So you are not including companies that no longer exist and that you would lose most money on. Yes, but your Excel shows that you can make $billions or $trillions, so automatically you can see that your calculations are incorrect because it is not possible to trade so much money without affecting the price, even when daily volume is $100M. You cannot buy all $100M of shares by yourself and still have $trillions to spend. Also, with $100M volume only $100K-$1M may be traded near the day's lowest price as many traders compete for the same shares. Anyway, for now your results are the same as everyone else's at one point or another. There are thousands of people posting the same results when they are learning. After several months or years you will fix some issues and understand more, but no else can look at your data and algo and find all the problems.
Slippage on market orders can be massive in small stocks (several percent even), and that's why those "inefficiencies" appear to exist in a naïve simulation. Passive limit orders or more advanced orders like peg to midpoint orders will not fill reliably (in fact, you will be the target of adverse selection). The fact that your estimated annuals return is >1000% makes no difference, slippage applies to every single transaction you make. Save yourself additional trouble for nothing and go implement a slippage simulation before doing anything else. You can test actual slippage on a real account, or piggyback the sim of e.g. Interactive Brokers' paper account, or otherwise have a look at the typical buy-ask spread wherever shows that. On a side note, environments with well tested slippage/fee simulations are massive time savers for beginners for this reason. You get to skip the fairytale phase where you dream of becoming a billionaire within a few years due to broken simulations - leaving that phase can take time since it's such an appealing fantasy.
The key issue with backtesting is static data. Depending on the lot size, the live market will react to the bid and ask.
When "psychological anxiety/money pressure" is in your equation, then your system is not completed, and you dont backtest enough, thats why you feel anxiety in your trading, because your system not completed that lead to your fear and distrust to your own system Yes thank you for your advise, i think i am need to recalculate my system again, i should start my calculation from Close candle 0, while it should be on Open candle +1 Can you tell me where i can get data on companies that no longer exist?? Since seems yahoo is not providing it thank you for your advise, what average spreads should i use, 0.5% from Close price is good enough?? since i am using excel to backtest i only can use static spreads
Hindsight is fine because you have the benefit of already seeing what happened. In real life trading, you have to make decisions on what to do in anticipation of what is going to happen in the future, 3 days from now, 30 days from now, 60 days from now, 1 year from now, etc. This is why paper trading for practical purposes is useless. Better to risk small amounts and test your trading system in real life conditions including, slippage which is substantial. Then, check your results based on your real life trades. This is more relevant and reflects what could have happened to your actual trades without the benefit of hindsight.
Do you discard trades when you don't have enough funds to cover all entries? If yes, that is a potential source of peeking into the future. For example dip buyers that rely on daily data. They will randomly pick trades on any given day up to the amount of equity available. Trouble is, during a down trending day with significant declines, you'd get positions that are completely unrealistic.
If you are using daily data (lower frequency data), you need to think whether your strategy is explainable or interpretable.
You can probably include the commission rates and a few other things because the difference between demo and live would be something else. If you want a proper analysis, you should probably include other factors in your assessment. Check if the platform is suitable for a particular strategy like the price action, range trading, scalping, swing trading, etc.