Three configurations, and which one to run
All three are the same script; only input values differ. The Default is what the script loads with. Prop Safe and XAUUSD are Load Parameters files, copied from their tabs below and pasted into TradingView’s settings dialog through Load parameters.
The first two are Micro Gold futures. XAUUSD is a different instrument — spot gold on a CFD feed — and is not comparable to the other two row for row: it runs at quantity 10 against a $1-per-ounce point value, so its dollar figures are a different size of position, not a better strategy.
| configuration | shape | last 365 days (TradingView) | run it when |
|---|---|---|---|
| Default | everything described under What changed below | 523 trades, 62.9% win, +$20,558, PF 1.64 |
a normal account: the most trades and the most profit |
| Prop Safe | the Default with three filters tightened: anti-liquidation 2.0 × ATR (from 3.5), maximum VWAP extension 2.2 standard deviations (from 2.7), short RSI reset 56 (from 39) | 270 trades, 68.2% win, +$12,658, PF 1.96 |
a funded evaluation, a prop account or a small account: higher win rate and less drawdown, for about half the profit |
| XAUUSD | a different instrument, not a third tuning — see Spot gold below | 447 trades, 67.6% win, +$24,105, PF 1.62 |
you trade spot gold on a CFD broker rather than Micro Gold futures |
What Prop Safe gives up, and what it buys. It skips entries straight after
an oversized candle, entries when price is already stretched far from VWAP, and
shorts that have not had a real bounce first. That removes about half the
trades. The ones that remain win more often and lose less: on the tuning window
below it is 81% winners against 75%, and over the full history since January
2024 its maximum drawdown is $1,140 against the Default’s $1,714, with ten
of eleven quarters positive against eight.
It is not the Default with shorts switched off. It still takes about one short for every three longs, and a shorts-off version measured worse on every window.
What changed since v2.2.4
Two new settings and nine moved defaults. No input was renamed or removed, so every saved v2.2.x configuration still loads.
The directional time gate
A new panel stops one side taking new entries for the late part of a session while the other side trades normally. A position already open is never touched — it still runs to its own target, stop or end-of-day close.
Stop taking LONG entries later in the session ships on, from 12:45 PM ET. Stop taking SHORT entries later in the session ships off. Each has its own time, and the window runs to the end of that session’s own trading window, so on an overnight session it cannot spill into the next evening’s open.
Gold tends to slump or sell off into the New York close, and late longs are the trades that suffer for it. Blocking late shorts looks appealing per trade — they are the weaker side after lunch — but it was worse for the account: it removed profit and increased drawdown, because those trades were offsetting losses elsewhere. Both switches ship so either can be tested; only the long one is on.
The 12:45 setting is not a fitted number. Every cut-off from 10:30 AM to 2:45 PM produces a profit factor of 1.41–1.42 over the full test period.
The nine moved defaults
| setting | v2.2.4 | v2.2.5 |
|---|---|---|
| Min VWAP Distance filter | off | on, 0.35 points |
| RSI Reset short threshold | 34.5 | 39 |
| Long take-profit | 0.52% | 0.375% |
| Breakeven | on | off, both sides |
| Long trail distance | 0.25% | 0.285% |
| Short trail distance | 0.35% | 0.385% |
| Short trail trigger | 0.35% | 0.365% |
Breakeven comes off and the trail, set slightly wider, does the protecting on its own. The long target comes in so winners are banked sooner. And the short entry’s RSI reset threshold finally does something — at 34.5 that condition was satisfied on 95.7% of session bars, which meant the short side was entering on little more than RSI below 50.
One visible consequence: the Flatten Breakeven exit label no longer appears in the trade list.
Spot gold
The XAUUSD tab is a different instrument, not a third tuning of the Micro Gold config. Spot gold and Micro Gold futures share an underlying but not a tick, a volume profile or a session structure — and because VWAP is volume-weighted, the same price series with a different volume series is a different indicator. The Micro Gold values do not transfer, and this configuration was built from scratch against a real spot-gold chart.
