SFRECOVERY EA v3.34 · PRESETS · XAUUSD · 10K USD ACCOUNT

SFRecovery EA presets: 7 .set files, input by input.

SFRecovery EA is a recovery EA for MT5: it has no entry signal and manages the positions you already have when they fall into drawdown. These are the seven presets that ship with v3.34 (the same inputs as v3.32), calibrated on XAUUSD for a 10,000 USD account. For each one: what it is for, what it watches, what launches a cycle, how the grid is built, where the basket exits and which loss caps stop it. Every value on this page is read from the .set files; inputs a preset does not change keep the EA default.

Seven presets, side by side.

Dollar amounts are for a 10,000 USD account and scale with your equity. Lots do not scale on their own.

READ FIRST

A preset is a set of limits, not a strategy. SFRecovery is a risk manager: averaging down turns many small losses into fewer, larger ones, so every live preset ships with a hard money cap on the basket and an absolute close on the account. Nothing on this page is a performance figure. A backtest does not predict live results; test the preset you pick on a demo account with the exact lot sizes you plan to run.

Input / behaviour Guardian Control AW_Zona Equilibrado Conservador Agresivo OPTIMIZAR
UseSmallest grid, start hereTakes the whole symbol at 2% DDZONE dual-side grid0.05 grid, mid clock0.05 grid, no clocks0.05 grid, fast clockStrategy Tester only
Mode InpModeATRATRATRKALMANKALMANKALMANATR
AccountHedging advisedHedging advisedHedging onlyHedging advisedHedging advisedHedging advisedTester
WatchesMagic 0Magic 0 + whole symbol at 2%Magic 0Magic 0Magic 0Magic 0Magic 0
Cycle trigger$15$15$15$50$50$50$15
Grid lots (6 adds)6 × 0.01 fixed6 × 0.01 fixed6 × 0.01 fixed per side0.05 ×1.5 = 1.010.05 ×1.5 = 1.010.05 ×1.5 = 1.016 × 0.01 fixed
Step growth×1.15 per level×1.15 per level×1.15 per levelconstantconstantconstant×1.15 per level
Clocks StepBars / CycleMaxBars96 / 2496 / 2496 / 2496 / 24off / off12 / 24off / off
Basket target$25$25$25$40$40$40$25
Hard money cap$250$250$250$400$400$400$250
Absolute close10%10%10%12%12%12%OFF
CB tiers3/4/6%3/4/6%3/4/6%2/4/6%2/4/6%2/4/6%3/4/6%
Max basket lotsoff2.02.0offoffoffoff
ZONE / CONTROL-CONTROL 2%ZONE----
Own magic7301737301737301737301737301737301730
Download.set.set.set.set.set.set.set
← scroll the table →
SAME INPUTS AS v3.32

The v3.34 files carry exactly the same inputs as the v3.32 ones; only the header comment changed. The engine is what changed: with these DCA_FIRST presets the basket target and the InpCycleMaxBars exit now fire, while on v3.32 they never did. Expect shorter cycles that close at their target. More in what is new in v3.34.

Pick by what you trade, not by the name.

The names come from the original Spanish files. What matters is the size of your positions and your account type.

01

Guardian

You trade XAUUSD by hand with small positions, or you are trying SFRecovery for the first time. Smallest grid, tightest caps.

0.01
02

Control

Several EAs or a webhook bridge trade the same symbol and you want one manager to take the whole book when the account is 2% below its peak.

CONTROL
03

AW_Zona

You have a hedging account and want a zone recovery grid, the AW Recovery-style construction, working both sides.

ZONE
04

Equilibrado · Conservador · Agresivo

Your positions are 0.10 lot or more on a 10k account. Same 0.05-lot geometric grid and caps; they differ only in the bar clocks: mid, none, fast.

0.05
05

OPTIMIZAR

Only in the Strategy Tester, as the base for your own optimization. Never on a live account as it ships.

TESTER

What every preset has in common.

These inputs carry the same value in all seven .set files.

Averaging, no lock

InpRecoveryStyle=0 (DCA_FIRST) and InpUseLock=false: the EA averages the losing side with a grid and never hedges it with a lock. The scale-out staircase (InpScaleTiers=4 and the other InpScale* inputs) only works in LOCK_FIRST, so here it is inactive; since v3.34 it no longer blocks the basket target.

What happens at launch

InpDeleteSLTP=true, InpCloseProfit=true and InpDeletePending=true: when a cycle starts, the EA removes the SL and TP of the watched losing positions (from then on only the engine exits them), closes the watched positions that are in profit and deletes the watched pending orders on the symbol.

Money trigger only

Cycles launch on the floating loss in dollars (InpDDMoney, per preset). InpDDPercent=0.0, InpInstantLaunch=false and InpLaunchPullbackBars=0. The first add opens at launch by design; the next ones need one more grid step or the bar clock.

