01
Scan
The engine watches EURUSD around the clock, five days a week, entering against the recent move only when its volatility, spread and news gates all agree the tape is calm enough for mean reversion.
How it works
Martingale, explained the way we'd want it explained to us.
Grit Markets trades EURUSD around the clock, five days a week. It opens a small position against the recent move (mean reversion), and if price keeps going the wrong way it adds a larger position every 2.1 pips — each leg ×1.21 the last — so that a partial reversal closes the whole basket 3.4 pips in profit. Frequent small wins, occasional deep drawdowns. Hard limits decide how deep. That is the entire strategy — no black box.
The sizing math
Every leg is ×1.21 the last. Depth is what does the damage.
A ×1.21 ladder grows gently at first — which is exactly why it can go deep. By leg 24 (the depth this strategy class has actually reached in practice) a single leg trades over 60 times the base size and the basket has committed more than 300 times the base exposure. This table is why the max-legs cap and the equity stop exist — and why we publish it.
| Level | Lot | Cumulative lots |
|---|---|---|
| 1 | 0.01 | 0.01 |
| 5 | 0.02 | 0.08 |
| 9 | 0.05 | 0.22 |
| 13 | 0.10 | 0.52 |
| 17 | 0.21 | 1.17 |
| 21 | 0.45 | 2.56 |
| 24 | 0.80 | 4.57 |
The engine's shipped ×1.21 geometry — multiplier and cap configurable
The lifecycle
Four states. Nothing hidden between them.
02
Enter at base lot
Every sequence starts at the smallest size — the base lot you set. A winning first trade closes the sequence with a small profit, which is the most common single outcome.
03
Recover with compounding size
If price moves 2.1 pips against the basket, the engine adds a leg ×1.21 the size of the last, so a reversal to a nearer price closes the whole basket 3.4 pips in profit. This is the Martingale core, and it is where the risk lives.
04
Hit a limit, take the loss
If floating losses reach your profile's equity stop, the engine flattens the basket, realises the loss, and refuses new baskets for the rest of the day. A bounded loss taken on purpose is the feature; unbounded averaging is the failure mode this control exists to prevent — which is why every Grit Markets risk profile keeps the stop armed.
What can go wrong
The failure modes, before you pay us — not after.
The long adverse run
A strong trend without meaningful pullbacks can walk a sequence straight to its maximum level. When that happens the realised loss is many times larger than the wins that preceded it. The simulator on the home page shows how often this occurs at your settings.
Gap risk
Weekend gaps and news spikes can jump past step prices and stop levels. The news filter reduces exposure to scheduled events; it cannot remove gap risk entirely.
Margin exhaustion
Deep sequences demand margin exactly when floating losses are largest. Undersized accounts get margin-called before the sequence can resolve — which is why the max-level cap and equity stop exist, and why base-lot sizing matters more than any other setting.
The controls
Seven controls stand between the ladder and your account.
Equity stop
ALWAYS ARMED- trigger
- floating loss % of balance
- action
- flatten basket + halt
- scope
- every risk profile
The master safety net: if floating losses reach your profile's percentage of balance, the engine closes the entire basket and stops for the day. Running a Martingale without a stop is how classic Martingale accounts die — so every Grit Markets risk profile keeps it armed. The percentage varies by profile; the stop itself is not optional.
Max-legs cap
PROFILE-SET- ladder
- capped in every profile
- sizing
- ×1.21 geometric per leg
- uncapped
- not offered
A hard ceiling on how deep one basket may stack. The ladder grows geometrically, so this cap decides your worst case. The uncapped ladder is the historic behaviour of this strategy class — we say so plainly, and no Grit Markets profile ships it. Your profile choice moves the cap between shallower and deeper, never to unlimited.
Daily loss limit
ENFORCED- counts
- equity-stop events / day
- action
- no new baskets today
- survives
- VPS restarts
After a stop-out, the engine refuses to open new baskets for the rest of the day. One bad regime day cannot compound into three.
Event filter
KEYWORD-TIERED- watches
- USD · EUR calendar
- tier 1
- 24h before → 2h after
- tiers 2–3
- 60/30m before → 90/60m after
Events are classified by what they are, not by the calendar's star rating — because those ratings misrank the events that matter (a Fed statement can carry a lower rating than an oil-inventory print). Rate decisions, NFP and central-bank pressers get the widest window; CPI and GDP the next; other high-impact releases the base tier, with the calendar's own rating as the fallback so nothing slips through unclassified. Blocks new baskets and new legs, never the management of open positions.
Volatility gates
WATCHING- hourly ATR cap
- skip wild markets
- daily range gate
- skip after big days
- spread cap
- skip wide spreads
Mean-reversion grids do their worst in fast, trending tape. The engine measures hourly volatility, yesterday's range versus the 5-day norm, and the live spread — and simply declines to start baskets when the tape is hostile.
Calendar protection
SCHEDULED- friday
- cutoff before weekend
- monday
- 90-min warm-up
- rollover
- midnight blackout
No new baskets into the weekend close (gap risk), none in the choppy first minutes of the weekly reopen, and none during the broker's midnight rollover spread-spike. Holiday markets (24 Dec – 2 Jan) are blocked outright.
No-trade calendar
MAINTAINED- built from
- 16.5-yr stop-out study
- encodes
- 131 dates · 12 windows · 22 wks
- updates
- delivered automatically
We studied 16.5 years of this strategy class's failures — 1,372 stop-out events across 3,037 test windows — and encoded the historically dangerous calendar dates, weekday windows and weeks of the year as tiers the engine enforces. The calendar is maintained centrally and every licensed EA receives updates automatically. It removes historically hostile periods; it does not remove risk, and no study of the past binds the future.
Risk controls bound losses; they do not eliminate them. Tighter profiles mean smaller worst cases and more frequent realised losing baskets; looser profiles do the reverse. You choose the profile — the equity stop stays armed in all of them. No configuration of Grit Markets removes the risk of losing capital.