Our Default Setting Was the Only Losing One: What 100% Real Ticks Told Us About Our Own EA

24 August 2026, 01:00
Kenichiro Sakamoto
0
18
Our five-market portfolio EA shipped with four risk tiers and the highest one, Ultra, was the default. On the long tester-modelled window it produced the best-looking numbers, which is exactly why it was the default. Then we re-measured everything on 100% real ticks and changed it.

Why the long window is always modelled
Brokers only store genuine tick-by-tick history for the most recent months. A multi-year 100% real-tick backtest is therefore structurally impossible at any broker, not merely difficult. Every multi-year figure you have ever seen, ours included, is tester-modelled from bar data. That is not fraud, but it is a different measurement from the one your live account will perform.

What the real ticks showed
On the longest window where 100% real ticks exist, with all five legs on a $10,000 account:
Defensive: +$393 / profit factor 1.15 / max equity drawdown 11.4%
Standard: +$534 / profit factor 1.11 / max equity drawdown 22.7%
Aggressive: +$885 / profit factor 1.09 / max equity drawdown 39.0%
Ultra: -$146 / profit factor 0.99 / max equity drawdown 70.9%
A second broker's independent real-tick window confirmed the Ultra result at profit factor 0.90 with 75.1% equity drawdown. Two findings we had to act on: real-tick drawdowns run materially deeper than the long-window figures at every tier, and Ultra was the only losing tier.

What we changed
In version 1.47 the default moved from Ultra to Standard, and the real-tick table above is published on the product page above the more flattering long-window one. Users who preferred the previous behaviour can still select Ultra; they now do so having read what it measured.

The general lesson
The setting that looks best on a modelled multi-year run is often the one whose risk the model understates most, because modelled bars smooth away the intrabar sequence that decides where a leveraged position gets stopped. When comparing risk presets, weight the short real-tick window more heavily than the long modelled one, even though it is less impressive.
The EA and its full measurement history, real-tick table included: https://www.mql5.com/en/market/product/182751

Our measured backtest data for every EA (profit factor, equity drawdown, trade count, year-by-year results) is published at fxea365.com/ea/ranking