Results

Every strategy we have tested: the scoreboard

One row per strategy, best first. The failures stay on the board — a list of winners only would be a sales page, not research.

39 strategies and strategy families from 15 articles · updated 2026-10-16 · download as CSV or JSON

How to read this: sort by out-of-sample Sharpe — return per unit of risk on years the strategy never saw — to see what held up. Every number is net of costs and comes straight from the article's own results file; hover over one to see which. The verdict is the article's own: Works Weak Failed. Column by column ↓

Show:

								Costs
Turnaround Tuesday	S&P 500 (SPY)	Daily	0.61	1.23	—	74.1%	-23.8%	—	Works
Turnaround Tuesday	Nasdaq-100 (QQQ)	Daily	0.89	1.17	—	75.7%	-19.5%	—	Works
200-day moving average, hold or cashBuy-and-hold: Sharpe 1.08, max drawdown -83.4%.	Bitcoin	Daily	—	—	1.16	—	-70.0%	—	Works
Two-rule mean reversion (worked example)	Nasdaq-100 (QQQ)	Daily	0.79	1.04	—	—	-20.3%	—	Works
RSI(2) dip-buying, exit above the prior high	Nasdaq-100 (QQQ)	Daily	0.52	0.92	0.73	70.4%	-18.0%	—	Works
RSI(2) dip-buyer (tripwire worked example)	Nasdaq-100 (QQQ)	Daily	0.88	0.86	0.87	69.0%	-20.5%	—	Works
Opening-move breakout, long-only with trend filterWeaker in the last two years of the test.	Nasdaq-100 futures (NQ)	5-min opening range, intraday	0.77	0.84	0.80	12.4%	-4.6%	—	Works
RSI(2) below 30, exit above the prior highIn the market 37% of the time.	Nasdaq-100 (QQQ)	Daily	—	—	0.82	—	-19.5%	—	Works
Connors TPS scale-in	Nasdaq-100 futures (NQ)	Daily	0.72	0.78	—	—	-11.6%	—	Works
200-day moving average, hold or cashBuy-and-hold: Sharpe 0.64, max drawdown -55.2%.	S&P 500 (SPY)	Daily	—	—	0.72	—	-29.4%	—	Works
RSI(3) dip-buying with trend filter	Nasdaq-100 futures (NQ)	4-hour	0.61	0.72	—	—	-13.6%	—	Works
Intraday momentum (Zarattini), long side	Nasdaq-100 futures (NQ)	Intraday	0.90	0.68	0.76	39.4%	-11.0%	—	Works
Buy after three down closes, sell the next close	Nasdaq-100 (QQQ)	Daily	0.73	0.67	0.67	56.7%	-18.7%	45 bp per trade before costs, 35 bp after	Works
RSI(3) dip-buying with trend filter	Nasdaq-100 futures (NQ)	2-hour	0.37	0.63	—	—	-14.1%	—	Works
Connors RSI(2) with 200-day filter	Nasdaq-100 futures (NQ)	Daily	0.68	0.56	—	—	-13.4%	—	Works
200-day moving average, hold or cashBuy-and-hold: Sharpe 0.46, max drawdown -83.0%.	Nasdaq-100 (QQQ)	Daily	—	—	0.53	—	-58.9%	—	Works
RSI 30/40155 trades. The article reports win rate and drawdown for this version, not Sharpe.	Nasdaq-100 (QQQ; NQ futures as the history proxy)	2-hour	—	—	—	70.3%	-11.5%	—	Works
Intraday momentum (Zarattini), both sidesThe headline number rides the short side's out-of-sample luck.	Nasdaq-100 futures (NQ)	Intraday	0.43	1.39	0.99	38.2%	-12.3%	—	Weak
Moving-average crossover, parameters picked in-sampleBuy-and-hold over the out-of-sample years: Sharpe 0.93.	Nasdaq-100 (QQQ)	Daily	0.77	0.77	—	—	-28.6%	—	Weak
Momentum rotation: hold the six-month winnerEqual-weight buy-and-hold of the same three: Sharpe 0.96.	Stocks, bonds, gold (SPY / TLT / GLD)	Monthly	0.52	0.70	0.61	—	-33.6%	—	Weak
RSI 30/40Only 32 trades in the whole test: precise, but too rare to lean on alone.	Nasdaq-100 (QQQ)	Daily	0.10	0.62	0.37	87.5%	-26.3%	—	Weak
RSI(14) below 30, hold five daysIn the market 3.3% of the time.	Nasdaq-100 (QQQ)	Daily	—	—	0.45	—	-21.2%	—	Weak
Fridays when VIX is above 20	Gold (GLD)	Daily, Thursday close to Friday close	—	—	0.23	51.3%	-22.0%	Sharpe 0.23 at 0.1% per round trip, 0.43 at futures-level costs	Weak
Intraday momentum (Zarattini), short sideNegative in-sample; profitable only out-of-sample.	Nasdaq-100 futures (NQ)	Intraday	-0.21	1.22	0.64	37.1%	-16.5%	—	Failed
200-day moving average, hold or cashBuy-and-hold: Sharpe 0.67, max drawdown -45.6%.	Gold (GLD)	Daily	—	—	0.58	—	-35.1%	—	Failed
Sell in May, bonds for the summerBuy-and-hold over the same years: Sharpe 0.66.	S&P 500 (SPY) and bonds	Seasonal, two switches a year	—	—	0.57	—	-39.6%	—	Failed
Sell in May, cash for the summerBuy-and-hold: Sharpe 0.64.	S&P 500 (SPY)	Seasonal, two switches a year	—	—	0.55	—	-36.7%	—	Failed
Deliberately overfit: best of 2160 parameter combinations	Nasdaq-100 (QQQ)	Daily	0.82	0.33	—	—	-24.6%	—	Failed
Naive opening range breakout, both directions	Nasdaq-100 futures (NQ)	5-min opening range, intraday	-0.27	0.32	0.11	31.0%	-43.1%	—	Failed
Every Friday (Thursday close to Friday close)The average Friday edge is about one round-trip commission.	Gold (GLD)	Daily	—	—	0.05	50.8%	-40.3%	—	Failed
200-day moving average, hold or cashBuy-and-hold: Sharpe 0.29, max drawdown -48.4%.	Long bonds (TLT)	Daily	—	—	0.03	—	-47.3%	—	Failed
Fridays above the 200-day average	Gold (GLD)	Daily	—	—	-0.02	49.7%	-32.5%	—	Failed
Santa Claus rally	S&P 500 (SPY)	Seasonal	0.31	-0.07	—	—	—	—	Failed
Hold every night (buy the close, sell the open)The overnight effect is real; trading it every night is not.	Nasdaq-100 (QQQ)	Daily, overnight hold	—	—	-0.79	—	-98.3%	Breaks even only below 2.8 bp per side	Failed
Classic book setups (Street Smarts, DeMark, Kaufman, Chande)17 variants tested, 0 passed the bar.	Futures	Various	—	—	—	—	—	—	Failed
Published catalog setups (DAX drift, EuroStoxx, natural gas, soybeans)4 tested, 0 passed: all negative out-of-sample.	European index and commodity futures	Daily and seasonal	—	—	—	—	—	—	Failed
Trend following, Carver-style acceleration20 variants tested, 0 passed.	Futures panel	Daily	—	—	—	—	—	—	Failed
Trend following, Clenow-style core panel24 variants tested, 0 passed: real, but needs far more markets than a retail account holds.	Futures panel	Daily	—	—	—	—	—	—	Failed
Turtle Soup (Street Smarts)4 markets tested, 0 passed; walk-forward mean Sharpe between -0.04 and -0.4.	S&P 500, Nasdaq, gold and bond futures	Daily	—	—	—	—	—	—	Failed