Grid spacing

InpUniversal=true: on metals the first step is 0.35 × the D1 ATR(14), widened by InpStepMult at each level. InpStep=200 is only a fallback when no ATR is available, and InpATRMult only applies with universal mode off. Up to InpMaxOrders=6 per grid, one order per bar.

Pair partials

Once InpOverlapAfter adds are open, pairs of 0.01 lot are closed when a recovery leg is in profit and at least 10 points from its entry, and the pair (that leg plus the worst watched loser) nets at least -$2 with commission (InpPartMinNet=-2.0): the basket shrinks during the cycle. No broker-side TP per leg (InpAvgTPPoints=0) and no auto-buyout.

Basket exit

The basket closes at break-even plus InpBasketMoney and a $1 exit-cost cushion (InpExitCostBuffer=1.0) so the close-side commission and spread do not turn a green exit red. InpBasketPerLot=0.0: the target does not grow with lots.

Kill scope and halt

InpKillScope=0 (MANAGED): a protection closes only the recovery basket, not your other EAs. InpHaltGuard=false. The halt is still published as the terminal global variable SFR_HALT_<symbol> (InpHaltGlobalVar=true). Circuit-breaker cooldown: InpCBCooldown=24.

Protections left off

Daily equity stop (InpDailyLossPct=0.0), day halt after N kills (InpMaxDailyKills=0), regime gate (InpGateRegime=false) and the worst-case refusal (InpMaxEquityRiskPct=0.0, InpMaxAccountWorstPct=0.0: the worst case is shown, never enforced). On a prop-firm account set InpDailyLossPct yourself.

Execution guards

Max slippage 30 points; new entries blocked above 100 points of spread (InpMaxSpreadEntry=100) and around the daily rollover (InpRolloverGuard=true). The spread guard on DCA adds is InpMaxSpreadProt=7500: on gold that is a $75 spread, so in practice it does not block adds.

Money auto-scale

InpAutoScale=true and InpAutoEquityScale=true: trigger, target, cap, partial threshold and exit cushion are multiplied by your equity / 10,000 and by the symbol type. Lots are not. See sizing.

Telegram

InpTelegram=false: no alerts until you add your bot token and chat ID and switch it on. Since v3.34 alerts are queued and sent from the timer, so they never block the tick.

Both directions

InpMultiDir=true: each cycle recovers whichever side is losing. Trades you open from the on-chart pad (the BUY / SELL buttons) carry magic 0 since v3.34, so these presets manage them.

Each preset, in detail.

Values from the v3.34 .set files, with the input name next to each one so you can check it in the EA window.

Guardian

START HERE

SFRecovery_v3.34_XAUUSD_Guardian_10k.set

What it is for

The smallest grid in the set and the default starting point: fixed 0.01-lot adds, an early $15 trigger, a $25 basket target and the tightest caps of the set ($250 on the basket, 10% absolute close, shared with Control and AW_Zona). It does the classic recovery job and nothing more: average the losing side a little, shrink the basket with pair partials, and get out at break-even plus a small target.

Who it suits

Manual traders on XAUUSD with small positions (0.01 to 0.10 lot on 10k), first-time users and anyone trying SFRecovery on a demo account.

Risks of this preset

  • A $15 trigger is close to everyday noise on gold: cycles launch often, and every launch removes the SL and TP of your losing watched trades and closes the watched trades that are in profit.
  • The grid is small next to a large position: if you trade 0.50 lot yourself, six adds of 0.01 barely move the break-even, and the $250 cap is what ends the cycle.
  • No lot cap on the basket (InpMaxBasketLots=0). The $250 cap and the circuit breaker re-arm after their cooldown; only the 10% absolute close stops the EA for good, so several capped cycles in a row add up.
  • v3.34 changes how it behaves: the $25 target and the 24-bar small-positive exit now fire. On v3.32 neither did, so cycles that stayed open for days now close much earlier.
Recovery styleDCA_FIRST: averaging grid on the losing side, no hedge lock (InpRecoveryStyle=0, InpUseLock=false)
Recovery modeATR: grid spaced by the daily ATR (InpMode=2)
Account typeHedging recommended: pair partials need opposite positions; on netting they switch off
What it watchesManual trades and on-chart pad trades (magic 0). Other EAs are ignored unless you add their magic numbers (InpMagicForRecovery=0)
Cycle triggerFloating loss of the watched positions ≥ $15; % trigger and instant launch off (InpDDMoney=15.0)
Grid lotsFixed 0.01 lot per add; 6 adds = 0.06 lot (InpLotMode=2, InpAvgLot=0.01; InpAvgMult not used)
Grid step0.35 × D1 ATR(14) at the first level, ×1.15 wider at each level. The 6th add sits about 2.71 D1 ATRs from the 1st if every add waits for distance (InpStepMult=1.15)
Max adds6 per grid, one order per bar; the first one opens at launch (InpMaxOrders=6)
Bar clockstime-based add after 96 bars without distance, only while the basket loses; after 24 bars in cycle, exits at a small positive (about $2) (InpStepBars=96, InpCycleMaxBars=24)
Partial closesOnce 2 adds are open: a profitable recovery leg is paired with the worst watched loser and 0.01 lot of each is closed if the pair nets at least -$2 (InpOverlapAfter=2, InpPartMinNet=-2.0)
Basket target$25 over break-even, plus a $1 exit-cost cushion (InpBasketMoney=25.0)
Hard money capBasket loss ≥ $250: flatten the basket, then cooldown (InpMaxLossMoney=250.0)
Absolute close10% equity drawdown from the true peak: close everything and stop the EA for good (InpAbsoluteClose=10.0)
Circuit breaker3% halves new lots · 4% stops new adds · 6% flattens + 24-bar cooldown, then re-arms (InpCBTier1=3.0)
Max basket lotsOff (InpMaxBasketLots=0.0)
ZONE / CONTROLBoth off (InpDualSide=0, InpTakeControlPct=0.0)
Own magic730173: a dedicated number, so after a reload the EA re-adopts its own legs instead of treating them as trades to recover (InpMagic=730173)
Panel, alerts, testerPanel on · launch alert on (Telegram stays off until you add a token and chat ID) · tester auto-entries off

