- What is the one idea behind Iridescence?
- That a market pushed away from where it was trading tends to come back, and that the push is usually short-lived. Most of the readings find that on the chart — price beyond a volatility band and turning, a momentum extreme, a spike that closed weakly after making a new high. Two find it somewhere other than price — a month-end flow that moves a pair for reasons unrelated to its chart, and the hour at which short-term direction turns over. It is the same event read seven ways, which is why the positions are short — half close within a day.
- How is this different from Lattice Weave?
- Weave takes one way of reading a market and applies it to seven pairs; positions there are held about twelve days on average. Prism reads a shorter-lived event, on faster timeframes, and half of its positions are closed within a day. They also share no trade — that is checked mechanically before either can be published, by comparing the two published ledgers row by row.
- Is Iridescence a martingale or grid EA?
- No. Each position opens at the lot you set, with a stop and a target attached when the order is sent. There is no averaging down, no grid, no recovery trade and no lot progression of any kind. What varies is how many of the seven readings happen to be in the market at the same time, not the size of any position.
- How was it tested?
- In the MetaTrader 5 strategy tester on M1 history, every tick generated from M1 bars, with the whole book running on one AUD/USD H1 chart at 0.1 lot per position on a $10,000 balance. The published backtest is the part of that run that ends at the 2026-04-01 cutoff; what the same build did after that date is shown separately as the verification window, so the two never share a trade. Every closed trade is published on this page and downloadable as CSV, including which symbol it was opened on.
- Does the backtest include swap?
- Yes, taken from the broker's own history — the per-deal swap is in the published trade ledger. Commission was zero in this run. Positions are short-lived by design, so swap matters less here than for a hold-for-weeks EA, but a broker with different rates will still produce a different result. One caveat worth stating plainly. A broker's stored swap history is revised over time, so re-running this exact build on the same dates does not necessarily reproduce the same figures to the cent. The figures on this page are the run whose report is published above.
- Why read it in seven places instead of one?
- Because in any single market this event is rare — several of the readings act only a handful of times a year, which is too thin to judge or to run on its own. Reading the same idea across five pairs and several timeframes is what turns it into something that trades often enough to be worth operating, and it means a quiet stretch in one market does not stop the account entirely.
- Can I turn parts of it off?
- No. The seven readings are compiled into the build and there is no switch for them, by design — the published record is one run with all of them working together, so any subset would be a different strategy with a different drawdown that nothing on this page describes. The lot is different — it is an input, so you can scale it to your own balance.
- My broker adds a suffix to its symbol names. Will it work?
- Set the symbol suffix input to whatever your broker appends, or leave it empty if the symbols have none — the build ships empty. A symbol that cannot be selected is skipped with a message in the Experts log and never trades, so a wrong suffix silently removes part of the book rather than trading the wrong thing.
- Why an AUD/USD chart when most of the trades are EUR/USD?
- Because the chart symbol is not traded. The EA opens all five pairs itself, so the chart it sits on only decides where it runs from — attach it to EUR/USD and it does exactly the same thing. AUD/USD is named because that is the chart the published record was measured on, and picking the same one is how you reproduce that record exactly. There is one measurable difference, and it is in the record rather than in the trading. Running the same test from a EUR/USD chart produces the identical 912 trades, the identical profit factor and the identical balance drawdown to the cent, but a shallower equity drawdown — 8.41% instead of 9.49% — because floating loss is sampled on the chart symbol's ticks, so which moments get looked at changes how deep the worst one appears. The deeper figure is the one a buyer would actually have sat through, so that is the run published here.
- Does it depend on my broker's server clock?
- Three of the readings do, because they trade a flow that happens at a particular time of day. The build converts server time to UTC assuming GMT+2 in winter and GMT+3 in summer, switching on the US daylight-saving dates — which is what the account the record was measured on does. There is no input to change it, so on a broker with a different offset those three act an hour away from where they were measured. The rest read price from the bar itself and behave the same anywhere.
- What happens when the spread widens?
- New entries are refused while the spread is wider than that reading's own guard. Open positions are never closed because of spread, so a widening spread can only stop an entry, not force an exit.
