We test what futures traders are told, and publish what fails.
Futures Playbook Research is an independent research firm for futures markets. Each study runs on CME futures data back to 2010, net of costs, and is published with its files: the pre-registration where one exists, the result and the code. Most claims fail, and the library says so.
Do the setups sold for passing prop evaluations beat costs?
- 15-minute opening-range breakout, 2× range targetInconclusiveInconclusive
- Opening-range breakout, let it runRefutedRefuted
- Initial-balance extensionRefutedRefuted
- Initial-balance failed-extension fadeRefutedRefuted
- Range fade after 10:00RefutedRefuted
- ICT Silver Bullet, New York 10:00–11:00 (sold as 77% at 1:2)InconclusiveRefuted
- ICT Silver Bullet, 3R targetInconclusiveRefuted
- ICT Silver Bullet, London 03:00–04:00RefutedRefuted
- VWAP ±2σ band fadeRefutedRefuted
- VWAP trend-day pullbackToo fewToo few
- EMA 9/21 crossoverRefutedRefuted
- News straddle (CPI, payrolls, FOMC)Too fewToo few
- Post-release spike fadeToo fewToo few
Latest studies
15 studies and 950 tested arms so far, numbered in the order they were written down.
- FPR-26-15The setups sold to prop tradersRefuted
Thirteen setups sold to prop traders, on NQ and ES micros: none of 26 market cells beats costs.
- FPR-26-14Day session, gaps, turn of month and release daysRefuted
Twenty-nine everyday claims, from being long the day session to gap follow-through and turn-of-month strength: none validated.
- FPR-26-13The overnight drift in three layers, including crowd attentionMixed
One of 19 arms passes, the pre-FOMC night, but it re-measures nights already known. Crowd attention as a signal is refuted.
- FPR-26-12Is the overnight drift just event nights?Inconclusive
Apart from the already-known pre-FOMC night, no identifiable event night carries the overnight drift.
- FPR-26-11Level breakouts: why the old daily picks lostRefuted
The old daily level-breakout picks lose after costs in every variant replayed. Found after the fact, on 2026 data only.
The record so far
7 of the 10 claims in the record are refuted and none is validated: one is in a forward test, one is a weak result found after the fact and one was only described. Separately, 13 setups sold to prop traders were tested on both index micros. With this many tests, a few passes by chance were expected. No study has yet been replicated outside the firm.
| Claim | Verdict | What the test found |
|---|---|---|
| Level breakouts | Refuted | Replayed after the fact with the firm's own entry mechanics on eight futures through 2026: about zero before costs, negative after. Not pre-registered; the multi-year opening-range test on the next line was. |
| Opening-range breaks | Refuted | Pre-registered: 320 variants with and without a relative-volume filter, plus a volatility-regime split. None held up after costs. |
| Gamma walls, cages and flow direction | Refuted | 730 days of strike-level dealer positioning. Every wall, cage and flow-direction rule failed. |
| Being long during the day session | Refuted | Eight futures, 2010–2026. None pays after costs; five of eight are negative. |
| Gap continuation in the first 30 minutes | Refuted | A coin flip after earnings nights and after quiet nights alike. |
| Turn-of-month strength | Refuted | Inverted: of the three calendar subsets tested, these four sessions are the only one with a negative NQ overnight average. |
| Crowd attention as a signal | Refuted | Pre-registered. Nights after attention spikes averaged more than the baseline, but the difference failed the multiple-testing bar (p = 0.117 against 0.0033). Not an edge. |
| Release days are wilder | Described, not tested | Described after the fact, not tested: CPI, PCE and payrolls mornings ran about 4% wider than an ordinary day, not multiples. |
| The pre-FOMC night on NQ | In forward test | Found in the 2020–2026 sample (+$972.70 a night, n = 53, t = 3.12, placebo p = 0.009) among about 544 cells tested that day. A frozen forward test needs 8 new FOMC nights before the firm calls it an edge; 1 is in. As studied it also crosses the 17:00–18:00 flat window. |
| The last half hour, 15:30→16:00 | Weak, post hoc | Found after the fact (t = 2.17) once about 25 variants had been screened, and negative on all four index futures in 2023–2026. Not an edge. |
| 13 setups sold to prop traders | 0 of 26 | Opening-range breakouts, ICT Silver Bullet, VWAP fades, EMA crosses and news straddles on NQ and ES micros, January 2019 – September 2026. None beat costs. |
These results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown.
How a study runs
The method is the product. The full standard, with its limits, is on the methods page.
- Write it down firstThe claim, the exact rule, the data, the costs and the bar a result has to clear are fixed before any result exists. Each pre-registration is now hashed before the run, so the order can be proved; the library shows which earlier studies fell short of this.
- Run it on the tapeMinute data split into in-sample and out-of-sample periods, commission and slippage on every fill, and controls such as a coin-flip version of the same trade, so a result has to beat luck.
- Publish the verdictValidated, refuted or too few trades, with the code and the limits, in the library. Most claims fail, and we say so.
Independent by design
The firm does not sell trade signals, market data, subscriptions or paid placement, and does not manage money. Every published study shows whether it was pre-registered and comes with its files, so you can check the verdict yourself. Market data stays in the lab; what is published is the verdict, the method and the code.
Questions
What is Futures Playbook Research?
An independent research firm for futures markets. It tests trading claims on CME futures data and publishes every study behind its claims with the files. Futures Playbook Research is a trade name of The Futures Playbook LLC, a Florida limited liability company.
Who runs it?
The firm was founded by Dr. Alex Djahankhah, a computer-science professor and U.S. Marine Corps veteran, who writes the research rules and runs the studies. About the firm.
Can we check your studies?
Yes. The library has the pre-registration, results and code for each of the 15 studies behind the record, including the ones that were not pre-registered and the defects found in the firm's own files. Market data is not included; you rerun them on your own.
Do you give trade signals or sell anything?
No. The studies report verdicts on rules, not recommendations, and the firm sells no signals, subscriptions or data. It retired its own daily picks in October 2026 when they showed no edge on the tape.
Your result differs from ours?
Rerun a study on your own data and tell us; a confirmed difference gets a dated correction in the library.