%	yes	1.1572%	1 / 60
2024-04-01 → 2024-07-01	-8.9066%	0.4087%	10.7393%	2.8661%	yes	1.15%	4 / 62
2024-07-01 → 2024-10-01	-8.3611%	1.6072%	11.6419%	8.8989%	yes	1.1068%	2 / 63
2024-10-01 → window end	-7.1466%	2.5401%	12.2539%	-2.708%	yes	1.0971%	2 / 63

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 4 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.2× · cost drag 0.12%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2024-01-01 → 2024-04-01	-8.9528%	-1.0562%	6.7103%	2.5516%	yes	0.9667%	1 / 60
2024-04-01 → 2024-07-01	-8.1619%	-0.6541%	7.5166%	1.0148%	yes	0.966%	3 / 62
2024-07-01 → 2024-10-01	-7.7094%	0.1885%	7.9802%	5.6504%	yes	0.9257%	0 / 63
2024-10-01 → window end	-6.3769%	0.9712%	8.1753%	-5.0523%	yes	0.8492%	2 / 63

Step 18 · 2025-01-02 → 2025-12-31

Trio (SPY, TLT, GLD), equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 3 names held · selection: reselect · 0.0% in cash · turnover 1.3× · cost drag 0.13%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2025-01-01 → 2025-04-01	-5.478%	3.4211%	11.7181%	4.5965%	yes	0.9902%	1 / 59
2025-04-01 → 2025-07-01	-5.3579%	2.0356%	10.7665%	4.3501%	yes	1.0062%	5 / 61
2025-07-01 → 2025-10-01	-6.4169%	2.5959%	11.5714%	7.6449%	yes	1.0167%	0 / 63
2025-10-01 → 2026-01-01	-5.0154%	3.9431%	12.8489%	3.3115%	yes	0.9935%	3 / 63
2026-01-01	no segment follows this rebalance, not scored

The full fifteen-fund shelf, equal risk contribution

Portfolio book, rebalanced quarterly · 5 constructions · 15 names held · selection: reselect · 0.0% in cash · turnover 1.3× · cost drag 0.13%

Rebalance	P5	Median	P95	Realized	In band	VaR 95 (1d)	Breaches
2025-01-01 → 2025-04-01	-5.157%	1.3771%	7.3387%	3.6886%	yes	0.7252%	0 / 59
2025-04-01 → 2025-07-01	-4.8758%	0.3886%	6.4679%	3.1682%	yes	0.7039%	5 / 61
2025-07-01 → 2025-10-01	-5.0019%	1.0208%	6.8464%	3.8055%	yes	0.6916%	0 / 63
2025-10-01 → 2026-01-01	-4.2134%	1.7211%	7.4541%	1.5955%	yes	0.6355%	1 / 63
2026-01-01	no segment follows this rebalance, not scored

QuanterLab · Study 8d6b33cb5ec2 · compiled August 30, 2026. Point-in-time constituents and hypothesis-registration timestamps are enforced by the platform. This report is generated from the frozen study artifact and is reproducible from the ledger above. Educational research, not investment advice: every result on this page is simulated, and nothing here is a recommendation to buy or sell any security.

Dalio's Holy Grail, Measured and Then Walked at Five Breadths: Three, Five, Seven, Ten and Fifteen Funds

1 Methodology

2 Results

2.1 Headline

2.2 Per-step results

2.3 Search accounting

2.4 The comparison

3 The circuit

The objective and the search

3.1 Universe

3.2 Price Loader

3.3 Portfolio Optimizer

3.4 Portfolio Backtest

3.5 Transaction Cost

4 Projection calibration, pooled across the walk

5 Discussion

5.1 Findings

5.2 Interpretation

5.3 Limitations

References

Appendix A Reproducibility in QuanterLab

Appendix A2 Registration record

Appendix B Per-step diagnostics

Step 1 · 2008-01-02 → 2008-12-31

Step 2 · 2009-01-02 → 2009-12-31

Step 3 · 2010-01-04 → 2010-12-31

Step 4 · 2011-01-03 → 2011-12-30

Step 5 · 2012-01-03 → 2012-12-31

Step 6 · 2013-01-02 → 2013-12-31

Step 7 · 2014-01-02 → 2014-12-31

Step 8 · 2015-01-02 → 2015-12-31

Step 9 · 2016-01-04 → 2016-12-30

Step 10 · 2017-01-03 → 2017-12-29

Step 11 · 2018-01-02 → 2018-12-31

Step 12 · 2019-01-02 → 2019-12-31

Step 13 · 2020-01-02 → 2020-12-31

Step 14 · 2021-01-04 → 2021-12-31

Step 15 · 2022-01-03 → 2022-12-30

Step 16 · 2023-01-03 → 2023-12-29

Step 17 · 2024-01-02 → 2024-12-31

Step 18 · 2025-01-02 → 2025-12-31

Run a study like this one

New research, as it publishes

One click left, check your inbox