I'm using the strategy of the Brin line plus the stop-gap line, which is a very high return on historical year data, but the real disk is losing money; and switching cycles, the results of the re-measurement will change a lot, is it because the strategy is deeply adapted to the past data?
This world is my only one.But I believe in myself. There is a return to the strategy. There will be a return to the strategy if the yield is high.
This world is my only one.Like me, I'm like you now, and I've been losing money for the last month.
[Translated from Chinese]In a nutshell, the parameters that are being tuned are over-fitting, and if your feedback and the real-world results are not too far apart, you may be making the mistake of using the past to determine the future, so let's say you're an alien and you come to Earth and you see people wearing shorts, and you're ready to come back to Earth in shorts for four months and then you find it's cold, and that's about how it feels. If there is a large discrepancy between the retest and the real-world results for the same period, this may be better solved by details such as delays, slippage, etc.
The baby dinosaurI'm going to suggest to Martin that you don't choose what to use, what to stop and stop.
Don 。。。。。。。