how to code flexible/adaptable strategies?

Last post deleted because I did not join?

I have been trying to code my day trading activities. I am a pretty profitable manual day trader. I usually scalp options. However, I do not have a rigid set of rules to go from.

I usually just look at level 2 data/tick or 1 min charts and hypotheses just pop into my mind. If I can validate the hypotheses in multiple instances(sometimes over as short as 20 mins period), I just act on it. Maybe I am just a degenerate who always see new patterns and was consistently lucky.

ie I see intermittent and consistent 100 buy market orders from exchange ADF that take out the closest ask, then when a bunch of sell market orders smashes multiple bids in the market depth, you will see buy market orders from ADF that smash many asks levels.

My hypothesis: Jane street/JPM are accumulating and keeping the price here. I sell a weekly CSP at the highest volume/OI strike and pick up some calls when large sell market orders hit. (I need to maker sure the option prices are reactive to underlying).


ie I see movements in a stock that always penetrates clear supports/resistances by 10 cents and stay for a couple of seconds and then bounce the other direction and the option market makers all cancel their bids/asks to avoid supposed toxic fills.

Hypothesis: these market makers are getting baited and I place limit orders that match MM's and do not cancel.

The point is that these strategies seem to come to me on the fly and needs only very few examples to convince me. I am almost certain that they do not generalize well in particular stocks nor over long time periods. Is it possible to code something that is as flexible as my execution? I have tried multiple strategies ie I have coded many strategies templates that if whichever validated the best for a day, use it the next day.

It just does not seem to produce consistent profits in paper trade nor small live test as I manually do them.

feel free to use the above two examples. I am pretty sure they are worthless over long periods Let me know though šŸ™‚

Background: I am an infra developer at a Bay Area Startup. I am familiar with Machine learning and finance in general, having won machine learning/DL competitions and some algo trading contests in conferences and uni.

Submitted October 20, 2020 at 01:37PM by Actual_Revolution_93

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s