It starts from a v2.2.5 gold configuration that was already running on an OANDA chart, and changes eleven things. Three of them are switches coming off:
- the MAE exit, which was closing trades that went on to recover
- the long trail, which was cutting long winners before their target
- the long time gate, which was blocking longs after 11:00 AM ET
The rest give the remaining trades more room: the long bracket widens to 0.75% / 0.775%, the short target tightens to 0.445% while its stop widens to 0.60%, the short trail triggers much later and follows much looser (0.49% / 0.485%), and the MFE exit moves out to 0.45%.
The largest single change is the last one. The minimum VWAP slope rises from 1.45 to 2.25, which stops the strategy taking trades while VWAP is essentially flat — the condition it has no edge in. That one setting is worth more than every exit change combined, and it holds across a wide band: every value from 2.00 to 2.50 beats 1.45 on both holdout windows, so it is a plateau rather than a lucky number.
Against the configuration it was derived from, on the same chart and the same 365 days:
| parent config | XAUUSD | ||
|---|---|---|---|
| net | +$19,203 |
+$24,105 |
+25.5% |
| maximum drawdown | $2,729 |
$1,688 |
−38% |
| profit factor | 1.492 | 1.622 | |
| win rate | 62.04% | 67.56% | |
| trades | 598 | 447 | −25% |
| net ÷ drawdown | 7.0× | 14.3× |
More money on a quarter fewer trades, with substantially less drawdown.
What was tried and rejected, because a list of wins without the losses is not evidence: the MACD settings and both RSI blocks were swept exhaustively and left exactly as they were. A faster MACD improved both holdout windows while every other setting sat at its original value — and made things worse once the exits were unconstrained, which is why it is not in here. A looser overbought level produced eleven positive quarters out of eleven and a deep-history profit factor of 1.78, on 363 trades against 788; a perfect record on a halved sample is the least durable result in a study, not the most.
Results
Micro Gold, 5-minute, 1 contract, a $50,000 account, $1.20 per side and 1 tick of slippage. Every row is a TradingView trade-list export: the Default’s taken on 21 September 2026, Prop Safe’s on 22 September. The selection window is 7 June 2026 onward — the range this configuration was tuned on, so read it as the best case, not as what to expect.
The XAUUSD rows are a different instrument and a different cost basis, and they are not comparable to the two above: spot gold on the OANDA CFD feed, 5-minute, quantity 10 against a $1-per-ounce point value, the same $50,000 account, $0.025 per contract per side — the real Skilling/TradeLocker fee — and 1 tick of slippage. Spread is not modelled, which at this size would cost roughly $3.60 a round turn; set your own broker’s in Properties before reading anything off it. Exported 24 September 2026, selection window 14 June 2026 onward — its own chart range, a week later than the Micro Gold runs.
| configuration · window | trades | net | win | profit factor | drawdown |
|---|---|---|---|---|---|
| Default · last 365 days | 523 | +$20,558 |
62.9% | 1.64 | $1,665 |
| Default · selection window | 159 | +$12,898 |
75.5% | 3.21 | $517 |
| Prop Safe · last 365 days | 270 | +$12,658 |
68.2% | 1.96 | $1,057 |
| Prop Safe · selection window | 80 | +$6,850 |
81.3% | 4.58 | $213 |
| XAUUSD · last 365 days | 447 | +$24,105 |
67.6% | 1.62 | $1,471 |
| XAUUSD · selection window | 129 | +$10,306 |
74.4% | 2.30 | $987 |
Drawdown here is measured trade to trade. TradingView’s own figure also counts
the swing inside an open trade, so it reads a little higher — $1,174 and $360
for the two Prop Safe rows, $1,688 and $1,059 for the two XAUUSD ones.
If you open the script on TradingView today these will not match to the dollar, and they are not meant to. The export is a snapshot; your chart keeps trading past it, and your date range, broker costs and data feed are your own. Treat every number here as the shape of the thing, not a quote.
