As seen on Quantocracy
All research
QuanterLab · Research

Two Kinds of Crash, Two Kinds of Protection: the 200-Day Line and Volatility Targeting, 2000 to 2025

How this study was run: the companies, the method, the dates
Universe · Single tickers (no stock selection, so index membership does not apply)
Method · Comparative: Volatility targeting vs The 200-day line
Manipulated variable ·
The one difference between the arms is how each guards against a fall. Arm A, the volatility target, always owns the fund. The volatility target sets the share of the position it holds at each close from how much the fund has been moving: the standard deviation of its daily returns over the last 21 … (full registered statement)The one difference between the arms is how each guards against a fall. Arm A, the volatility target, always owns the fund. The volatility target sets the share of the position it holds at each close from how much the fund has been moving: the standard deviation of its daily returns over the last 21 trading days (recent moves) against the same over the past five years (normal moves; a fund with less history reads what it has, from its first year on). It holds normal divided by recent of the position, never more than all of it, so twice as jumpy as normal holds half. Arm B is the 200-day line. The 200-day line holds the whole book in the fund while its close is above its 200-day average and sells to cash at the first close below it; it buys back at the first close above. Both arms trade the S&P 500 fund (SPY) from each first of January for a year. Nothing is fitted: every setting is fixed before the walk, the same in every window. Signals and the volatility reading are taken at the close and held from that close; cash earns nothing; both pay 0.02% of the traded value on every trade, from the same Transaction Cost card, and the volatility target pays it on every resize too; prices leave out dividends, for both arms and for buy and hold alike, which the forward tests print beside the SPY benchmark.
Step size · 1 year per forward window
Out-of-sample span · 2000-01-03 → 2025-12-31
Compiled · October 05, 2026
Search family · the paper's 28 walks (N = 28, every member reported)
Abstract

Anyone who held stocks through 2008 or the Covid crash has asked the same question: is there a simple rule that gets you out before the worst of it? Two answers are famous. The 200-day line sells when the price ends the day below its average of the last 200 trading days. Volatility targeting holds less of the fund when its daily moves grow larger than usual, and more when they calm down.

We tested both rules, and both together, on SPY, the fund that follows the S&P 500, from 2000 to 2025, and on 13 other stock, bond and gold funds. Each rule had one setting, fixed before any test ran. We asked what protection means for a person who holds through a fall: how deep the fall goes, how long it takes to get back where it started, and how much growth the rule gives up over the years.

The two rules protect against different crashes. In the two long bear markets, the dot-com crash and the 2008 crisis, the 200-day line fell less than half as far as holding SPY and was back where it started years sooner. In the sudden crashes, volatility targeting fell less than holding every time, and in 2018 and 2020 it was back months before the line. Over the whole test, $1 held in SPY became $4.69, against $4.88 with volatility targeting and $3.10 with the 200-day line.

The six falls of SPY since 2000, side by side. Each line starts at 100 on the day SPY set its record and runs until SPY was back at it. Grey: holding SPY. Blue: the 200-day line. Green: volatility targeting. Amber: both together. In the two great bear markets the blue line falls much less than the grey one. In 2018 and 2020 the green line falls less than holding and is back months before the blue one. In 2022 and 2025 the amber line lies on top of the blue one: both together held almost exactly what the line held.
Figure 1. The six falls of SPY since 2000, side by side. Each line starts at 100 on the day SPY set its record and runs until SPY was back at it. Grey: holding SPY. Blue: the 200-day line. Green: volatility targeting. Amber: both together. In the two great bear markets the blue line falls much less than the grey one. In 2018 and 2020 the green line falls less than holding and is back months before the blue one. In 2022 and 2025 the amber line lies on top of the blue one: both together held almost exactly what the line held.

Why this one

Most investors who use the 200-day line think of it as crash protection. In this study it protected in the two long bear markets, the dot-com crash and the 2008 crisis. In a sudden crash it sold after much of the fall had happened and bought back after most of the rebound. Volatility targeting, the rule fund managers use to decide how much to hold by how much the market is moving, made the sudden crashes shallower.

In this paper protection has to do three things at once: make the fall smaller, make the way back shorter, and give up little growth over the years. A rule that sells after every dip shows a small worst fall, and it also misses the rises that make most of the long-run growth.

The 200-day line is the best known trend rule in investing. Chart readers watch it and financial news quotes it, and our earlier paper tested it against a Kalman filter, a method that smooths the price to find its trend. Volatility targeting is a professional's rule: fund managers use it, and index providers publish versions of the S&P 500 that hold less when the daily moves grow. We ran both on the same funds and years as the Kalman paper, so the two papers can be read side by side.

How the two rules work

The 200-day line watches the price. It holds the whole position while the fund ends the day above its average price of the last 200 trading days. It sells all of it the first day the fund ends below that average, and buys it all back the first day it ends above. So it acts only after the price has fallen far enough to cross its average.

Volatility targeting watches the daily moves. At the end of every trading day it compares how much the fund moved over the last month, 21 trading days, with how much it moved over the past five years. When the daily moves have been twice as large as usual, it holds half the position; three times as large, a third. It never holds more than the whole position, so in calm times it simply holds the fund. It can act before the price has fallen far, because the daily moves often grow first.

Both together uses the 200-day line to decide whether to own the fund, and volatility targeting to decide how much of it to own. So there are three ways to protect, the two rules and both together, and each is compared with simply holding the fund. Every change of position, the small daily ones included, pays a trading cost.

Six falls

SPY fell 15% or more from an all-time high six times in these years. Three falls were slow: the dot-com crash, the 2008 crisis and the 2022 slide. The first two were long bear markets, years long and deep; the 2022 slide was shorter and shallower. Three falls were sudden and hit bottom within about three months: the 2018 drop, the Covid crash and the 2025 tariff drop. The figure at the top of the paper shows all six side by side. For each one we measured how far each way of holding fell, and how long it took to get back where it started, to its value on the day of SPY's high. The tables further down have every number (Table 1).

In the two long bear markets the 200-day line protected far better than volatility targeting. In the 2008 crisis holding SPY fell by more than half; the 200-day line fell by about a sixth, and it was back where it started more than three years sooner. Volatility targeting fell by more than a third. In the dot-com crash volatility targeting did not help at all. That fall was slow, and its daily moves were not much larger than usual, so volatility targeting held most of SPY all the way down. It was back where it started only after the 2008 crisis, later than any other way of holding.

