Research

What we tested, and what we found

Nothing is sold until a strategy holds up on recorded prices and then on a demo account. This is the record so far, losses included.

See what the bot is scanning now

Recorded 28 September 2026. A test on past prices: not a track record, and not a forecast.

Conclusion

On gold, the strategy lost money on every timeframe tested, even before costs, and its setups went the right way no more often than a coin toss. A second idea, the London-open breakout, did not hold up either. Until a strategy holds up on recorded prices and then on a demo account, Vantrex offers no gold signal service or EA built on it.

How it was tested

Market
Gold (OKX XAU-USDT-SWAP perpetual)
Prices
OKX public market data, listed 9 April 2025
Method
The bot's own setup pipeline replayed one closed bar at a time, with its reward:risk gate of 2.0, before the AI review and without session quotas.
Costs
0.05% fee plus 0.05% slippage per side

Replay results

15m
3 Mar – 28 Sep 2026
Trades
26
Per trade, after costs
-2.68R
Per trade, before costs
-0.03R
Profit factor
0.03
1H
16 May 2025 – 28 Sep 2026
Trades
56
Per trade, after costs
-2.04R
Per trade, before costs
-0.11R
Profit factor
0.09
4H
16 May 2025 – 28 Sep 2026
Trades
24
Per trade, after costs
-1.87R
Per trade, before costs
-0.25R
Profit factor
0.15

Did the setups move their own way?

All 224 hourly setups, measured from the entry. Near 50% means no edge.

After 1 hour
45%
went the right way
After 4 hours
47%
went the right way
After 12 hours
47%
went the right way

The market-gap sequence

A liquidity sweep, a shift in structure, then an entry at an open fair-value gap: 136 signals on the hourly chart. The rules were fixed before the data was examined, and the prices from 13 Apr 2026 onward were kept aside until the end. On those, entries at the gap went the right way in their first hour 37% of the time.

A different idea: the London-open breakout

Each weekday, the first 15-minute close outside the Asia range (00:00 to 07:00 UTC) between 07:00 and 10:00 UTC, with the stop at the middle of the range and the target at twice the risk.

First 11 months
16 May 2025 – 13 Apr 2026
Trades
107
Per trade, before costs
+0.04R
Per trade, after costs
-0.44R
Last 5 months, kept aside
14 Apr – 24 Sep 2026
Trades
55
Per trade, before costs
+0.41R
Per trade, after costs
+0.05R

It worked only in the later period, and there it barely covered costs, so it did not pass the test set before it ran.