To measure the performance of program trading system, there must be a virtual trading report (also called “paper trading”), showing each trading result and overall monthly (or seasonal) performance. Thus, an initial algotrading report starts in August, 2022.
To make it simple, there must be a few assumptions:
- initial invest amount is set to $200,000 with 0 stock on hold on first day.
- each stock is trade as $10,000. That means, the system can maximum hold 20 stocks. If there are more buy signals, the system (currently is human decision) will suggest to terminate the worst performance transaction.
- transaction is 1 buy 1 sell. That is, sell 100% if sell signal occurs.
- transaction is buy or sell on next day open, when signal occurs. Transaction period minimum is 1 day.
- due to bear market in Hong Kong stock market, cut loss is set to global 7.5% (each stock can set to different cut loss value). No cut profit is set (it is still working on).
- All transaction cost is set to fixed 0.75%. To make it simple, it is set to -$75 after complete sell transaction.
- To make it more interesting, I will manually add columns for “my decision(Y/N)” – decision to buy on next day or not. If yes, then follow the buy signal to buy on next day. Otherwise, this signal is skip until next transaction.
- the monthly performance should compare to stock-market index.
- As of current, there are only 111 HK stocks are included in the algotrading system. Why, because not all stocks are ok to give satisfied result. If result is not good enough, the stock should not put in the algo system. User can review the criteria on designing algo trading system here.
Result will be post on every weekend and month end. All algorithms are subject to change without notice. After 3 or 4 months, the system should show if it is good enough or not. Currently, the system is incomplete and missing risk management. Please read the disclaimer and terms and conditions.