Casino GOLD (ICM) (By Hallimax )
| Gain : | +169.49% |
| Drawdown | 78.84% |
| pips: | 13549.0 |
| Trades | 1502 |
| Won: |
|
| Lost: |
|
| Type: | Real |
| Leverage: | 1:500 |
| Trading: | Automated |
Casino GOLD (ICM) Discussion
This is an EA I am still working on and testing slightly different trade trigger mechanisms. I am currently seeking to limit the draw down risks and further testing will highlight if it is feasible or not. I plan to open this for copy trading by next March. It would be interesting for me to see if this is the type of strategy people would be interested in following and any other insights you care to offer. Many thanks for taking the time to respond.
Casino GOLD v2 Update
Version 2 has been redesigned to reduce drawdown severity while maintaining the same profit targets.
Testing results so far:
Maximum observed drawdown: ~20% (significantly improved from v1)Win rate and edge remain consistent with original systemNo total account loss across any version since testing began October 2025Important caveats:
Sample size remains too small for statistical confidenceExtended testing required before live deployment with copiersThe absence of a total loss does not eliminate the possibility - the market simply hasn't presented the specific conditions that would cause one
Risk management:
Daily profit withdrawal remains mandatory - this is the primary protection mechanism for copiersThe system is high-risk by design; the goal is frequent wins while limiting catastrophic loss probability
Why live testing, not backtesting?
Backtests are useful for directional strategies but misleading for systems like this. They don't capture spread widening during volatility, execution slippage, or requotes - all of which directly impact profitability on XAU/USD. A backtest that shows 95% wins means nothing if real-world execution turns 10% of those into losses.
Forward testing on live market conditions - first on demo, now on a real account with real money at risk - provides accurate performance data. It's slower, but it's honest.
Target launch for copy trading: March 2026, subject to continued positive performance.
Hallimax posted:Casino GOLD v2 Update
Version 2 has been redesigned to reduce drawdown severity while maintaining the same profit targets.
Testing results so far:
Maximum observed drawdown: ~20% (significantly improved from v1)Win rate and edge remain consistent with original systemNo total account loss across any version since testing began October 2025Important caveats:
Sample size remains too small for statistical confidenceExtended testing required before live deployment with copiersThe absence of a total loss does not eliminate the possibility - the market simply hasn't presented the specific conditions that would cause one
Risk management:
Daily profit withdrawal remains mandatory - this is the primary protection mechanism for copiersThe system is high-risk by design; the goal is frequent wins while limiting catastrophic loss probability
Why live testing, not backtesting?
Backtests are useful for directional strategies but misleading for systems like this. They don't capture spread widening during volatility, execution slippage, or requotes - all of which directly impact profitability on XAU/USD. A backtest that shows 95% wins means nothing if real-world execution turns 10% of those into losses.
Forward testing on live market conditions - first on demo, now on a real account with real money at risk - provides accurate performance data. It's slower, but it's honest.
Target launch for copy trading: March 2026, subject to continued positive performance.
March 2026 sounds reasonable if the discipline around withdrawals and risk stays as tight as you describe.
Hallimax posted:Casino GOLD v2 Update
Version 2 has been redesigned to reduce drawdown severity while maintaining the same profit targets.
Testing results so far:
Maximum observed drawdown: ~20% (significantly improved from v1)Win rate and edge remain consistent with original systemNo total account loss across any version since testing began October 2025Important caveats:
Sample size remains too small for statistical confidenceExtended testing required before live deployment with copiersThe absence of a total loss does not eliminate the possibility - the market simply hasn't presented the specific conditions that would cause one
Risk management:
Daily profit withdrawal remains mandatory - this is the primary protection mechanism for copiersThe system is high-risk by design; the goal is frequent wins while limiting catastrophic loss probability
Why live testing, not backtesting?
Backtests are useful for directional strategies but misleading for systems like this. They don't capture spread widening during volatility, execution slippage, or requotes - all of which directly impact profitability on XAU/USD. A backtest that shows 95% wins means nothing if real-world execution turns 10% of those into losses.
Forward testing on live market conditions - first on demo, now on a real account with real money at risk - provides accurate performance data. It's slower, but it's honest.
Target launch for copy trading: March 2026, subject to continued positive performance.
This reads disciplined and refreshingly honest. Cutting drawdown while keeping the edge is the real win here, and I like that you’re upfront about sample size and tail risk instead of overselling it.
Update – March 2026
Casino GOLD has been paused.
The short version: the algo ran into a sustained market condition that broke it. Gold moved in a way that the system wasn't designed to handle, and forward testing exposed exactly the kind of failure that backtesting alone didn't surface. That's the whole point of forward testing — and honestly, I'm glad it happened in testing rather than at scale.
I've identified the structural weakness and part of the algo will be rewritten based on what I learned from those failures. I'm not going to rush that process. This has always been a labour of love, and I'd rather take the time to get it right than push something out just to show activity or chase profit.
Backtesting showed promising results, but live markets have a way of humbling you. The conditions that caused the failure were real and the lessons are clear. The core concept still has merit — the execution needs to be smarter.
No timeline on the next version. When it's ready, it's ready. I'll update here when there's something worth showing.
Thanks to everyone who's been following along and reading. Appreciate the interest.