The sudden crashes were different. In the Covid crash, volatility targeting began to cut within a week of SPY's high and held only about a fifth of SPY at the bottom. Holding fell by a third; volatility targeting fell by about a sixth, and it was back where it started a little before holding. The 200-day line sold only when SPY was already well below its high, sold and bought back again during the worst days, and bought back for good only after most of the rebound. It was back where it started five months after volatility targeting. In the 2018 drop the two rules fell about as far as each other, and volatility targeting was back months before the line, though holding itself was back sooner than both.

Two falls did not follow this pattern. In the 2025 tariff drop the 200-day line fell less than volatility targeting, and both rules were back later than holding. In the 2022 slide, with only moderate daily moves, all three ways to protect fell almost as far as holding, and the 200-day line sold and bought back again and again.

Both together had the smallest fall in five of the six falls. After the 2018 drop and the Covid crash, though, it was back months later than volatility targeting alone.

What each rule gave up

Over the whole test, $1 held in SPY became $4.69, against $4.88 with volatility targeting: it kept all of holding's growth and a little more. A rule that never holds more than the whole fund can end with more because it held less on the most turbulent days, when the losses were biggest, and a smaller loss needs a smaller gain to get back. With the 200-day line $1 became $3.10, and with both together $3.51, both less than holding. The line was out of the market more than a quarter of the time.

Part of what the line gave up came in years when SPY moved sideways near its average. Then the line sold and bought back again and again, nearly always at a higher price than it sold. In the calm years before the 2008 crisis, SPY rose by about a fifth while the line gained almost nothing.

So did anything beat simply holding SPY? On growth, volatility targeting did, by a small margin. On the size of the worst fall, all three ways to protect did, and measured against how bumpy the ride was, all three did too (Table 2).

The 13 other funds show a similar result. On all 14 funds, SPY included, volatility targeting made the worst fall smaller, on gold by almost nothing, and on nine of the 14 it grew at least as fast as holding. The 200-day line also cut the worst fall on all 14, by more than volatility targeting on 11 of them, but it grew more slowly than holding on 13 of the 14. The one exception is the financials fund, where being out of the market in the 2008 crisis was worth more than everything the line missed (Tables 3 and 4).

How much of SPY each rule held, 2000 to 2025. Top: what $1 in SPY became, with the six falls shaded. Bottom: the 200-day line is all in or all out (blue shading); volatility targeting (green line) holds less as the daily moves grow, down to about a fifth in 2008 and 2020, and never much below half in 2000 to 2002.
Figure 2. How much of SPY each rule held, 2000 to 2025. Top: what $1 in SPY became, with the six falls shaded. Bottom: the 200-day line is all in or all out (blue shading); volatility targeting (green line) holds less as the daily moves grow, down to about a fifth in 2008 and 2020, and never much below half in 2000 to 2002.
1  Methodology, in detail (click to open)

1  Methodology

What we tested. Four ways of holding a fund. Holding it every day. The 200-day line: the whole position while the close is above its 200-day average, nothing while it is below. Volatility targeting: always in the fund, holding the normal size of its daily moves divided by the recent size, never more than the whole position. The recent size is the standard deviation of daily returns over the last 21 trading days; the normal size is the same over the past five years, or since the fund began when it is younger. Both together: the 200-day line decides whether to hold, volatility targeting how much.

Funds and years. SPY from 2000, the Nasdaq-100 fund QQQ from 2002, gold (GLD) from 2007, and from 2006 the Russell 2000 fund IWM, long Treasury bonds (TLT) and the nine sector funds of the S&P 500. These are the funds and years of our Kalman paper. All tests end on 31 December 2025.

How the tests ran. Each year was its own test, run forward from the first trading day of January to the last of December, with every setting fixed before any test ran and the same every year; the platform locked each year's test before it ran and keeps every result. Two comparisons per fund, volatility targeting against the 200-day line and both together against the 200-day line, make 28 tests; none was dropped. The 200-day line ran in both comparisons and came out the same to the cent each time. All 28 tests were locked the same way, each year before it ran, and every test also prints holding its fund beside its two rules. The record further down lists all 28 in numbers, under Every test, in numbers. The ledger in the appendix is this page's own test, volatility targeting against the 200-day line on SPY; each of the other 27 tests keeps a ledger of its own on the platform.

Prices and costs. Daily closing prices without dividends, from a licensed commercial data provider. A signal is read at the close and the position is held from that close. Every purchase and sale pays 0.02% of the amount traded, the small daily resizes of volatility targeting included. Cash earns nothing.

One continuous path. Each yearly test starts in cash and buys at the close of its first trading day, so the yearly tests joined end to end would miss the first trading day of every year, for every rule and for holding too. Our figures and tables carry each rule through those days with the share of the fund it held the day before, so holding SPY matches SPY's own record. The yearly results kept on the platform, which the headline and the record below use, leave those days out. That lowers growth by 0.3 to 0.4 points a year on SPY, deepens the worst falls of holding and volatility targeting by 1.5 points, and changes none of the comparisons in this paper. The Sharpe ratios in the SPY table are the platform's, read on those yearly records, so they match the headline. Our Kalman paper printed SPY on the yearly records too: holding 5.7% a year and the 200-day line 4.1%, against 6.1% and 4.5% here.

Measuring a fall. A fall is a drop of 15% or more from a record close of SPY. It lasts from the record close until SPY closes back at that level. It is fast when SPY's lowest close came within 100 calendar days of the record, slow otherwise. For each way of holding, the depth is the lowest value inside that span, measured from its own value on the record day, and the time back is the number of trading days from the record until it is back at that value again, counted after its lowest point. The longest wait for a new high in the SPY table is measured from each way of holding's own high to its next higher close, whenever that high was.

2  Results

Findings

The two long bear markets: the 200-day line protected far better than volatility targeting. In the dot-com crash and the 2008 crisis it fell less than half as far as holding SPY and was back where it started years sooner (Table 1). Volatility targeting made the 2008 fall shallower but did not help in the dot-com crash.

Sideways years: the 200-day line sold and bought back again and again. In the calm years before the 2008 crisis SPY rose by about a fifth while the line gained almost nothing.

The sudden crashes: volatility targeting fell less than holding in all three. In the 2018 drop and the Covid crash it fell about as far as the 200-day line or less, and it was back where it started months before it. In the Covid crash holding fell by a third and volatility targeting by about a sixth.

Two exceptions. In the 2025 tariff drop the 200-day line fell less than volatility targeting, and all three ways to protect were back later than holding. In the 2022 slide all three ways to protect fell almost as far as holding.

