Overdraft

@Abomination81 This is for a bot trading options on the us stock exchange

@Overdraft_Ent Yeah, main problem is polymarket is geo locked. Closest two locations are Montreal and Ireland

Lets fix your BOT! You have problems, lets work on it. To do it I need the following. What market and time youre on, and the basic outline of what youre doing. With what problems you have! Lets get those things printing! #polymarket

@Abombination81 Your tweets over this week has added so much value to all my bots, still massive learning curve but data data data is the goal.

@Overdraft_Ent I'm being honest, if people arent bookmarking that page... well they are not going to make profitable bots

This is the single most valuable post I have made to date. And that is really saying something. I debated posting this... but fuck it. If this isn't the most bookmarked and shared post in history... well Jesus can't help you. You backtested, you lost money, you tuned, you tweaked... lost more. Said F' this and tried something new. Only to see the same thing happen again and again. Most failures are savable, most strategies "work" if tuned right. I'm up to 320 live working strategies, and 200 others that work, I just don't use for various reasons. I'm not a genius, but you start learning a recipe for a good bot. ------------------------------------------------------------ Some facts that make the suck, suck less #1 Regime changes. This is an easy win. Different days do different things. M-F are retail investors, Saturday and Sunday are casuals with time off. Check any bot you have, literally the opposite thing works on weekends that works weekdays most of the time. Or just skipping weekends or weekdays all together can fix it. #2 Time. Same idea. Certain markets behave differently... asia as a whole, sucks. So when certain countries come online, things work completely differently. Many strategies can be brute forced to work by adjusting the times it trades. #3 Negative signals. If you have a bot and its losing money... do the opposite. That simple... oh a mean reversion that is eating shit.. flip to momentum. Oh over x bps, try under etc. If you have a test that shows oh... only a 42% wr.. oh no, how shitty.. guess what, the opposite is a 58% wr. #4 Skip losses. Most losses cluster. Logic like (If I lose two in a row, skip the next 4 windows) can fix alot of low PNL bots. Losses and wins cluster, stopping some losses is big big. #5 Panic Sales. I posted a csv with the % chance something will drop in price. There is a point in every window, Time vs BPS, that it will not flip. You can set a logic to dump anything FAK with 100% slippage to save some of the loss. #6 Latency. Cant fill? Sitting on books unfilled? Heavy slippage? I promise reducing latency to the true minimum will fix it. ALL backtests assume INSTANT fills, so if youre slipping, the backtest is invalid. #8 Same strategy, different coin and time. If it logically makes sense.. try it on everything in the 5m and 15m crypto space... BTC, SOL, ETH, XRP, BNB, DOGE, HYPE all act simmilar but different. Strategies that might not work for stable btc, might do great for psycho hype. #9 Simple. Simple. Simple. If youre hedging, rebalancing, using more than 2 signals, splitting, merging, cross window arb, mutli window consensus... etc.. You fucked up. Ai tricks you into these complicated schemes. NONE of the top bots use that crap. Its all simple brutally well execution. #10 Try not to win, try not to lose. You can tune anything, but its hard if youre eating shit every 5m. Work on your downside. Can you do a gtc stop loss here? Can you fire an arb there? What if we skip windows that have Z signal etc. #11 Passive > Active. Picking the windows you enter will always work better than exiting shitty windows. Buying the right side will always be better than exiting the wrong. GTC bids will always beat out FAK for reliability. The less you have to worry about laser focused execution the better. At this point if you're not following, you don't deserve the sweet mana from the #PolyMarket gods.