Dollar amounts at 10,000 USD equity on XAUUSD; they scale with your equity. Bar counts follow the chart timeframe.

Control

WHOLE SYMBOL BOOK

SFRecovery_v3.34_XAUUSD_Control_10k.set

What it is for

Guardian plus CONTROL mode. Until the account is 2% below its equity peak it behaves like Guardian and only watches manual trades. At 2% it takes control of every position on the symbol, whatever EA opened it, forces the ZONE mirror grid and works the whole book until the basket exits at its target or a cap fires. A 2.0-lot cap bounds the doubled exposure.

Who it suits

Traders who run several EAs or a webhook bridge on the same symbol and want one manager to take the wheel when the account as a whole is in drawdown, not just one trade.

Risks of this preset

  • Once engaged it treats every position on the symbol as part of the basket: it strips their SL and TP, closes the ones in profit at launch and closes them all if a protection fires. Your other EAs keep running and can keep opening trades into that book.
  • ZONE works both sides at once: more legs, more spread and swap. The 2.0-lot cap counts your own positions too: if the book is already near 2 lots, no add opens.
  • Kills while in control wait only 6 bars before it re-engages. Control is released only when the managed book is flat and the drawdown is back under 1%. The 10% absolute close is the only stop for good.
  • The 2% threshold is account equity drawdown: losses on other symbols count toward it, and then it takes over this symbol.
Recovery styleDCA_FIRST: averaging grid on the losing side, no hedge lock (InpRecoveryStyle=0, InpUseLock=false)
Recovery modeATR: grid spaced by the daily ATR (InpMode=2)
Account typeHedging recommended: on netting the takeover still happens, but without the mirror grid and without pair partials
What it watchesManual trades and on-chart pad trades (magic 0). Other EAs are ignored unless you add their magic numbers. In CONTROL it takes every position on the symbol, any magic (InpMagicForRecovery=0)
Cycle triggerFloating loss of the watched positions ≥ $15; % trigger and instant launch off (InpDDMoney=15.0)
Grid lotsFixed 0.01 lot per add; 6 adds = 0.06 lot (InpLotMode=2, InpAvgLot=0.01; InpAvgMult not used)
Grid step0.35 × D1 ATR(14) at the first level, ×1.15 wider at each level. The 6th add sits about 2.71 D1 ATRs from the 1st if every add waits for distance (InpStepMult=1.15)
Max adds6 per grid, one order per bar; the first one opens at launch (InpMaxOrders=6)
Bar clockstime-based add after 96 bars without distance, only while the basket loses; after 24 bars in cycle, exits at a small positive (about $2) (InpStepBars=96, InpCycleMaxBars=24)
Partial closesOnce 2 adds are open: a profitable recovery leg is paired with the worst watched loser and 0.01 lot of each is closed if the pair nets at least -$2 (InpOverlapAfter=2, InpPartMinNet=-2.0)
Basket target$25 over break-even, plus a $1 exit-cost cushion (InpBasketMoney=25.0)
Hard money capBasket loss ≥ $250: flatten the basket, then cooldown (InpMaxLossMoney=250.0)
Absolute close10% equity drawdown from the true peak: close everything and stop the EA for good (InpAbsoluteClose=10.0)
Circuit breaker3% halves new lots · 4% stops new adds · 6% flattens + 24-bar cooldown, then re-arms (InpCBTier1=3.0)
Max basket lots2.0 lots, your watched positions included: no add once exceeded (InpMaxBasketLots=2.0)
ZONE / CONTROLCONTROL at 2% equity drawdown from the peak: ZONE forced, 6-bar cooldown after a kill, released when the book is flat and drawdown is below 1% (InpTakeControlPct=2.0)
Own magic730173: a dedicated number, so after a reload the EA re-adopts its own legs instead of treating them as trades to recover (InpMagic=730173)
Panel, alerts, testerPanel on · launch alert on (Telegram stays off until you add a token and chat ID) · tester auto-entries off

Dollar amounts at 10,000 USD equity on XAUUSD; they scale with your equity. Bar counts follow the chart timeframe.