The whole test on SPY: volatility targeting grew a little faster than holding, with a smaller worst fall. The 200-day line had less than half of holding's worst fall and ended with about two thirds of holding's money (Table 2).

The 14 funds: volatility targeting cut the worst fall on all 14 and grew at least as fast as holding on nine of the 14. The 200-day line cut the worst fall on all 14, by more than volatility targeting on 11 of the 14, and grew more slowly than holding on 13 of the 14 (Tables 3 and 4).

2.1  Headline

Volatility targeting, Sharpe
0.46
day by day, every year the test ran, 2000 to 2025
The 200-day line, Sharpe
0.42
day by day, every year the test ran, 2000 to 2025
SPY, Sharpe
0.38
the same days, stitched across the windows
The result
In the 2008 crisis holding SPY fell by more than half and the 200-day line by about a sixth. In the Covid crash holding fell by a third and volatility targeting by about a sixth.
The 14 funds, from the first year in the tables to the end of 2025. Left: the worst fall, where further right is a smaller fall. Right: growth a year. Volatility targeting (green) had a smaller worst fall than holding (grey) on every fund and grew about as fast. The 200-day line (blue) cut the fall on every fund, on most by more than volatility targeting, and grew more slowly on all but financials.
Figure 3. The 14 funds, from the first year in the tables to the end of 2025. Left: the worst fall, where further right is a smaller fall. Right: growth a year. Volatility targeting (green) had a smaller worst fall than holding (grey) on every fund and grew about as fast. The 200-day line (blue) cut the fall on every fund, on most by more than volatility targeting, and grew more slowly on all but financials.
What $1 in SPY became from January 2000 to December 2025, prices without dividends, after trading costs. The scale is logarithmic: equal steps up the side are equal percentage gains.
Figure 4. What $1 in SPY became from January 2000 to December 2025, prices without dividends, after trading costs. The scale is logarithmic: equal steps up the side are equal percentage gains.
Sections 2.2 to 3, the full record: every year, every test, and how each one was run (click to open)

The registration names the two groups compared The volatility target and The 200-day line; this paper calls them Volatility targeting and The 200-day line.

Table 1. The six falls of SPY: how far each fell, and how long until it was back. From the day SPY set its record close to the day it closed back at it: the deepest each way of holding fell below its own value on the record day, and the time from the record until it was back at that value. Fast: SPY reached bottom within 100 days of the record.

FallKindHolding SPYThe 200-day lineVolatility targetingBoth together
2000 to 2002slow−49%, back in 7.2 years−22%, back in 3.8 years−50%, back in 10.8 years−19%, back in 3.7 years
2007 to 2009slow−57%, back in 5.4 years−17%, back in 2.0 years−38%, back in 3.1 years−15%, back in 23 months
2018fast−20%, back in 7 months−14%, back in 16 months−15%, back in 9 months−12%, back in 16 months
2020fast−34%, back in 6 months−20%, back in 11 months−16%, back in 6 months−14%, back in 9 months
2022slow−25%, back in 2.0 years−21%, back in 2.2 years−23%, back in 23 months−20%, back in 2.2 years
2025fast−19%, back in 4 months−10%, back in 6 months−16%, back in 6 months−10%, back in 6 months

Table 2. SPY from January 2000 to December 2025. Prices without dividends, 0.02% a trade, cash earns nothing. Longest wait for a new high: from a high of that way of holding to its next higher close. Sharpe ratio: return for each unit of risk, on the platform's yearly records like the headline; on the continuous path the other columns follow, each is 0.02 to 0.03 higher, in the same order.

Way of holding$1 becameA yearWorst fallLongest wait for a new highShare of SPY heldSharpe, as in the headline
Holding SPY$4.696.12%−56.5%7.2 years (2000 to 2007)100%0.38
The 200-day line$3.104.45%−25.7%6.8 years (2000 to 2006)72% of the days0.42
Volatility targeting$4.886.28%−50.2%10.3 years (2000 to 2011)92% on average0.46
Both together$3.514.95%−21.6%4.0 years (2000 to 2004)70% on average0.48

Table 3. The 14 funds: growth a year. From the year shown to December 2025, counted as in the table for SPY.

FundFromHoldingThe 200-day lineVolatility targetingBoth together
S&P 500 (SPY)20006.1%4.5%6.3%5.0%
Nasdaq-100 (QQQ)200212.0%9.8%11.9%9.7%
Russell 2000 (IWM)20066.6%3.5%7.0%3.5%
Long Treasuries (TLT)2006−0.3%−0.5%−0.6%−1.1%
Gold (GLD)200710.2%7.3%9.5%7.6%
Materials (XLB)20065.5%−0.7%5.5%−0.1%
Energy (XLE)20062.7%1.7%3.2%1.5%
Financials (XLF)20063.8%5.8%6.5%5.5%
Industrials (XLI)20068.3%6.1%8.9%6.1%
Technology (XLK)200613.9%11.3%13.7%11.4%
Consumer staples (XLP)20066.2%0.0%6.2%0.8%
Utilities (XLU)20065.0%0.0%5.7%0.6%
Health care (XLV)20068.2%3.1%8.0%3.7%
Consumer discretionary (XLY)200610.4%8.4%10.9%8.4%

Table 4. The 14 funds: the worst fall. The deepest fall from a high, from the year shown to December 2025.

FundFromHoldingThe 200-day lineVolatility targetingBoth together
S&P 500 (SPY)2000−56.5%−25.7%−50.2%−21.6%
Nasdaq-100 (QQQ)2002−53.6%−27.0%−51.5%−25.4%
Russell 2000 (IWM)2006−59.9%−34.1%−42.2%−32.2%
Long Treasuries (TLT)2006−51.8%−38.7%−48.2%−38.1%
Gold (GLD)2007−45.6%−34.6%−45.5%−32.4%
Materials (XLB)2006−60.7%−47.2%−41.1%−38.4%
Energy (XLE)2006−76.7%−47.4%−68.4%−48.1%
Financials (XLF)2006−83.7%−25.8%−58.9%−27.2%
Industrials (XLI)2006−63.3%−21.3%−44.3%−20.4%
Technology (XLK)2006−53.5%−25.6%−38.7%−21.7%
Consumer staples (XLP)2006−34.2%−31.3%−21.7%−27.8%
Utilities (XLU)2006−48.8%−36.5%−38.3%−31.0%
Health care (XLV)2006−40.6%−31.6%−30.3%−23.0%
Consumer discretionary (XLY)2006−60.1%−23.3%−42.9%−21.6%

