Backtesting is inaccurate when the test includes any drawdown.#527

I found what I consider to be a backtesting issue that critically undermines its accuracy. For Eg, DCA Short bots, when I give the bot 500USD to work with, and include a stop loss… If the backtesting ever encounters an actual loss and goes into drawdown (negative profit - indicated as red) then the backtesting will continue to assume funds for new trades that it technically would not have available in real life, for eg, if the funds available drop below 500USD, to say, 400USD it will keep backtesting as if it had 500USD to invest even after stop loss, this is incredibly inaccurate.

I noticed this for a long time so I try to work around this, but it would be really great if you could fix this on your end to make sure the backtesting accounts for actual available funds.

If I have 500USD and stop loss and have 400USD then it should do further backtesting investments with only 400USD for new trades and so on, not assuming I have unlimited money, since in real life, and if I run the bot with the demo mode, it will correctly invest 400USD as that would be the new bot investment total…

I noticed the backtesting does behave properly so long as there is + profit… so if there are any stop loss events, say if the profit is 500USD + 100USD, then stop loss back to 550USD total, then it will invest as normal and expected with the correct amount of funds assumed.

So there is a disparity here in the backtesting which makes zero sense, and does not apply to real bot behaviour at all.

Please consider looking into this and adjusting the logic so it will respect the actual available funds after trade loss events if the investment funds go into the negative from their starting point.

Also adding another column showing total investment value (investment + profit tally) would be very helpful not only the profit column.