‹ ALL STRATEGIES
About

How we test, and who we are

Tested and published by TradingBite Research·Updated 06 October 2026·How we test, and who we are·Tell us we got it wrong
Who runs this

TradingBite Research is one person: a futures trader who built the testing system behind this site — the data pipeline, the tests and the replay charts you can drive on every page. Not a broker, not a prop firm, not a course seller. Nothing on this site is sold to you for trading with anyone, and no result here pays us if you place a trade.

That matters more than it sounds. Almost every page that ranks for “best trading strategy” is published by someone who gets paid when you keep trading: brokers, prop firms, indicator vendors, course sellers. None of them can publish “this does not work”. We can, and mostly do — of the 21 strategies tested here, one has passed.

Reach us at hello@tradingbite.net. That address is read by the person who runs the tests.

How a strategy gets tested
  1. The rules come from the teacher, not from us. They are taken from a public video or write-up and coded exactly as stated — their entry, their stop, their target, their management, on the market they actually trade. If we think a rule is bad, it still gets tested as taught. A few pages test a textbook method (moving averages, RSI) using its standard definition, and a few test our own research; those are labelled as such.
  2. Where a teacher gives no number, we test the range. If someone says “a recent level” or “a tight stop”, every sensible value is run and all of them are published — not the one that looks best. The page says which version the replay is showing.
  3. Anything needing judgement is left out and named. Hand-drawn levels, “wait for confluence”, a target chosen by eye: we cannot code those honestly, so each page lists what was left out rather than guessing and calling it their rule.
  4. Costs are charged on every trade. Commission per round turn, on the contract stated. A strategy that only works before costs has not worked.
  5. Results are measured against a control wherever one can be run — the same trades taken on a coin-flip side, or a random entry held the same time. A rule that cannot beat that is measuring the market, not the method. Where no control has been run yet, the page says so.
  6. A range, not a single number. Wherever the test produces one, a result carries the 95% range for its average. If that range crosses zero, the honest answer is “unproven”, however good the average looks.
MarketsNasdaq (MNQ), gold (GC) and, on two pages, Bitcoin (CME) futures
Data1-minute bars from a commercial market-data vendor, front month by traded volume; the two order-flow pages use tick data
Costs$3.00 a round turn on MNQ, $25.00 on GC, $10.00 on CME Bitcoin
Stated per pagemarket, period, number of trades, session, bar size, costs
Slippagenot modelled — assume results are a little optimistic
What the three verdicts mean
Passedmakes money after costs, beats its own control, and still stands after allowing for how many versions were tried
Unprovendoes not clearly lose, but falls short of passing — the range crosses zero, it does not beat its control, it does not survive the count of versions tried, or its control is not yet run
Failedit clearly loses after costs, clearly loses to its own control, or the teacher's specific claim is clearly contradicted

“Failed” is used sparingly and only where the numbers force it. A strategy that merely fails to prove itself is called unproven, not failed — those are different things and the difference is the reader's money.

What we are not claiming

We are not testing the trader. A teacher can be a good trader and still teach rules that do not survive being followed literally. Their skill lives in the judgement they cannot write down. We test what they wrote down, because that is what you would be acting on.

We are not saying anyone is dishonest. No claim here is about anyone's honesty, their results or their character. If a teacher's rules were read incorrectly, that is our error to fix — tell us how they should be read and we will re-run the test and publish the outcome either way.

A backtest is not a promise. It says what those rules would have done on past data, with the costs stated, on one market. It cannot tell you what happens next, and it cannot tell you whether you will follow the rules when it matters.

What would change a verdict: a corrected reading of the rules, a longer or different data sample, a market the teacher actually trades that we have not tested, or a forward test that disagrees with the backtest. Any of those, and the page changes.

How this was produced

The rules come from the teacher's own public video or write-up — or, for a textbook method or our own research, the page says so. They are coded as stated and run over years of futures data from a commercial market-data vendor, with commission charged on every trade; slippage is not modelled. Where a teacher gives no number, every value in the plausible range is tested and all of them are shown — not only the best one. Each page states the market, the period, the sample and the costs used.

Written with software. The tests are code, and the code and much of the writing were produced with AI assistance. Every result comes from that research code. The words around the numbers are written from those results — if you find one that disagrees with its own numbers, tell us and it gets fixed.

If a number here is wrong, say so. Email hello@tradingbite.net with the page and what you think is wrong. If a teacher believes their rules were read incorrectly, tell us how they should be read and we will re-run the test and publish the result, whichever way it goes.

This is not advice. These are tests of publicly taught methods on historical data, published so you can check them yourself. Past results do not predict future results. Trading futures can lose you more than you put in. Terms and full disclaimer ›