2.2  Per-step results

Table 5. 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 Volatility targeting SR The 200-day line SR
1 2000-01-03 → 2000-12-29 -0.44 -1.18
2 2001-01-02 → 2001-12-31 -0.85 n/a
3 2002-01-02 → 2002-12-31 -1.23 -1.55
4 2003-01-02 → 2003-12-31 1.31 1.89
5 2004-01-02 → 2004-12-31 0.80 0.21
6 2005-01-03 → 2005-12-30 0.39 -0.09
7 2006-01-03 → 2006-12-29 1.18 0.88
8 2007-01-03 → 2007-12-31 0.27 -0.34
9 2008-01-02 → 2008-12-31 -1.25 -1.03
10 2009-01-02 → 2009-12-31 0.97 1.08
11 2010-01-04 → 2010-12-31 0.76 -0.01
12 2011-01-03 → 2011-12-30 -0.08 -0.96
13 2012-01-03 → 2012-12-31 0.94 0.83
14 2013-01-02 → 2013-12-31 2.22 2.22
15 2014-01-02 → 2014-12-31 1.04 0.96
16 2015-01-02 → 2015-12-31 -0.07 -0.88
17 2016-01-04 → 2016-12-30 0.88 0.89
18 2017-01-03 → 2017-12-29 2.58 2.58
19 2018-01-02 → 2018-12-31 -0.42 -0.76
20 2019-01-02 → 2019-12-31 1.85 1.21
21 2020-01-02 → 2020-12-31 0.93 0.26
22 2021-01-04 → 2021-12-31 2.00 2.02
23 2022-01-03 → 2022-12-30 -0.90 -2.46
24 2023-01-03 → 2023-12-29 1.77 0.82
25 2024-01-02 → 2024-12-31 1.77 1.78
26 2025-01-02 → 2025-12-31 0.84 1.00
Out-of-sample equity: normalised growth (1.00x = break even)0.65x0.99x1.34xbars into the window →
Figure 5. Volatility targeting: every step's out-of-sample curve overlaid, each rebased to 1× at its own start. Read alongside the per-step table: consistent shape across steps is the walk-forward's evidence; a single lucky leg is not.
Out-of-sample equity: normalised growth (1.00x = break even)0.75x1.04x1.33xbars into the window →
Figure 6. The 200-day line: the same windows, the other arm. Compare shape-for-shape with the previous figure: the two arms trade the identical out-of-sample windows.

2.2b  Every test, in numbers

Every test this paper registered, two rows each, the paper’s own test first.

WalkWindowsSpanGrowth CAGRWorst drawdownPooled Sharpe
SPY · volatility targeting (this paper, the one the lab opens) 26 2000-01-03 → 2025-12-31 +341.9% +5.9% -51.7% 0.46
SPY · the 200-day line (this paper, the one the lab opens) 26 2000-01-03 → 2025-12-31 +187.1% +4.1% -25.7% 0.42
TLT · volatility targeting 20 2006-01-03 → 2025-12-31 -9.0% -0.5% -47.3% 0.03
TLT · the 200-day line 20 2006-01-03 → 2025-12-31 -5.8% -0.3% -36.7% 0.03
QQQ · volatility targeting 24 2002-01-02 → 2025-12-31 +1212.3% +11.3% -51.5% 0.66
QQQ · the 200-day line 24 2002-01-02 → 2025-12-31 +836.2% +9.8% -25.9% 0.70
TLT · both together 20 2006-01-03 → 2025-12-31 -16.3% -0.9% -36.1% -0.06
TLT · the 200-day line, second run 20 2006-01-03 → 2025-12-31 -5.8% -0.3% -36.7% 0.03
QQQ · both together 24 2002-01-02 → 2025-12-31 +810.1% +9.6% -24.3% 0.72
QQQ · the 200-day line, second run 24 2002-01-02 → 2025-12-31 +836.2% +9.8% -25.9% 0.70
SPY · both together 26 2000-01-03 → 2025-12-31 +225.4% +4.6% -21.6% 0.48
SPY · the 200-day line, second run 26 2000-01-03 → 2025-12-31 +187.1% +4.1% -25.7% 0.42
XLB · volatility targeting 20 2006-01-03 → 2025-12-31 +178.8% +5.3% -41.7% 0.37
XLB · the 200-day line 20 2006-01-03 → 2025-12-31 -14.4% -0.8% -46.7% 0.02
GLD · volatility targeting 19 2007-01-03 → 2025-12-31 +382.0% +8.6% -48.2% 0.62
GLD · the 200-day line 19 2007-01-03 → 2025-12-31 +251.0% +6.8% -34.6% 0.55
IWM · volatility targeting 20 2006-01-03 → 2025-12-31 +266.5% +6.7% -42.3% 0.43
IWM · the 200-day line 20 2006-01-03 → 2025-12-31 +88.5% +3.2% -34.2% 0.29
XLB · both together 20 2006-01-03 → 2025-12-31 -3.8% -0.2% -38.0% 0.06
XLB · the 200-day line, second run 20 2006-01-03 → 2025-12-31 -14.4% -0.8% -46.7% 0.02
GLD · both together 19 2007-01-03 → 2025-12-31 +270.1% +7.1% -32.4% 0.62
GLD · the 200-day line, second run 19 2007-01-03 → 2025-12-31 +251.0% +6.8% -34.6% 0.55
IWM · both together 20 2006-01-03 → 2025-12-31 +89.9% +3.3% -32.2% 0.30
IWM · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +88.5% +3.2% -34.2% 0.29
XLE · volatility targeting 20 2006-01-03 → 2025-12-31 +64.9% +2.5% -70.0% 0.23
XLE · the 200-day line 20 2006-01-03 → 2025-12-31 +37.2% +1.6% -45.5% 0.18
XLV · volatility targeting 20 2006-01-03 → 2025-12-31 +336.1% +7.6% -30.3% 0.61
XLV · the 200-day line 20 2006-01-03 → 2025-12-31 +78.0% +2.9% -31.8% 0.32
XLK · volatility targeting 20 2006-01-03 → 2025-12-31 +1137.5% +13.4% -38.6% 0.80
XLK · the 200-day line 20 2006-01-03 → 2025-12-31 +745.2% +11.3% -25.6% 0.77
XLE · both together 20 2006-01-03 → 2025-12-31 +33.0% +1.4% -46.2% 0.17
XLE · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +37.2% +1.6% -45.5% 0.18
XLF · volatility targeting 20 2006-01-03 → 2025-12-31 +215.6% +5.9% -58.6% 0.39
XLF · the 200-day line 20 2006-01-03 → 2025-12-31 +182.2% +5.3% -26.5% 0.45
XLV · both together 20 2006-01-03 → 2025-12-31 +96.9% +3.4% -23.2% 0.38
XLV · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +78.0% +2.9% -31.8% 0.32
XLK · both together 20 2006-01-03 → 2025-12-31 +758.8% +11.4% -20.1% 0.82
XLK · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +745.2% +11.3% -25.6% 0.77
XLF · both together 20 2006-01-03 → 2025-12-31 +163.4% +5.0% -27.9% 0.43
XLF · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +182.2% +5.3% -26.5% 0.45
XLY · volatility targeting 20 2006-01-03 → 2025-12-31 +653.8% +10.6% -43.4% 0.69
XLY · the 200-day line 20 2006-01-03 → 2025-12-31 +375.3% +8.1% -25.4% 0.64
XLP · volatility targeting 20 2006-01-03 → 2025-12-31 +237.7% +6.3% -21.2% 0.59
XLP · the 200-day line 20 2006-01-03 → 2025-12-31 +6.3% +0.3% -31.1% 0.08
XLI · volatility targeting 20 2006-01-03 → 2025-12-31 +431.5% +8.7% -44.5% 0.60
XLI · the 200-day line 20 2006-01-03 → 2025-12-31 +224.9% +6.1% -20.5% 0.52
XLY · both together 20 2006-01-03 → 2025-12-31 +381.5% +8.2% -22.9% 0.66
XLY · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +375.3% +8.1% -25.4% 0.64
XLP · both together 20 2006-01-03 → 2025-12-31 +22.1% +1.0% -27.5% 0.16
XLP · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +6.3% +0.3% -31.1% 0.08
XLI · both together 20 2006-01-03 → 2025-12-31 +224.0% +6.1% -20.1% 0.54
XLI · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +224.9% +6.1% -20.5% 0.52
XLU · volatility targeting 20 2006-01-03 → 2025-12-31 +213.2% +5.9% -38.8% 0.46
XLU · the 200-day line 20 2006-01-03 → 2025-12-31 +6.6% +0.3% -33.7% 0.09
XLU · both together 20 2006-01-03 → 2025-12-31 +20.3% +0.9% -28.5% 0.14
XLU · the 200-day line, second run 20 2006-01-03 → 2025-12-31 +6.6% +0.3% -33.7% 0.09
platform reference (SPY) (benchmark) 2000-01-03 → 2025-12-31 +322.4% +5.7% -58.0%

