So I been working on this trading algo for the past few months. Im a data engineer, so coding it wasnt the hard part, but I have basically zero real trading experience. I cant read chart etc. This is genuinely my first time doing anything like this.
The strategy is crafted by GPT 5.5, which is mean reversion, when a stock drops hard, im betting it bounces back soon after. I only buy if the bigger trend is still up, so I'm not catching stocks that are just falling apart, just short dips. To find the dip I use something called RSI(2) method. When its oversold enough, I buy the next morning. I sell when it hits a small profit target, hits a stop loss, or just hasnt moved after a few days (plus a backup stop in case it gaps down overnight). How much I buy depends on how risky the trade looks, so bad setups get less money.
Backtest results (8 years of data, around 11,000 trades):
\-Won about 71% of the trades.
\-The losing trades are bigger than the winning trades on average, but overall it still made a small profit.
\-Average holding time is around 4-5 days.
\-6 of the 8 years made money, while 2 years lost money.
\-There was one period where it lost a decent amount before recovering.
Paper trading (about 7 weeks so far):
\-Around 50 trades.
\-About an 80% win rate so far.
\-The results are actually a bit better than the backtest, but I know 50 trades probably isn't enough to mean much yet.
\-I also had one trade gap down overnight and lose more than I expected.
The thing I'm stuck on is my account size. My account is only in the low four figures. Once i include commissions and trading fees, they eat up a lot of the profits because each trade is pretty small. If I pretend I had a much bigger account, the fees barely matter and the strategy looks much better.
So now I'm wondering if the strategy is actually okay and my account is just too small, or if that's a sign the strategy isn't as good as I think it is.
Queston:
\-Does a 71% win rate sound good, or should I be worried that the losing trades are bigger than the winning ones? It feels like one really bad trade could wipe out a lot of small wins.
\-What are some mistakes beginners usually make when testing a strategy? I'm sure there are things I havent even thought about yet.
\-Is around 50 paper trades still way too early to know if a strategy actually works? How long did you paper trade before you started using real money?