AW_Zona

ZONE · HEDGING ONLY

SFRecovery_v3.34_XAUUSD_AW_Zona_10k.set

What it is for

Guardian with the ZONE dual-side grid always on. After the first add, a mirror grid opens on the opposite side with the same lots and steps, so the basket works both directions of the swing: the zone recovery construction that AW Recovery made popular, here as an AW Recovery alternative for MT5. A 2.0-lot cap bounds the total exposure.

Who it suits

Traders with a hedging account who know zone recovery and want to compare an AW Recovery-style setup on MT5.

Risks of this preset

  • Hedging account required. On netting the EA switches ZONE off and warns in the journal: you get a one-sided Guardian grid with a lot cap.
  • Two grids of up to six adds each: up to twelve own legs paying spread and swap on both sides.
  • In a strong one-way trend one side fills while the other books little; the $250 cap and the 10% absolute close are what end that.
  • The 2.0-lot cap includes your own watched positions.
Recovery styleDCA_FIRST: averaging grid on the losing side, no hedge lock (InpRecoveryStyle=0, InpUseLock=false)
Recovery modeATR: grid spaced by the daily ATR (InpMode=2)
Account typeHedging only: on a netting account the EA switches ZONE off and warns in the journal
What it watchesManual trades and on-chart pad trades (magic 0). Other EAs are ignored unless you add their magic numbers (InpMagicForRecovery=0)
Cycle triggerFloating loss of the watched positions ≥ $15; % trigger and instant launch off (InpDDMoney=15.0)
Grid lotsFixed 0.01 lot per add; 6 adds = 0.06 lot per side (InpLotMode=2, InpAvgLot=0.01; InpAvgMult not used)
Grid step0.35 × D1 ATR(14) at the first level, ×1.15 wider at each level. The 6th add sits about 2.71 D1 ATRs from the 1st if every add waits for distance (InpStepMult=1.15)
Max adds6 per grid, one order per bar; the first one opens at launch (InpMaxOrders=6)
Bar clockstime-based add after 96 bars without distance, only while the basket loses; after 24 bars in cycle, exits at a small positive (about $2) (InpStepBars=96, InpCycleMaxBars=24)
Partial closesOnce 2 adds are open: a profitable recovery leg is paired with the worst watched loser and 0.01 lot of each is closed if the pair nets at least -$2 (InpOverlapAfter=2, InpPartMinNet=-2.0)
Basket target$25 over break-even, plus a $1 exit-cost cushion (InpBasketMoney=25.0)
Hard money capBasket loss ≥ $250: flatten the basket, then cooldown (InpMaxLossMoney=250.0)
Absolute close10% equity drawdown from the true peak: close everything and stop the EA for good (InpAbsoluteClose=10.0)
Circuit breaker3% halves new lots · 4% stops new adds · 6% flattens + 24-bar cooldown, then re-arms (InpCBTier1=3.0)
Max basket lots2.0 lots, your watched positions included: no add once exceeded (InpMaxBasketLots=2.0)
ZONE / CONTROLZONE on: a mirror grid opens on the opposite side after the 1st add, same lots and steps (InpDualSide=1)
Own magic730173: a dedicated number, so after a reload the EA re-adopts its own legs instead of treating them as trades to recover (InpMagic=730173)
Panel, alerts, testerPanel on · launch alert on (Telegram stays off until you add a token and chat ID) · tester auto-entries off

Dollar amounts at 10,000 USD equity on XAUUSD; they scale with your equity. Bar counts follow the chart timeframe.

Equilibrado (balanced)

0.05 GRID · MID CLOCK

SFRecovery_v3.34_XAUUSD_Equilibrado_10k.set

What it is for

A heavier geometric grid for bigger positions: the first add is 0.05 lot and each next one is 1.5 times larger, with constant spacing. It launches at a $50 loss, targets $40 over break-even and is capped at $400 on the basket and 12% on the account. Middle clock: a time-based add after 96 bars without distance and a small-positive exit after 24 bars.

Who it suits

Traders whose losing positions are 0.10 lot or more on a 10k account, who want the grid to have weight next to that position and accept bigger swings than Guardian.