Growth and CAGR above are each walk over its own windows, so they are not comparable across walks with different window counts: a walk that excluded a window did not live through it. The figure rebases every line on the session all of them share.

2.3  Search accounting

This paper's search is a declared family: the paper's 28 walks, counted at N = 28 evaluated books. Every member is either a registered walk with its own hypothesis and frozen record, or a derived average computed from those frozen records; every member is reported, in the family table, 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 rVolatility targeting − rThe 200-day line 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, registered per window before scoring, is not multiplied by it.

In the table: Arm A = Volatility targeting · Arm B = The 200-day line.

Table 6. Window-by-window paired comparison. Δ is the growth gap (Arm A − Arm B) over the window's paired dates.
#WindowPaired bars Arm AArm B ΔLeader
1 2000-01-04 → 2000-12-29 251 -9.8% -20.8% +11.0 pp Arm A
2 2001-01-03 → 2001-12-31 247 -16.3% +0.0% -16.3 pp Arm B
3 2002-01-03 → 2002-12-31 251 -24.9% -4.4% -20.5 pp Arm B
4 2003-01-03 → 2003-12-31 251 +21.7% +23.2% -1.5 pp Arm B
5 2004-01-05 → 2004-12-31 251 +8.7% +1.6% +7.1 pp Arm A
6 2005-01-04 → 2005-12-30 251 +3.5% -1.2% +4.7 pp Arm A
7 2006-01-04 → 2006-12-29 250 +11.8% +7.2% +4.6 pp Arm A
8 2007-01-04 → 2007-12-31 250 +2.7% -5.4% +8.0 pp Arm A
9 2008-01-03 → 2008-12-31 252 -19.4% -0.9% -18.6 pp Arm B
10 2009-01-05 → 2009-12-31 251 +18.7% +13.7% +5.0 pp Arm A
11 2010-01-05 → 2010-12-31 251 +12.2% -1.0% +13.2 pp Arm A
12 2011-01-04 → 2011-12-30 251 -3.7% -11.0% +7.3 pp Arm A
13 2012-01-04 → 2012-12-31 249 +11.7% +9.9% +1.8 pp Arm A
14 2013-01-03 → 2013-12-31 251 +26.4% +26.4% +0.0 pp tie
15 2014-01-03 → 2014-12-31 251 +11.6% +10.3% +1.4 pp Arm A
16 2015-01-05 → 2015-12-31 251 -1.9% -9.1% +7.1 pp Arm A
17 2016-01-05 → 2016-12-30 251 +10.0% +8.2% +1.8 pp Arm A
18 2017-01-04 → 2017-12-29 250 +18.5% +18.5% +0.0 pp tie
19 2018-01-03 → 2018-12-31 250 -5.9% -10.2% +4.3 pp Arm A
20 2019-01-03 → 2019-12-31 251 +21.3% +13.5% +7.9 pp Arm A
21 2020-01-03 → 2020-12-31 252 +15.6% +3.0% +12.7 pp Arm A
22 2021-01-05 → 2021-12-31 251 +28.5% +28.8% -0.3 pp Arm B
23 2022-01-04 → 2022-12-30 250 -18.6% -16.1% -2.5 pp Arm B
24 2023-01-04 → 2023-12-29 249 +24.8% +9.4% +15.4 pp Arm A
25 2024-01-03 → 2024-12-31 251 +23.9% +24.0% -0.1 pp Arm B
26 2025-01-03 → 2025-12-31 249 +12.1% +11.0% +1.1 pp Arm A

Paired Sharpe of the difference track: 0.20 · block bootstrap (2000 paths, block 10, seed 1234): P(Volatility targeting beats The 200-day line) = 87.0%.

Window win-rate. Volatility targeting led 17 of 26 windows (65.4%), The 200-day line led 7, and 2 windows were ties, and the mean window gap of +2.09 pp points the same way. Widest single window: 2002 at -20.5 pp.

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

The sentence below is the registered design with its rules in plain words; the exact registration record is in Appendix A2; the authored description of the design is Section 1.

