No trades Friday; trying a similar program today (buying 2554.5 given price hits 2552.5 first, without hitting 2545)
No entries, and I'm very happy about that! Even though the price went down to where I wanted to buy, there was no support, so the algorithm didn't buy. Eliminating entries like this makes a huge difference in long-term profitability. My goal for this week is to get the breakout-then-pullback algorithm going.
No trades yesterday (almost...) Trying to buy support around 2547 today. I'm finishing up my test code for breakouts and then it's on to implementing multiple entry possibilities.
I'm selling any resistance 2550-2560 today. No trades yesterday. Breakout code has passed all the tests; now on to implementing multiple possible entries. I'll be averaging several trades a day once I get that in.
I'll be buying any support in the 2565-2570 area today. I'm about 1/3 finished implementing multiple entry possibilities.
It's been an interesting week. There was a missed trade both Monday and this morning, where local low was exactly the price the algo was waiting for. However, since the algo polls the market only every so often, it missed the ticks and never entered. My first thought was, well, there's always going to be a boundary where you'll just miss a trade for one reason or another. I still think that's a valid point. However, I also think a better design is to have the prices *inform the trading algo* when the level of interest is touched. In other words, use callbacks instead of polling.
Last week was all testing and coding. I got a bit carried away but this next update should take me pretty far: No more price polling-- prices notify trading programs when the market gets to places of interest can now trade breakouts multiple trade possibilities instead of just one idea per day Running on the simulator this week just to make sure the program operates properly. Back to real $ on Monday.
Despite having passed all the unit tests and integration tests, the new program keeps finding support prematurely. So, I'm collecting NQ ticks right now to use for more realistic, automated tests-- basically an offline market simulation. I have my code base designed for such a feature so I should be done with it relatively quickly.