Weak-day reversal (lower high, lower low)
Buy the break of a weak day's high. Small, positive, and not distinguishable from noise.
Why it failed: Best of six pre-set variants: t 1.67. Long-only on rising indices, so even that is mostly drift. See the scorecard
Cumulative result
R per trade (bid/ask fills, financing). The best of the three exits we fixed in advance; the other two are weaker.
Show the data as a table
| Year | Trades | Sum R | Mean R |
|---|---|---|---|
| 2005 | 32 | −4.6 | −0.144 |
| 2006 | 30 | −3.9 | −0.130 |
| 2007 | 33 | +10.8 | +0.328 |
| 2008 | 34 | −4.0 | −0.117 |
| 2009 | 35 | +3.7 | +0.107 |
| 2010 | 35 | +4.6 | +0.133 |
| 2011 | 33 | +2.0 | +0.060 |
| 2012 | 33 | +1.1 | +0.032 |
| 2013 | 30 | −2.9 | −0.097 |
| 2014 | 30 | +2.1 | +0.069 |
| 2015 | 31 | +1.7 | +0.055 |
| 2016 | 28 | +0.6 | +0.023 |
| 2017 | 28 | +7.1 | +0.252 |
| 2018 | 29 | +11.6 | +0.399 |
| 2019 | 33 | +8.1 | +0.245 |
| 2020 | 28 | +3.7 | +0.133 |
The claim
After a day with a lower high and a lower low than the day before, buy when price breaks back above that weak day's high; stop at its low.
Source: Shared on X as 'Strategy #24' (AlgomaticTrade, October 2026). The exits are behind a paywall, so this is not the author's full strategy: we fixed three exits ourselves before the run.
What we tested
- Day D-1 has a lower high and a lower low than day D-2 (sessions run 18:00 to 17:00 New York).
- On day D, the first 1-hour close above the high of D-1 triggers a long entry at the ask on the next bar's open.
- Stop at the low of D-1; 1R is the distance from entry to stop.
- Exits fixed in advance: X1 the first daily close above the entry (at most 5 sessions); X2 the close of the third session; X3 a +1R target (at most 5 sessions).
Costs and fills
- Bid/ask fills from the minute file, stop first when stop and target fall in the same bar, gaps through the stop filled at the bar open, financing of 2.1% a year on the notional for each night held.
Data
- US100 and US500 CFD minute bid/ask, true clock, aggregated to 1-hour bars
Six variants, none near a bar
Three exits on two indices gave six results, all printed. The best, holding for three sessions on the Nasdaq-100, made +0.083R a trade over 502 trades, t 1.67. The same exit on the S&P 500 made +0.130R on 200 trades, t 1.57. Both halves of the Nasdaq sample are positive.
None of them is close to the 3.43 bar for a single idea, let alone the higher bar for six tries.
Why a small positive number is not an edge here
The rule only buys, and it holds for several days on indices that rose strongly over the sample (the Nasdaq-100 roughly eightfold from 2005 to 2020). A long-only rule should beat the market's own upward drift over the same holding period before it is called a setup. We did not run that benchmark, so even the small positive mean is not evidence of anything specific to the setup.
The 12-gate scorecard
3 passed · 1 failed · 8 not run or not reached. A gate is "not reached" when the strategy had already failed a veto: there is nothing left for it to test.
Is the backtest real?
Absolute veto. These gates look for bugs and physical limits, not for statistical evidence, so one failure ends the test.
True clock from the lab's loader; session days defined in New York time.
Bid/ask fills, gap-through stops at the open, stop first on a same-bar tie, overnight financing included.
Is the edge more than luck?
G12, one pre-registered holdout, is the only statistical veto. G1 to G5 and the regime test G6 are always reported; G6 vetoes only a regime that is significantly negative. A strategy with no edge after costs never reaches G12, because there is nothing to confirm.
Best of six variants: t 1.67, against a bar of at least 3.43.
No drift benchmark (random-day longs with the same holding period). The positive mean may be index drift.
All three exits positive on Nasdaq (t 1.26 to 1.67); none strong.
Both halves of the Nasdaq sample positive for every exit.
Validation and holdout years (2021 onwards) were not opened.
Can it survive a prop-firm account?
Absolute veto, for strategies meant for funded accounts. The account rules are simulated event by event, not at the end of the day.
Long-only on an index needs a drift benchmark
Almost any buy rule on the Nasdaq-100 since 2005 shows a profit. The question is whether it beats buying on random days with the same stop and holding period. Without that comparison, a positive backtest says more about the index than about the setup.
Receipts
Every number on this page comes from one of these files: the test script, its saved output, the trade list and, where the test was pre-registered, the registration and its hash lock. Download them and check our work. Numbers that also appear in the chart are recomputed from the trade list when this page is built, and the build fails if they disagree.
| File | Kind | Size | Status | sha256 |
|---|---|---|---|---|
| _algomatic24_lhll_20261019.py experiments/_algomatic24_lhll_20261019.py | Script | 6 KB | Complete | 4610463e6c1e |
| _algomatic24_lhll_20261019.out experiments/_algomatic24_lhll_20261019.out | Saved output | 903 B | Complete | 8f9ac009ccf7 |
| _algomatic24_lhll_20261019_trades.csv experiments/_algomatic24_lhll_20261019_trades.csv | Trade list | 163 KB | Complete | 9d6163ae3e57 |
| data_loader.py prop_gpu_miner/data_loader.py | Lab module | 10 KB | 3 lines withheld (private strategy) | 62ae717b4bf1 |
Price data is not included: the vendors' licences do not allow it, so the scripts run only with your own copy of the data, and the trade lists show dates, direction, R and costs, not prices. Lines that refer to the lab's private live strategy are replaced by a marker; the receipts page explains the rules and lists the sha256 of every original file.
Related tests
Intraday momentum into the close
The effect from the papers is there before costs: about 1.4 basis points a day. The spread is 0.6.
We mined 10,248 strategy rules on a GPU
Breakouts, sweeps, VWAP reversion and engulfing bars on five indices. The best result was weaker than what noise produces.