A COMPARATIVE study: The volatility target vs The 200-day line, walked on the same registered out-of-sample windows. The volatility target: SPY, bought when always, every trading day; sell when never, and forward-tested out-of-sample from the anchor: the rule set is frozen at the anchor, with no in-sample re-optimization. The 200-day line: SPY, bought when the close is above its 200-day average; sell when the close is below its 200-day average, and forward-tested out-of-sample from the anchor: the rule set is frozen at the anchor, with no in-sample re-optimization. The arms differ in 2 nodes across 6 recorded settings: Signal Module, config (entry · indicator: every_day → sma; entry · operator: above → price_above; entry · length:, → 200; entry · value: 0.5 →, ; …); Volatility Target, removed in The 200-day line. NOTE: with more than one difference, an out-of-sample gap cannot be attributed to any single change. The contrast under test: whether The volatility target generates better risk-adjusted returns than The 200-day line over the identical out-of-sample windows.

Every block in this study is a card from the platform's catalog: the fund and its price loader, the Signal Module (the 200-day average, or Every trading day for volatility targeting, which is always in the fund), the Signal Forward Test, one year at a time, the Transaction Cost card, and the new Volatility Target card, which sets how much of the fund to hold from the size of its recent and normal daily moves. A reader can rebuild every test and change any setting.

The frozen circuit, data flows left to rightticker: click for detailstickerticker price loader: click for detailsticker price loadersignal module: click for detailssignal modulebacktest validator: click for detailsbacktest validatortransaction cost: click for detailstransaction costvolatility target: click for detailsvolatility targetticker: click for detailstickerticker price loader: click for detailsticker price loadersignal module: click for detailssignal modulebacktest validator: click for detailsbacktest validatortransaction cost: click for detailstransaction costThe volatility targetThe 200-day lineshared
Figure 7. The frozen circuit, every node a primitive, every wire a typed data-flow; the two arms are colour-coded (Volatility targeting green, The 200-day line 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
Ticker, A single instrument symbol, the seed of a single-name circuit.
Ticker Price Loader, Single-ticker OHLCV loader, the deep window a signal needs.
Signal Module, The entry / exit rule, turn indicators into a per-bar trade signal.
Transaction Cost, Charge for trading, slippage + commission on every turn.
Backtest Validator, Forward-test the winning rule on unseen, out-of-sample data.
Volatility Target, Hold less when the fund gets jumpy, all of it when it is calm.

The objective and the search

The volatility target

UniverseSingle ticker SPY with 365 days (~1.0y) of price history.
Signal generationbuy when always, every trading day; sell when never.
Validation & out-of-samplesignal forward test (1y horizon from the anchor); overlays: Transaction Cost, Volatility Target.

The 200-day line

Signal generationbuy when the close is above its 200-day average; sell when the close is below its 200-day average.
Validation & out-of-samplesignal forward test (1y horizon from the anchor); overlays: Transaction Cost.

Every other specification row is identical to The volatility target's table above.

What differs between the arms, 2 differences; more than one thing changes at once:

  • paramSignal Module, config
    • · buy when: always, every trading day (Volatility targeting); the close is above its 200-day average (The 200-day line)
    • · sell when: never (Volatility targeting); the close is below its 200-day average (The 200-day line)
  • removedVolatility Target, removed in The 200-day line

Reader's note. With 2 settings changed at once across 2 nodes, an out-of-sample gap between the arms cannot be attributed to any single change, the arms are compared as whole packages, and any causal reading of one ingredient is unsupported by this design.

Cost elements are wired into the circuit.

Show the mathematics, 6 primitives, formulas and parity notes

3.1  Ticker

A single instrument symbol, the seed of a single-name circuit.

No math. It just names one stock and hands the symbol to a Ticker Price Loader, which fetches its price history.

What it is

A constant: one ticker string. The computation lives downstream in the loader and the signal.

3.2  Ticker Price Loader

Single-ticker OHLCV loader, the deep window a signal needs.

Same as the bulk loader but for one name, fetching a deep lifecycle window so an indicator or stochastic signal has enough history to warm up.

What it loads

One ticker's OHLCV up to the anchor, length set by the strategy's warm-up requirement.

P = \{(o,h,l,c,v)_\tau : \tau \le t\}

3.3  Signal Module

The entry / exit rule, turn indicators into a per-bar trade signal.

Composes indicators (RSI, moving averages, …) with comparison and logic operators into a rule that says enter, exit, or hold each bar. The rule is emitted as a portable config the optimizer tunes and the walk-forward validates, so what you design is exactly what gets traded.

Boolean rule → position state
\text{signal}_t = \begin{cases} +1 & \text{entry rule true} \\ 0 & \text{exit rule true} \\ \text{hold} & \text{otherwise}\end{cases}
e.g. enter when RSI < 30, exit when RSI > 50.

3.4  Backtest Validator

Forward-test the winning rule on unseen, out-of-sample data.

Takes the wired rule config (the Walk-Forward validated config wins, else the optimized config, else the raw signal config) and trades it FORWARD on the out-of-sample window to the right of the anchor, data it never saw during optimization, re-deriving the regime as-of each bar. It produces the true out-of-sample equity curve, trades and statistics: the signal-path twin of the Portfolio Forward Test, not an in-sample replay.

Apply the frozen rule forward (OOS)
E_t = E_{t-1}\,(1 + r_t),\qquad \text{Sharpe} = \frac{\bar r - r_f}{\sigma_r}\sqrt{252}
Config frozen from optimization / walk-forward, then replayed bar-by-bar on the forward window it has never seen, with cost + risk overlays applied.

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

3.6  Volatility Target

Hold less when the fund gets jumpy, all of it when it is calm.

Each close it compares how much the fund moved over the recent days with how much it normally moves, and holds normal divided by recent of the position over the next session. Twice as jumpy as normal holds half; it never holds more than the cap, so it never borrows. Every resize pays the transaction cost. It reads the fund's own price, not the strategy's results.

Share of the position held
w_{t+1} = \min\!\left(\frac{\sigma_{\text{normal},t}}{\sigma_{\text{recent},t}},\; c\right)
Each sigma is the standard deviation of daily returns up to the close t: recent over the last days, normal over the past years; c = the cap.

4  Discussion

4.1  Interpretation

The six falls came in two kinds, and in the clearest cases each rule protected better against one kind. In a long bear market the price falls for years and stays far below its average. The 200-day line suits that case: it waits for the price to cross its average, and in a long fall there is time to wait. In a sudden crash like the Covid crash the daily moves jump first and the price follows within days. Volatility targeting suits that case. It reads the daily moves, so it already held less while SPY was still near its high, and it bought back step by step as the moves calmed down.

