I have been working with trade automation since the early 1980s as you will know if you have been a regular reader of this blog.
As the markets changed, full automation became more and more of a challenge. Re-optimizing monthly became weekly and optimizing weekly became optimizing daily. Once it came to optimizing daily the algos became less stable due to the volatility fluctuations and also impossible to share due to the time constraints.
There had to be a better way. And there was.
If you click on the chart above and you enlarge it you will see Flo's fully automated trades today. I was out getting my Covid vaccination and happily could not interfere 😊😊😊 The ELgo that Flo uses is never optimized, ever. Its just trading order flow. The chart used the LogikUltimateRenko bars to get rid of th noise. Renko bars are very difficult to backtest so one needs to be very careful in algo development as you can easily fool yourself that something terrible is profitable.
Here's the Strategy Performance Report for today trading just one contract of the front month NQ.
NinjaTrader now has a more granular strategy reporting system. You can break down the strategy report into much smaller chunks as small as 30 minutes. The pic below shows every hour for today.
Interestingly, it shows that I would have made even more profit had I stopped trading after the first hour and a half of the RTH session. Problem is that you cannot know in advance how the day will unfold. Having said that, if you monitor you algo while it is trading automatically there is quite a lot you can do to know when to stop trading for the day. Markets that open away from the previous day's close create a need for people to do something as the market is out of balance. Using that and Market Profile I make decisions when to stop for the day. On days when I am not there, like today, I am still confident that Flo can handle the day well enough. As time goes by I intend to add that contextual info into Flo so she will be able to stop trading for the day when conditions and context say so.
No comments:
Post a Comment