finished building and backtesting an ea that looks solid on history. backtested on 5 years of tick data, good results across different market conditions. now i'm forward testing on demo.
question: how long is a statistically meaningful forward test period? some people say 3 months, some say you need a full year. at what point is the forward test evidence actually useful?
can confirm the unwelcome backtest from experience, i counted my go-to pattern and the real win rate was nowhere near my memory of it, though still slightly positive when paired with trend context. that count changed me from trading the pattern alone to using it only in alignment with the higher-timeframe direction, which is where its modest edge actually lived. counting didnt kill the pattern, it located the narrow condition where it was real.
IG provides daily market analysis,technical setups, economic event coverage, and educational content. its not 'advice' in a legal sense but rather research output that helps you contextualize markets. they have decent analysts on staff. compare to most brokers which only offer marketing material disguised as analysis.
log per trade, review weekly for tactical patterns, review monthly for the bigger picture. the per-trade entry captures the raw data while its fresh. the weekly review catches recurring mistakes early. the monthly review is where you see the real patterns, which setups actually make money and which youve been trading out of boredom. all three rhythms do different jobs.
fixed fields over prose, completely agree. and record the trades you did NOT take but considered, with why you passed. the skipped trades are half the lessons, especially the good ones you talked yourself out of and the bad ones you wisely avoided. a journal of only taken trades misses your decision quality.
one more beginner-safe habit: deposit a small amount first and do a small test withdrawal early, before you have any real balance built up. confirming you can actually get money out while the stakes are tiny is the cheapest trust test there is. do it in week one.
yes, search the regulators name directly, for example the fca register or asic connect, go to the official governmment or regulator domain, and search the brokers legal entity name or license number there. confirm the status is active and the activities listed match what the broker does. if you cant find them on the regulator they claim, thats your answer, walk away.
classic SR uses prior swing highs/lows.supply/demand zones look at the base before a strong impulse (a consolidation that broke out hard). theory is that level had unfilled orders that may still react when price returns. in practice they overlap heavily and the distinction is less important than the concept of 'price reactive levels'.
+1. alpha and fundednext are the safest mid-tier props i've used. fundednext has stricter rules but ive been paid every cycle. alpha is more flexible. blueguardian is newer but encouraging signs so far.
+1. transition periods are where to watch closely, not where to immediately leave. fxpro has handled the transition cleanly from the outside. if they start nickel and diming with fees in 2026 that would be the signal to reconsider.
