Thanks everyone for the replies.
As for the comments about plan,
LF, yes got that covered. Actually it was the seminal portion of the work. It is well integrated for scaling, which is central to the effort.
SML, I had Expectancy, Average Risk, Average Reward and I understand them mathematically, but am having a harder time translating them FULLY (drilling through) to actionable changes in trading. It seems to optimize them, requires fully modifying the plan versus tweaking it slightly. Got the drawdown too. Just working on making it useful, more a priori versus a posteriori. Could you clarify MEA, and MFE? I think I know but want to be clear. (MFE= Max Favorable Excursion and Maximum Adverse Excursion? MAE not MEA) Lastly, I will do a more full plan evaluation at some point, but my sense is I am close on this plan-method, so this is an effort to validate the plan, test it for robustness and limits before it gets stressed by scaling and slippage.
DD, and for all the comments about more discretionary metrics, e.g. following plans, I feel pretty good about that part. I created a two level semantic model that links directly into the trade facts of the matter, then drove that into metrics that give session feedback after each trade, so close to real-time during the day. Then made is dead simple to interpret into trading behavior. I have an three "meters" Tilt, Cash and Optimal. Between them all I can manage the discretionary part during the day easily and automatically.
EM, not really. Probably compared to most retail traders, it looks rock solid, but it is not algorithmic. I am more a Heisenberg principle person: The more precise you are, the less you are closer to reality in regard to elemental facts of the matter. On the other hand, all the more eloquent systems, in all disciplines, allow for "fraying at the edges" as new truths are learned. ( this is more Quine, "two dogmas of empiricism" stuff). Also I know the bots and the algo rule, and the plan is premised around them. Lastly, big picture wise, I am trying a more AI model without the A part this time around because I know I can adapt to a Bot-Algo over time faster then remodeling when they change. I traded ES from 1999-2012, through the spoofing, and birth of the bots and the higher volume. I left when it seemed everyone went to FOREX. Besides my system was only BE after comms. I consider that time learning phase.
But to be honest everyone, I am trying to stick to the goal. Consistently over time (low and high VIX) make modest amounts of money per contract, with acceptable risk and effort, and enjoy the journey. Thus I don't want to go off into the metric-weeds too much, but still want to be diligent wrt a continuous improvement model.
Thanks everyone, constructive comments.