Risks of this preset

  • Six adds total 1.01 lots, about 17 times the Guardian grid, and each add is bigger than the last one.
  • No lot cap on the basket, and the 5-lot per-order cap never binds at this size: the $400 cap and the 12% absolute close are the real limits.
  • Constant spacing (StepMult 1.0): in a trend the grid fills faster than in the Guardian family, where each step is 15% wider.
  • KALMAN mode does not protect the grid here: the filter only gates pad and tester entries, not DCA adds against the trend.
Recovery styleDCA_FIRST: averaging grid on the losing side, no hedge lock (InpRecoveryStyle=0, InpUseLock=false)
Recovery modeKALMAN: the same ATR grid, plus a Kalman trend filter on entries from the on-chart pad and the tester. DCA adds are not filtered (InpMode=3, InpDcaKalmanGate=false)
Account typeHedging recommended: pair partials need opposite positions; on netting they switch off
What it watchesManual trades and on-chart pad trades (magic 0). Other EAs are ignored unless you add their magic numbers (InpMagicForRecovery=0)
Cycle triggerFloating loss of the watched positions ≥ $50; % trigger and instant launch off (InpDDMoney=50.0)
Grid lotsGeometric from 0.05 lot, ×1.5 per add: 0.05 / 0.07 / 0.11 / 0.16 / 0.25 / 0.37; 6 adds = 1.01 lots. Per-order cap 5 lots (InpLotMode=0, InpMaxAvgVolume=5.0)
Grid step0.35 × D1 ATR(14) at the first level, same spacing at every level. The 6th add sits about 1.75 D1 ATRs from the 1st if every add waits for distance (InpStepMult=1.0)
Max adds6 per grid, one order per bar; the first one opens at launch (InpMaxOrders=6)
Bar clockstime-based add after 96 bars without distance, only while the basket loses; after 24 bars in cycle, exits at a small positive (about $2) (InpStepBars=96, InpCycleMaxBars=24)
Partial closesOnce 1 add is open: a profitable recovery leg is paired with the worst watched loser and 0.01 lot of each is closed if the pair nets at least -$2 (InpOverlapAfter=1, InpPartMinNet=-2.0)
Basket target$40 over break-even, plus a $1 exit-cost cushion (InpBasketMoney=40)
Hard money capBasket loss ≥ $400: flatten the basket, then cooldown (InpMaxLossMoney=400.0)
Absolute close12% equity drawdown from the true peak: close everything and stop the EA for good (InpAbsoluteClose=12.0)
Circuit breaker2% halves new lots · 4% stops new adds · 6% flattens + 24-bar cooldown, then re-arms (InpCBTier1=2.0)
Max basket lotsOff (InpMaxBasketLots=0.0)
ZONE / CONTROLBoth off (InpDualSide=0, InpTakeControlPct=0.0)
Own magic730173: a dedicated number, so after a reload the EA re-adopts its own legs instead of treating them as trades to recover (InpMagic=730173)
Panel, alerts, testerPanel on · launch alert off (Telegram stays off until you add a token and chat ID) · tester auto-entries off

Dollar amounts at 10,000 USD equity on XAUUSD; they scale with your equity. Bar counts follow the chart timeframe.

Conservador (no clocks)

0.05 GRID · NO CLOCKS

SFRecovery_v3.34_XAUUSD_Conservador_10k.set

What it is for

The same 0.05-lot geometric grid, trigger, target and caps as Equilibrado, without the clocks: it adds only when price moves a full step against the basket, and the basket always waits for its full $40 target. Conservative here means it never adds just because time passed; it does not mean lower risk.

Who it suits

Traders who want the grid to react only to price, not to time, and accept long cycles.

Risks of this preset

  • Without the small-positive exit, a basket that hovers just under its target in a sideways market stays open, paying swap, until it gets there or a cap fires.
  • Same 1.01-lot grid and the same caps as Equilibrado: the exposure is not smaller, it just builds on distance only.
  • Long cycles are more exposed to news and weekend gaps.
Recovery styleDCA_FIRST: averaging grid on the losing side, no hedge lock (InpRecoveryStyle=0, InpUseLock=false)
Recovery modeKALMAN: the same ATR grid, plus a Kalman trend filter on entries from the on-chart pad and the tester. DCA adds are not filtered (InpMode=3, InpDcaKalmanGate=false)
Account typeHedging recommended: pair partials need opposite positions; on netting they switch off
What it watchesManual trades and on-chart pad trades (magic 0). Other EAs are ignored unless you add their magic numbers (InpMagicForRecovery=0)
Cycle triggerFloating loss of the watched positions ≥ $50; % trigger and instant launch off (InpDDMoney=50.0)
Grid lotsGeometric from 0.05 lot, ×1.5 per add: 0.05 / 0.07 / 0.11 / 0.16 / 0.25 / 0.37; 6 adds = 1.01 lots. Per-order cap 5 lots (InpLotMode=0, InpMaxAvgVolume=5.0)
Grid step0.35 × D1 ATR(14) at the first level, same spacing at every level. The 6th add sits about 1.75 D1 ATRs from the 1st if every add waits for distance (InpStepMult=1.0)
Max adds6 per grid, one order per bar; the first one opens at launch (InpMaxOrders=6)
Bar clocksadds only on price distance; always waits for the full basket target (InpStepBars=0, InpCycleMaxBars=0)
Partial closesOnce 1 add is open: a profitable recovery leg is paired with the worst watched loser and 0.01 lot of each is closed if the pair nets at least -$2 (InpOverlapAfter=1, InpPartMinNet=-2.0)
Basket target$40 over break-even, plus a $1 exit-cost cushion (InpBasketMoney=40)
Hard money capBasket loss ≥ $400: flatten the basket, then cooldown (InpMaxLossMoney=400.0)
Absolute close12% equity drawdown from the true peak: close everything and stop the EA for good (InpAbsoluteClose=12.0)
Circuit breaker2% halves new lots · 4% stops new adds · 6% flattens + 24-bar cooldown, then re-arms (InpCBTier1=2.0)
Max basket lotsOff (InpMaxBasketLots=0.0)
ZONE / CONTROLBoth off (InpDualSide=0, InpTakeControlPct=0.0)
Own magic730173: a dedicated number, so after a reload the EA re-adopts its own legs instead of treating them as trades to recover (InpMagic=730173)
Panel, alerts, testerPanel on · launch alert off (Telegram stays off until you add a token and chat ID) · tester auto-entries off

