Overfit. Oversized. Overconfident. Blew accounts so you don't have to. Maker of Seven Sins (free MT5 discipline tool), EUR FORGE (MT5) & EUR EDGE (cTrader).
@NewMoonAlgos Good point. Moving the cutoff until the out-of-sample looks good is just optimising the split. Pick the date before the first test and leave it alone.
A profitable backtest is not validation.
تحقيق أرباح في الاختبار العكسي لا يعني صلاحية الاستراتيجية
Develop → Freeze → Out-of-sample → Other markets → Forward test
Full example on my MQL5 blog ↓
https://t.co/lzROV7KZRj
If I have to keep changing the rules to make every test profitable, I'm probably fitting the past.
#Backtesting #AlgoTrading #Forex
@SJosephBurns The hardest part for me is a losing run, not a single loss. A strategy that wins half its trades can easily lose 6 in a row, and that's exactly when I want to step in.
A better forward test doesn't always mean a better system. Sometimes it just means kinder costs.
Same trade, 10-pip stop, 20-pip target:
No costs → break even at 33.3% wins
1-pip spread → 36.7%
Freeze the costs along with the rules, or the comparison means nothing. Full post in the reply ↓
اختبار مستقبلي أفضل لا يعني دائماً نظاماً أفضل، أحياناً يعني فقط تكاليف أقل
#Backtesting #AlgoTrading #Forex
I built the tool I needed back when I was blowing accounts.
Seven Sins: a free MT5 utility that calls you out on the chart when you break your own rules. It never trades.
On your phone? Save this, open it on your PC.
أداة مجانية لميتاتريدر 5 تنبّهك كلما كسرت قواعدك
@NewMoonAlgos Agreed. Otherwise a better forward result might just mean a kinder cost model, not a better system. Same rules, same costs, or the comparison means nothing.
@SJosephBurns Interesting that strategy only comes at number 6. I spent years doing it the other way round: hunting for a better strategy while my position sizing was the thing blowing the accounts.
Seven Sins Trading Discipline utility tool is designed to assist with your discipline, which is the hardest thing most retail traders face. If you call discipline a strategy then yes it is. Otherwise it’s nit a trading EA/bot.
Have a look at the link attached for more info.
And if you’re after a trading EA , you can have a look at:
https://t.co/2oICluX4Nx
Lust is the sin I built Seven Sins for.
When EUR FORGE has a losing run, I want to step in and close trades early, even though losing runs are normal for it. Now my own tool calls me out when I do: "It was behaving exactly as tested."
Seven Sins is free, and EUR FORGE has a free demo. Both run in MT5 on a PC. Links in the reply ↓
هل تغلق صفقات الروبوت قبل أن يُنهيها بنفسه؟ هذه الخطيئة صنعت الأداة من أجلها
#TradingPsychology #MT5 #Forex
@SJosephBurns Executing. I had my rules written down for years and still broke them in the moment: one more trade after a loss, a stop nudged further away. Finding the edge was never my problem. Following it was.
أداة مجانية جديدة لميتاتريدر 5 (MT5) اسمها Seven Sins (الخطايا السبع).
تراقب حسابك وتنبّهك على الشارت كلما كسرت إحدى قواعدك: صفقة انتقامية بعد خسارة (Revenge trade)، صفقة بدون وقف خسارة (Stop Loss)، الإفراط في التداول (Overtrading)، ملاحقة السعر (Chasing)، أو التدخل في عمل الروبوت (EA).
لا تفتح ولا تغلق أي صفقة، فقط تراقبك.
تعمل على ميتاتريدر 5 على الكمبيوتر فقط، وليس على تطبيق الهاتف.
https://t.co/dKq5k0wtXM
I built the tool I needed back when I was blowing accounts.
Seven Sins: a free MT5 utility (PC) that calls you out on the chart when you break your own rules. Revenge trades, no stop, overtrading, chasing, meddling with your EA.
It never trades. It just watches you.
https://t.co/dKq5k0wtXM
أداة مجانية لميتاتريدر 5 على الكمبيوتر، تنبّهك على الشارت كلما كسرت قواعدك
#TradingPsychology #MT5 #Forex
Depends on holding time. On short-term trades, news-day spreads and slippage can decide the result. On swing trades it matters much less: one release is just noise over a multi-day hold. The one thing I still check is whether the stop sits close enough for a single news spike to take it out.
@mql5com Faster walk-forward runs are welcome. Worth checking in the results: which settings did each window pick? If the best stop is around 20–25 pips every time, that's a stable sweet spot. If it jumps 12 → 40 → 18 → 55, each window is probably just fitting its own noise.