Column by column

Sharpe in-sample and out-of-sample. Return per unit of risk on the years a strategy was built on, and on the years it never saw. The second number is the one that matters: a strategy that collapses out-of-sample was fitted to the past. Each Sharpe ratio is annualised at that strategy's own trading frequency.
Sharpe full test. For articles that test the whole history in one piece, without a held-out period. Read those rows with more caution.
Win rate and max drawdown cover the full test where the article reports one. Where an article reports only separate in-sample and out-of-sample figures, the table shows the worse of the two. Hover over any number to see exactly which results file and field it comes from.
Costs. Every result is already net of costs — 0.05% per side on ETFs, 0.01% per round trip on index futures. The column shows how sensitive a result is to costs, where the article measured it.
Verdict is what the article itself concluded. Works held up on data it never saw, net of costs. Weak is real but small, conditional, or beaten by simply holding the market. Failed lost money, collapsed out-of-sample, or did not survive costs.
Graveyard rows are whole families screened in our research log against a stricter bar, so they show how many variants were tried rather than one Sharpe ratio. The strategy graveyard explains the bar.

Two honest limits. Sharpe ratios from different markets, periods and trade frequencies are not perfectly comparable, so use the table to find what is worth reading, then read the article. And every row is a backtest: hypothetical, however carefully it was run.

Nothing on this page is typed by hand. A script reads each article's own results file and rebuilds the table whenever a new test is published, so a number here always matches the article it links to. The rules behind every test are on how we test.

By topic: Nasdaq · Gold · ICT concepts · Classic books · All of them: the scoreboard

Download a free strategy

Simple and robust — 70% winners across 155 trades over 16 years, net of costs. The free PDF gives you the market, the timeframe and the exact rules in plain English, plus every single trade it has taken.

Instant delivery. One PDF, no spam. Unsubscribe anytime.

Backtest notice: every result linked here is a backtest on historical data, net of trading costs (0.05% per side on ETFs, 0.01% per round trip on index futures). Backtested performance is hypothetical and does not represent live trading. Past performance does not guarantee future results. This is research, not financial advice.