Dollar amounts at 10,000 USD equity on XAUUSD; they scale with your equity. Bar counts follow the chart timeframe.

Agresivo (fast clock)

0.05 GRID · FAST CLOCK

SFRecovery_v3.34_XAUUSD_Agresivo_10k.set

What it is for

The same grid, trigger, target and caps as Equilibrado with the fastest clock: if price has not moved a full step after 12 bars and the basket is still losing, it adds anyway, and after 24 bars it takes any small positive. The grid reaches full size much sooner.

Who it suits

Experienced users who want the grid to keep working in slow grinds and accept reaching the full 1.01-lot grid quickly.

Risks of this preset

  • Time-based adds do not wait for price: in a slow trend all six adds can be open about 60 bars after launch (2.5 days on H1) without price moving a full step between them.
  • The fastest route to the largest exposure of the set; the $400 cap and the 12% absolute close are what stop it.
  • Same KALMAN caveat as Equilibrado: DCA adds are not filtered by the trend.
Recovery styleDCA_FIRST: averaging grid on the losing side, no hedge lock (InpRecoveryStyle=0, InpUseLock=false)
Recovery modeKALMAN: the same ATR grid, plus a Kalman trend filter on entries from the on-chart pad and the tester. DCA adds are not filtered (InpMode=3, InpDcaKalmanGate=false)
Account typeHedging recommended: pair partials need opposite positions; on netting they switch off
What it watchesManual trades and on-chart pad trades (magic 0). Other EAs are ignored unless you add their magic numbers (InpMagicForRecovery=0)
Cycle triggerFloating loss of the watched positions ≥ $50; % trigger and instant launch off (InpDDMoney=50.0)
Grid lotsGeometric from 0.05 lot, ×1.5 per add: 0.05 / 0.07 / 0.11 / 0.16 / 0.25 / 0.37; 6 adds = 1.01 lots. Per-order cap 5 lots (InpLotMode=0, InpMaxAvgVolume=5.0)
Grid step0.35 × D1 ATR(14) at the first level, same spacing at every level. The 6th add sits about 1.75 D1 ATRs from the 1st if every add waits for distance (InpStepMult=1.0)
Max adds6 per grid, one order per bar; the first one opens at launch (InpMaxOrders=6)
Bar clockstime-based add after 12 bars without distance, only while the basket loses; after 24 bars in cycle, exits at a small positive (about $2) (InpStepBars=12, InpCycleMaxBars=24)
Partial closesOnce 1 add is open: a profitable recovery leg is paired with the worst watched loser and 0.01 lot of each is closed if the pair nets at least -$2 (InpOverlapAfter=1, InpPartMinNet=-2.0)
Basket target$40 over break-even, plus a $1 exit-cost cushion (InpBasketMoney=40)
Hard money capBasket loss ≥ $400: flatten the basket, then cooldown (InpMaxLossMoney=400.0)
Absolute close12% equity drawdown from the true peak: close everything and stop the EA for good (InpAbsoluteClose=12.0)
Circuit breaker2% halves new lots · 4% stops new adds · 6% flattens + 24-bar cooldown, then re-arms (InpCBTier1=2.0)
Max basket lotsOff (InpMaxBasketLots=0.0)
ZONE / CONTROLBoth off (InpDualSide=0, InpTakeControlPct=0.0)
Own magic730173: a dedicated number, so after a reload the EA re-adopts its own legs instead of treating them as trades to recover (InpMagic=730173)
Panel, alerts, testerPanel on · launch alert off (Telegram stays off until you add a token and chat ID) · tester auto-entries off

Dollar amounts at 10,000 USD equity on XAUUSD; they scale with your equity. Bar counts follow the chart timeframe.

