095% partial trading implementation on backtesting side effects - Printable Version +- Gekko Forum (https://forum.gekko.wizb.it) +-- Forum: Gekko (https://forum.gekko.wizb.it/forum-13.html) +--- Forum: Technical Discussion (https://forum.gekko.wizb.it/forum-23.html) +--- Thread: 095% partial trading implementation on backtesting side effects (/thread-57858.html) |
095% partial trading implementation on backtesting side effects - Piff_Sara - 02-10-2019 Hi everybody, on papertrade and livetrade Gekko always buys asset using only the 95% of your total balance. The reason why this is necessary it's clear --> https://github.com/askmike/gekko/issues/2434 The inconsistency I tried to solve regards the backtesting function, beacuse it assumes you always trade the 100% of your balance. So the backtesting "brain" works differently from the papertrade and live trade "brain". The side effect is in the image attached: in this particular case there is an overestimation of the results by the backtesting function. I didn't considered fees for the sake of simplicity. Now we got a slight overestimation of the total profit, but we got just 3 roundtrips... what would happen if we would have thousands of them?? [attachment=341] So i modified the plugins\paperTrade.js file and I put a 0.95 factor at line 83 and 84, in order to make the backtesting "brain" working as the livetrade/papertrade one. My question is: do you think am I messing up/forgetting something else? Code: cost = (1 - this.fee) * this.portfolio.currency * 0.95; Hope this is gonna be helpful also for other people that are not aware of this 0.95 stuff. Thanks! |