I just ran 3 live tests with 100% success rate and made 45% profit. Yes there must be luck involved. I just wanted to share a progress update and good news on my first live test.tl;dr; I made a AI strategy pokemon evolution type evolve thingie , that tests strategies and then ranks them, then applies them to solana dexscreener tokens. This is a progress update thread as Im nearly a first stable prototype. Timeline:1 week: Program applies AI strategies ( hardcoded list generated by AI) against live tokens, logs successful simulated trades.2nd week: Create a report function so I can see which strategies are clearly working. Move the program to a VM so I can get to run in an automated scheduled environment. ( And not piss me off ).3rd Week:Connected it to Discord, so everyone can monitor all the testing updates live for free (h6DgNf2H)Added private channel which is simply the filtered "winning strategies" 4th Week: Tons of bugs, fixing these. Unstable. Runs paused. Added details for tokens ebgin traded (IE link to charts / twitter / etc). Stabilised code that looks up token on the API. Aimign to fix these , leave it running for a couple days. And presto. Status now: First live test. 3 trades, recommended by the AI. 10 min HODL. Total risked $230. Total after: $325. 45% increase. It should hold for amounts I assume as high as 5k trades, as liquidity for these tokens is far above 10k. Thats all for now, just wanted to share my progress and thanks to everyone who has supported me. also thanks for the (yous) and for listening to this autist.
This is what the strategy tests look like
And here are the better strats
>>58430320THe first trade on each was a test gone wrong, bu t the other percentages apparently!?! are real ( i fixed the bug) I dunno lol, but its the same strat I used tonight on my live test.
>>58430323Anon this is all very interesting. Could you not use the AI to also help you make decisions about the most effective strategies over time? You could have it continue testing all strategies it comes up with and start to surface "winners" for you to evaluate.