OPTIMIZAR (optimize)

TESTER ONLY

SFRecovery_v3.34_XAUUSD_OPTIMIZAR.set

What it is for

Not a live preset: a starting point for your own Strategy Tester optimizations. It keeps the Guardian grid and the $250 basket cap, turns the tester's automatic entries on (a buy every 50 bars, so runs are reproducible), removes the clocks and the panel, and leaves the absolute close OFF so the optimizer sees full runs.

Who it suits

Users who want to tune the inputs on their own symbol and dates in the MT5 Strategy Tester. Load it from the tester's Inputs tab, use real ticks, tick the inputs you want to vary and judge the passes by drawdown and by how often the caps fire, not by profit: the tester entries have no edge, and the best pass of an optimization is fitted to that period. Re-run it on other dates before trusting it.

Risks of this preset

  • Absolute close OFF: on a live account the only hard limits would be the $250 cap and the circuit breaker, which re-arm.
  • Own magic 0, the same magic it watches: the EA warns about it at start, and after a reload its own legs could be taken as trades to recover.
  • Before going live with an optimized set: give InpMagic its own number, set InpAbsoluteClose, switch the panel on and run it on a demo first. The automatic entries do nothing on a live account.
Recovery styleDCA_FIRST: averaging grid on the losing side, no hedge lock (InpRecoveryStyle=0, InpUseLock=false)
Recovery modeATR: grid spaced by the daily ATR (InpMode=2)
Account typeStrategy Tester (use a hedging account model)
What it watchesManual trades and on-chart pad trades (magic 0). Other EAs are ignored unless you add their magic numbers (InpMagicForRecovery=0)
Cycle triggerFloating loss of the watched positions ≥ $15; % trigger and instant launch off (InpDDMoney=15.0)
Grid lotsFixed 0.01 lot per add; 6 adds = 0.06 lot (InpLotMode=2, InpAvgLot=0.01; InpAvgMult not used)
Grid step0.35 × D1 ATR(14) at the first level, ×1.15 wider at each level. The 6th add sits about 2.71 D1 ATRs from the 1st if every add waits for distance (InpStepMult=1.15)
Max adds6 per grid, one order per bar; the first one opens at launch (InpMaxOrders=6)
Bar clocksadds only on price distance; always waits for the full basket target (InpStepBars=0, InpCycleMaxBars=0)
Partial closesOnce 2 adds are open: a profitable recovery leg is paired with the worst watched loser and 0.01 lot of each is closed if the pair nets at least -$2 (InpOverlapAfter=2, InpPartMinNet=-2.0)
Basket target$25 over break-even, plus a $1 exit-cost cushion (InpBasketMoney=25.0)
Hard money capBasket loss ≥ $250: flatten the basket, then cooldown (InpMaxLossMoney=250.0)
Absolute closeOFF (InpAbsoluteClose=0.0)
Circuit breaker3% halves new lots · 4% stops new adds · 6% flattens + 24-bar cooldown, then re-arms (InpCBTier1=3.0)
Max basket lotsOff (InpMaxBasketLots=0.0)
ZONE / CONTROLBoth off (InpDualSide=0, InpTakeControlPct=0.0)
Own magic0: the same magic it watches. Fine in the tester; on a live account the EA warns that its own trades will be swept into recovery (InpMagic=0)
Panel, alerts, testerPanel off · launch alert off (Telegram stays off until you add a token and chat ID) · tester auto-entries on, buys only every 50 bars

Dollar amounts at 10,000 USD equity on XAUUSD; they scale with your equity. Bar counts follow the chart timeframe.

Inputs → Load, then read the journal.

A .set file only fills the input fields. Nothing trades until the EA is attached and Algo Trading is on.

01

Download the .set

Pick one preset above or the zip with all seven. Optional: copy the files into the MT5 presets folder (File → Open Data Folder → MQL5 → Presets) so the Load dialog finds them straight away.

.set
02

Attach and load

Drag SFRecovery EA onto the XAUUSD chart. In the window that opens go to the Inputs tab, click Load, pick the .set and press OK; in the Common tab allow Algo Trading. If the EA is already running, press F7 on that chart to reopen the same window.

F7
03

Point it at your trades

The presets watch manual trades (magic 0). To manage an EA, put its magic number in InpMagicForRecovery (several: comma-separated; * = every position on the symbol). Never list 730173, the presets' own magic.

MAGIC
04

Read the journal

The Experts tab prints the build stamp (v3.34) and a CONFIG SUMMARY: the trigger in dollars after auto-scale, every cap as ON or OFF, and a NOT USED line with the inputs your setup ignores. Check those numbers before you leave it running.

LOG
STRATEGY TESTER

Settings tab: SFRecovery EA, XAUUSD, H1, modelling Every tick based on real ticks. Inputs tab: right-click → Load → pick the .set. The six live presets ship with InpTestAuto=false, so in a non-visual test nothing opens cycles: switch it on to let the tester open random entries (they have no edge; the result shows how the manager behaves, not a strategy), or trade from the pad in visual mode.