Each rule also has a weakness. The 200-day line is late in a sudden crash. It sells only after the price has already fallen to its average, it can sell and buy back several times while the price jumps around that average, and it buys back for good only after most of the rebound. In a market that moves sideways near its average, or falls slowly with rallies above it as in the 2022 slide, it sells and buys back again and again, losing a little each time. Volatility targeting protects only when the daily moves jump. In the dot-com crash they stayed close to their normal size most of the time, and it did not help: it took longer than any other way of holding to get back where it started. In the autumn of 2008 the moves jumped to several times their normal size, and it made that fall shallower. Big up days count as moves too, so after the lows of the 2018 drop and the 2025 tariff drop it held less in the first weeks of the rebound, and holding was back first.

For managing risk this gives a plain choice. If the fear is a long bear market that takes years to recover from, the 200-day line protects: in both long bear markets it was back where it started years before holding. It gives up growth to do it, in the rebounds it misses and in the sideways years. If the fear is a sudden crash that makes people sell at the bottom, volatility targeting makes it shallower, and on SPY it gave up no growth at all. Both together had the smallest fall in five of the six falls, and gave up most of the growth the line gave up.

A smaller fall also matters for a reason the tables do not show: it is easier to hold through, and an investor who sells at the bottom takes the whole loss.

Where this study stops. Each rule ran with one setting, chosen before the tests and never tuned. Other settings may protect differently, and anyone can try them in the lab. Dividends and interest on cash are not counted: holding would gain more from dividends, and the line, out of the market more than a quarter of the time, would have earned interest in the years when interest rates were high. The next questions are the same test with dividends and interest counted, and the same test on markets outside the US.

4.2  Limitations

Prices without dividends. Every rule and holding itself use prices without dividends. With dividends, holding would grow faster, and the rules that are partly in cash would miss part of the dividends.

Cash earns nothing. The 200-day line and both together spend long stretches in cash. Interest on that cash would have helped them in the years when interest rates were high.

One setting each, fixed in advance: the last 21 trading days of daily moves against the past five years, never more than the whole position, and the 200-day average. Nothing was tuned on the results, so other settings may show more or less protection.

No borrowing. The volatility-managed portfolios of Moreira and Muir (2017) hold more than the whole position in calm times, with borrowed money. Ours never does, so it can only hold less.

Six falls. Six large falls in a quarter century is a small number to judge rules by. The 13 other funds add breadth, but they lived through the same crashes.

Young funds. QQQ, long Treasuries and gold were only a few years old when their tests began, so their normal size of daily moves came from a short stretch. For QQQ that stretch was the most turbulent years of the dot-com era, so in the crash that followed, volatility targeting held almost all of QQQ while it kept falling.

Trading. Volatility targeting resizes its holding a little on every day it holds less than the whole position. At the small cost per trade used here, these resizes took very little growth on SPY; a trader who pays five times as much per trade would give up about 0.2% of growth a year. Taxes are not counted; in a taxable account every sale can mean tax to pay, and both rules sell more often than holding does.

Counting every day. The platform runs one test per year, and each yearly test skips the first trading day of its year. This paper fills those days in, with what each rule held the day before, so the lines run without gaps. That is why every way of holding grows a little faster here than in our Kalman paper, and why a year rebuilt in the lab can differ a little from this paper's figures.

Enter the lab

Press Open the platform and this study's test opens in the lab, our research workspace: SPY and its prices, volatility targeting and the 200-day line side by side, the one-year test and the trading cost. Volatility targeting is new in the lab, and its panel shows how much of the fund it holds when the daily moves grow.

Things to try. Run it on another fund. Let volatility targeting look at the last 10 trading days instead of 21, or at three years of normal moves instead of five, or let it hold at most 80% of the fund. Or put it under another rule of your own, or under the 200-day line itself, which is the both-together test of this paper. The lab runs one year at a time, so a year rebuilt there can differ a little from this paper's figures (see Counting every day, under Limitations).

References

Author’s references?
  1. Moreira and Muir (2017), Volatility-Managed Portfolios, Journal of Finance 72(4): 1611-1644. Holding less of the market after its moves have been large raised the return for each unit of risk. https://doi.org/10.1111/jofi.12513
  2. Fleming, Kirby and Ostdiek (2001), The Economic Value of Volatility Timing, Journal of Finance 56(1): 329-352. https://doi.org/10.1111/0022-1082.00327
  3. Harvey, Hoyle, Korgaonkar, Rattray, Sargaison and Van Hemert (2018), The Impact of Volatility Targeting, Journal of Portfolio Management 45(1): 14-33. For stocks, volatility targeting made the worst falls smaller and raised the return for each unit of risk. https://doi.org/10.3905/jpm.2018.45.1.014
  4. Cederburg, O'Doherty, Wang and Yan (2020), On the Performance of Volatility-Managed Portfolios, Journal of Financial Economics 138(1): 95-117. Tested in real time, most volatility-managed portfolios did not beat holding. https://doi.org/10.1016/j.jfineco.2020.04.015
  5. Bongaerts, Kang and van Dijk (2020), Conditional Volatility Targeting, Financial Analysts Journal 76(4): 54-71. Cutting only when the moves are unusually large, and never adding, kept most of the benefit. https://doi.org/10.1080/0015198X.2020.1790853
  6. Faber (2007), A Quantitative Approach to Tactical Asset Allocation, Journal of Wealth Management 9(4): 69-79. A ten-month moving average rule, the monthly cousin of the 200-day line. https://doi.org/10.3905/jwm.2007.674809
  7. QuanterLab (2026), A Stochastic Model Against a Geometric Rule: the Kalman Filter and the 200-Day Line, the same funds and years. https://quanterlab.com/research/a-stochastic-model-against-a-geometric-rule-the-kalman-filter-and-the-200-day