Read the holdouts
The table above is the window the configuration was chosen on. These two are not:
| window | Default PF | Default net | Prop Safe PF | Prop Safe net | note |
|---|---|---|---|---|---|
| Jun 2025 – May 2026 | 1.28 | +$7,969 |
1.63 | +$7,130 |
the year before the export |
| Jan 2024 – May 2025 | 1.02 | +$329 |
1.21 | +$1,288 |
v2.2.4 was 0.97 here, and lost money |
XAUUSD, measured on a different feed and stated separately for that reason. TradingView depth-caps the OANDA 5-minute series at about 101 days, so there is no way to run a holdout on the chart’s own data; these come from Dukascopy spot gold through the engine twin, which orders configurations reliably but does not quote what the chart would have made — it puts only 68% of its entries on the same bar. On that feed, against the configuration it was derived from:
| window | parent PF | parent net | XAUUSD PF | XAUUSD net |
|---|---|---|---|---|
| Jun 2025 – Jun 2026 | 1.32 | +$10,535 |
1.61 | +$18,782 |
| Jan 2024 – Jun 2025 | 1.22 | +$2,463 |
1.65 | +$4,789 |
Over the full January 2024 to September 2026 span it is 616 trades,
+$29,277, profit factor 1.63, maximum drawdown $1,490, and eleven of
eleven quarters positive with a worst quarter of +$22 — against the parent’s
nine of eleven and −$1,573.
Across the full 1 January 2024 to 21 September 2026 span: 904 trades,
+$20,765, 58.7% win, profit factor 1.42, maximum drawdown $1,714, eight of
eleven quarters positive, worst quarter −$850. The same span on v2.2.4 was
1,004 trades, +$18,786, 61.7% win, profit factor 1.34, drawdown $2,226, six
of eleven quarters, worst quarter −$1,084. Prop Safe over the same span:
463 trades, +$14,998, 63.7% win, profit factor 1.75, drawdown $1,140, ten of
eleven quarters positive, worst quarter −$310.
The trade-off, stated plainly: win rate falls from 61.7% to 58.7%, and it trades about a hundred fewer times a year. Breakeven is what used to turn a reversing trade into a scratch; without it those become small losses instead. Profit factor, net, drawdown and quarterly consistency all improve — win rate is the price, and it is a real one. Size by the holdout figures, not by the selection window.
Verification
This release is a paired export: the TradingView trade list and the Save
Parameters file it was generated under, so the settings under test are the
settings that were live. The Python engine twin reproduces the fills at 95.4%
entry-by-entry recall over the Default’s 523 trades and 95.2% over Prop
Safe’s 270 (98% of matched trades exit on the same bar). The XAUUSD
configuration reproduces at 100% — 122 of 122 entries, with matching
precision and a $25 difference in net profit across $9,397; a CFD feed has
no continuous-contract roll to disagree about. The residual is the continuous-contract
data difference described on the
v1.5.0 page, not an engine gap.
Known limits
- The tuned window is the last hundred days. The Default was selected on the 7 June 2026 onward range, and the holdout table is the honest statement of what to expect beyond it.
- Two instruments, tuned separately. Micro Gold and spot gold. Percent brackets travel; the anti-liquidation and candle-range filters are in ATR and travel too; the session hours, the time-gate cut-off and anything volume-weighted do not — which is why XAUUSD is its own configuration rather than the Micro Gold one pointed at a different symbol.
- XAUUSD’s holdouts are on a substitute feed. TradingView will not serve more than about 101 days of the OANDA 5-minute series, so the out-of-sample figures come from Dukascopy and share only 68% of their entry bars with the chart. The 365-day and selection-window rows are the real chart; the holdout table is a proxy.
- Spread is not modelled on XAUUSD. At quantity 10 it would cost roughly
$3.60a round turn against the$0.50of commission that is modelled. - The holdouts are engine figures. The twin is verified at 95.4% entry-by-entry, but only the year and selection tables are TradingView runs.
- Prop Safe was chosen with the older windows in view. Its three settings were picked because they beat the Default on the selection window and on both holdout windows at once, so those holdouts are no longer untouched for it; the forward test starts now.
- Split Sessions is not on this page. It remains on the v2.2.4 page at its v2.2.4 settings, and its Load Parameters file predates the time gate — read the note there before loading it onto this version.