FREE RESOURCES

NO SIGNUP

Free resources. Start here.

Free tools to use before you trade, our research notebook with every demo trade shown, and free guides to keep. Pick a tab.

FREE TOOLS

RUNS IN YOUR BROWSER

Free tools.
Use them before you trade.

Every tool runs in your browser. Nothing is uploaded and nothing is stored. Educational and analytical use only — not financial advice.

THE LAB · RESEARCH IN PROGRESS

We test our own ideas in public. Failures included.

This is our research notebook: the questions each system is trying to answer, the data it records, what broke, and what we changed. It is the same evidence standard our books use. Most of what we test does not work, and we show that too.

  • Research, not a service. Nothing here is a signal, a track record or advice.
  • Delayed. Results are published after the week closes, never live.
  • Only broker data counts. Paper and shadow results are not used as evidence.

Updated with data to Fri 23 Oct 2026

01 · THE SYSTEMS

What each system is asking

MT5 BROKER DATA · EVIDENCE

Currency strength tracker

Which currencies are strongest and weakest, and does that persist?

Records 16 currencies every 15 minutes, with the market context alongside (yields, gold, oil, volatility).

Since 3 May 2026 · 19,000+ snapshots

MT5 BROKER DATA · EVIDENCE

Structure Lab

Do common market-structure ideas hold up on real broker prices?

Logs structure events on 16 markets from M5 to monthly charts. It records and measures; it places no orders.

Since 25 Jun 2026 · 41,000+ observations

AI-BUILT · TRADING SWITCHED OFF

ATLAS

Can a system built entirely by AI, with no human judgement in its decisions, find an edge?

So far, no. Every model it has tried has been retired (see below). It now runs for data and logging only, and each new idea must pass a test written down before the numbers are run.

Trading off since 17 Oct 2026

MT5 BROKER DATA

Market recorder

What did the week actually do, on the prices a trader really sees?

Weekly open, high, low and close for the 28 major pairs, gold, silver, US indices and crypto. It feeds our free weekly market review.

Weekly since Oct 2026

DEMO ACCOUNT · EXPERIMENT

RuleGate

Does a strict, rules-only EA hold up once it trades a real broker feed?

Rules only, 1% risk per trade, on a demo account. Every closed trade is shown below, losses included.

Demo since 19 Oct 2026 · 9 trades so far

02 · SCAR TISSUE

What broke, and what we changed

Most trading brands only show the wins. This is the log of what didn't work.

WhenSystemWhat failedWhat we changed
5 Oct 2026Paper and shadow logsResults varied widely with the broker environment, so they were not reliable enough to learn from.Dropped as evidence. Only records from real broker prices now count, here and in our books.
7 Oct 2026RuleGateResults in R were not being measured correctly.R is now the net result divided by the amount risked at entry, on every trade.
17 Oct 2026ATLASIts trend-pullback and range-fade models lost money over 47 trades.Both retired, and trading switched off.
18 Oct 2026Currency strength trackerDuplicate recorders were writing two or three identical rows at a time.Reduced to one recorder; duplicates are removed before any analysis.
20 Oct 2026ATLASDaily momentum failed its pre-registered test on three markets it had never seen (−0.27R on the held-out data).Retired, with no re-cut of the test to make it pass.
Oct 2026Market feed (energy)Our feed's oil and gas quotes did not match exchange settlements, because of contract rolls and stale quotes.Published energy figures now come from exchange settlements confirmed by two sources.
25 Oct 2026RuleGate statisticsTen minutes of readings from a different account made the drawdown show as 99.9%.Those readings are now filtered out. The real figure is 5.5%.

03 · RULEGATE · DEMO ACCOUNT

Every closed trade, a week behind

Demo account, 1% risk per trade, tracking since 19 Oct. Results in R: one R is the amount risked on that trade. No broker or account details.

Closed trades9
Won · lost3 · 6
Total−1.14R
Account change−1.19%
Worst drawdown5.5%

Cumulative result, trade by trade · dots: green = winning trade, red = losing trade

0R−1.14R

9 trades is far too few to judge anything. Several were on the same market on the same day, so the sample is even smaller than it looks. We will keep publishing every week, whatever the result.

#Closed (UTC)MarketSideResultExit
1Tue 20 OctUS100Short+2.02RTake profit
2Tue 20 OctEUR/CADShort+0.10RStop loss
3Wed 21 OctUS30Long−1.01RStop loss
4Wed 21 OctUS30Long+1.89RTake profit
5Wed 21 OctUS30Long−1.05RStop loss
6Thu 22 OctUS30Long−1.03RStop loss
7Thu 22 OctAUD/JPYShort−1.00RStop loss
8Fri 23 OctUS100Short−1.04RStop loss
9Fri 23 OctXAU/USDShort−0.02RStop loss

04 · WHY WE DO THIS

The same rule as our books

A claim needs its source, its sample and its period. When the evidence is not there, we say so rather than filling the gap. The Lab holds us to that standard on our own ideas first.

Research only. Not a forecast, a signal or financial advice. These are results from our own research systems and a demo account. They are not a track record, and nothing here recommends buying or selling any instrument or using any system. Past results do not predict future results. Trading leveraged products carries a high risk of losing money.

FREE GUIDES

FREE DOWNLOAD

Free guides.
Read before you spend.

All free guides ↗

Complete, evidence-led guides you can keep. Add one to your basket and check out at £0: the download link arrives by email, with no payment details needed.

Educational content only, not financial advice.

GO DEEPER

Short editions cover one job each, most with worksheets or a spreadsheet, and each comes free with its full book. The full library has the whole picture.