Appendix A  Reproducibility in QuanterLab

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

#CommitReportAnchorOOS window
1 7c11bec189b5 13821 2000-01-01 2000-01-03 → 2000-12-29
2 4c22e2479a1c 13824 2001-01-01 2001-01-02 → 2001-12-31
3 43efeb312dc0 13827 2002-01-01 2002-01-02 → 2002-12-31
4 a1ed70f4b60d 13830 2003-01-01 2003-01-02 → 2003-12-31
5 61daf3244673 13833 2004-01-01 2004-01-02 → 2004-12-31
6 d3e39d41d8b7 13836 2005-01-01 2005-01-03 → 2005-12-30
7 dd9489cadc65 13839 2006-01-01 2006-01-03 → 2006-12-29
8 8f886149ee68 13842 2007-01-01 2007-01-03 → 2007-12-31
9 dd10dfc6bf21 13845 2008-01-01 2008-01-02 → 2008-12-31
10 b5616fa8a381 13848 2009-01-01 2009-01-02 → 2009-12-31
11 c7d73fda6e9b 13851 2010-01-01 2010-01-04 → 2010-12-31
12 3350d167e0c4 13854 2011-01-01 2011-01-03 → 2011-12-30
13 7fb0c499efa0 13857 2012-01-01 2012-01-03 → 2012-12-31
14 705efbaf263b 13860 2013-01-01 2013-01-02 → 2013-12-31
15 c2a85a6ca879 13863 2014-01-01 2014-01-02 → 2014-12-31
16 6bf57c8b6834 13866 2015-01-01 2015-01-02 → 2015-12-31
17 8f55216a014e 13869 2016-01-01 2016-01-04 → 2016-12-30
18 91f87ed1c17f 13872 2017-01-01 2017-01-03 → 2017-12-29
19 23dfc30c4edb 13875 2018-01-01 2018-01-02 → 2018-12-31
20 5dbf8a455826 13878 2019-01-01 2019-01-02 → 2019-12-31
21 afea05508958 13881 2020-01-01 2020-01-02 → 2020-12-31
22 57659defea51 13883 2021-01-01 2021-01-04 → 2021-12-31
23 a409b7e4fe89 13885 2022-01-01 2022-01-03 → 2022-12-30
24 760fc0ece91d 13887 2023-01-01 2023-01-03 → 2023-12-29
25 e236041a6ec8 13889 2024-01-01 2024-01-02 → 2024-12-31
26 1b14e88b3517 13891 2025-01-01 2025-01-02 → 2025-12-31
QuanterLab reference architecture

The methods the platform implements, for every study made on it, rest on these works.

  1. 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.
  2. 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
  3. Lo, A. W. (2002). The Statistics of Sharpe Ratios. Financial Analysts Journal, 58(4), 36–52. doi:10.2469/faj.v58.n4.2453

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 registration 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 registration precedes its own run, and each run precedes the next registration. A study whose registrations all post-date its runs would show it here. Wall-clock spacing between registrations varies with the author's schedule and queue latency; the ordering, not the tempo, is the claim.

“A COMPARATIVE study: The volatility target vs The 200-day line, walked on the same registered out-of-sample windows. The volatility target: SPY, bought when always, every trading day; sell when never, and forward-tested out-of-sample from the anchor: the rule set is frozen at the anchor, with no in-sample re-optimization. The 200-day line: SPY, bought when the close is above its 200-day average; sell when the close is below its 200-day average, and forward-tested out-of-sample from the anchor: the rule set is frozen at the anchor, with no in-sample re-optimization. The arms differ in 2 nodes across 6 recorded settings: Signal Module, config (entry · indicator: every_day → sma; entry · operator: above → price_above; entry · length:, → 200; entry · value: 0.5 →, ; …); Volatility Target, removed in The 200-day line. NOTE: with more than one difference, an out-of-sample gap cannot be attributed to any single change. The contrast under test: whether The volatility target generates better risk-adjusted returns than The 200-day line 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 registration precedes its own run, and each run precedes the next registration.
#AnchorRegistered at (UTC)Run completed (UTC)
1 2000-01-012026-10-05 09:02:53 2026-10-05 09:02:54
2 2001-01-012026-10-05 09:02:54 2026-10-05 09:02:55
3 2002-01-012026-10-05 09:02:55 2026-10-05 09:02:56
4 2003-01-012026-10-05 09:02:57 2026-10-05 09:02:58
5 2004-01-012026-10-05 09:02:58 2026-10-05 09:02:59
6 2005-01-012026-10-05 09:02:59 2026-10-05 09:03:00
7 2006-01-012026-10-05 09:03:00 2026-10-05 09:03:01
8 2007-01-012026-10-05 09:03:01 2026-10-05 09:03:02
9 2008-01-012026-10-05 09:03:02 2026-10-05 09:03:03
10 2009-01-012026-10-05 09:03:04 2026-10-05 09:03:05
11 2010-01-012026-10-05 09:03:05 2026-10-05 09:03:06
12 2011-01-012026-10-05 09:03:06 2026-10-05 09:03:07
13 2012-01-012026-10-05 09:03:07 2026-10-05 09:03:08
14 2013-01-012026-10-05 09:03:09 2026-10-05 09:03:10
15 2014-01-012026-10-05 09:03:10 2026-10-05 09:03:11
16 2015-01-012026-10-05 09:03:11 2026-10-05 09:03:12
17 2016-01-012026-10-05 09:03:12 2026-10-05 09:03:13
18 2017-01-012026-10-05 09:03:13 2026-10-05 09:03:15
19 2018-01-012026-10-05 09:03:15 2026-10-05 09:03:16
20 2019-01-012026-10-05 09:03:16 2026-10-05 09:03:17
21 2020-01-012026-10-05 09:03:17 2026-10-05 09:03:18
22 2021-01-012026-10-05 09:03:18 2026-10-05 09:03:19
23 2022-01-012026-10-05 09:03:19 2026-10-05 09:03:20
24 2023-01-012026-10-05 09:03:20 2026-10-05 09:03:21
25 2024-01-012026-10-05 09:03:22 2026-10-05 09:03:23
26 2025-01-012026-10-05 09:03:23 2026-10-05 09:03:24
QuanterLab · Study deb2c4f47878 · compiled October 05, 2026. The universe is a fixed list of funds, so there are no constituents to resolve; 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.

Run a study like this one

Everything above was produced inside QuanterLab, the registration, the walk, the statistics and the paper itself. Build the circuit on a canvas, register the hypothesis before you score it, and the platform enforces the rest.

Reading the research needs no account; creating one is free. Follow the research and we'll tell you when the next study publishes.

As seen on Quantocracy

A note on AI. QuanterLab is a quantitative finance research platform, and every number in this study comes from a run on the platform. The hypothesis, the parameter choices, the validation design and the conclusions belong to the author. Runs execute on point-in-time data with walk-forward validation, and each study ships with its methodology and logs, so a reader can reconstruct the result instead of trusting it. I use AI to edit and structure the prose; it does not generate results, produce numbers, or decide what a study concludes.