UPDATING FROM v3.32

With a cycle open, v3.34 picks it up again from your open positions and from then on applies the basket target: if the basket is already above it, it closes as soon as v3.34 manages it. A stopped EA (absolute close or daily halt) stays stopped; to start again after an absolute close, remove the EA from the chart and attach it again.

Calibrated for 10k on gold: how to scale.

The money inputs follow your equity on their own. Lots do not: those you resize by hand.

Scales by itself: money

Trigger, basket target, hard cap, partial threshold and exit cushion are multiplied by equity / 10,000, measured when the EA starts (attach, reload or a change of inputs) and kept between 0.1× and 10×. After a large deposit or withdrawal, reload the EA. The panel footer and the CONFIG SUMMARY show the scale in use.

Scale by hand: lots

InpAvgLot, InpMaxAvgVolume, InpMaxBasketLots, InpPartClose and the pad lot (InpLotSize, InpTestLot) are absolute. Rule: preset lot × your equity / 10,000, rounded down to your broker's lot step. Below the minimum lot the EA uses the minimum, so a 0.01 grid on a 2,000 USD account weighs five times more than on 10k while its dollar caps shrink to a fifth.

Stays the same: percentages

InpAbsoluteClose, the circuit-breaker tiers (InpCBTier1 to InpCBTier3) and InpTakeControlPct are percentages of equity, so they mean the same on any account size. Leave them as they are unless you want tighter limits.

Equity at startMoney scaleGuardian / Control / AW_Zona: trigger · target · cap0.05 presets: trigger · target · capFirst-add lot, hand-scaled (0.01 presets)First-add lot, hand-scaled (0.05 presets)
$2,000×0.2$3 · $5 · $50$10 · $8 · $800.01 (minimum lot)0.01
$5,000×0.5$7.50 · $12.50 · $125$25 · $20 · $2000.01 (minimum lot)0.02
$10,000×1$15 · $25 · $250$50 · $40 · $4000.010.05
$25,000×2.5$37.50 · $62.50 · $625$125 · $100 · $1,0000.020.12
$50,000×5$75 · $125 · $1,250$250 · $200 · $2,0000.050.25
$100,000×10$150 · $250 · $2,500$500 · $400 · $4,0000.100.50

Computed with the EA's own formula from the .set values (money × equity / 10,000) and the lot rule above, for a 0.01 lot step. Control and AW_Zona: scale InpMaxBasketLots (2.0 at 10k) the same way.

Another symbol

The EA multiplies the money inputs by the symbol type (metals 1×, forex and indices 0.1×, crypto 0.01×, other commodities 0.5×) and spaces the grid with that symbol's own D1 ATR. It does not resize lots: 0.01 lot of gold and 0.01 lot of EURUSD move very different amounts of money. Start from Guardian, size the first-add lot so the worst case on the panel is a share of equity you accept, and set InpMaxEquityRiskPct to refuse cycles above it.

Timeframe

The bar clocks (InpStepBars, InpCycleMaxBars, InpCBCooldown, InpControlCooldown) count bars of the chart the EA runs on. Our test runs of these presets used XAUUSD H1, where 24 bars are about one trading day; on M15 the same clocks run four times faster.

Then test it

Run your resized set in the Strategy Tester with real ticks on your own dates, then on a demo account with the lot sizes you will use live. Watch where the caps fire, not the profit line.

Preset questions.

Email soporte@signalforge-ai.com if yours is not here.

Because the money inputs auto-scale to your equity and symbol. This page shows them at 10,000 USD on XAUUSD; at 20,000 USD the Guardian trigger is $30 and its cap $500. The CONFIG SUMMARY prints the input, the scale and the result.
Not as shipped: all seven watch magic 0, which is manual and pad trades. Add your EA's magic number to InpMagicForRecovery, or use "*" for every position on the symbol. Control is the exception: at 2% account drawdown it takes every position on the symbol on its own.
AW_Zona no: ZONE needs opposite positions and the EA switches it off on netting. The others run, but the pair partials (and the lock, which these presets do not use) switch off, and Control takes over without the mirror grid. A hedging account is recommended for all of them.
Only after you add the firm's limits. The presets ship with the daily equity stop off (InpDailyLossPct=0); set it below the firm's daily limit (the input comment suggests 40-60% of it, e.g. 2% for a 5% rule) and check that the absolute close sits inside the firm's maximum drawdown. Many firms also restrict grid and averaging strategies: read their rules first.
The v3.32 files stay online with the same names (SFRecovery_v3.32_...). Their inputs are identical to v3.34; on the v3.34 EA they behave as described here.

Keep reading: the SFRecovery EA product page, every input in the inputs manual, the step-by-step setup guide, the math of a zone recovery EA and our AW Recovery alternative comparison.

Caps first. Then the grid.

SFRecovery EA v3.34 and its seven presets: drawdown recovery for MT5, with the worst case shown before every cycle.

POSITION MANAGER · NO ENTRY SIGNAL · TEST ON DEMO FIRST