what a ■■■ he takes his eggs out and makes the box unbalanced
Can wire multiple vpns through cli. Otherwise youre stuck with shitty desktop apps that work per machine.
That’s cool. I think I’m gonna get some tofu and make miso soup.
1 Like
The guy who raped me and the grad student have teamed up. Awesome.
good jokes m8 real funny see u at fuck youj
1 Like
I am done with my robot trading fleet. 5 bots on 5 futures symbols, all doing high trades so I expect to be able to determine if the backtest are bullshit or we have found real statistical edge
It's honestly been so long since I felt like I had a real win so it would be awesome if over the next 2 months the numbers actually line up and it performs well and I hammer out all of the potential execution so I can try it live.
Each model averages anywhere between 5-30k a year so we're looking at a range of maybe 15-50k as income over 12 months and if that actually happens I will be blown away.
However I have a feeling it won't happen. Lol. There's always some shit I missed or fucked the models so we need to do it all over again, but I haven't felt this excited since I first went live in IBKR when all I had at that time were my backtest showing I was making like 1 million a year during the period of time my datasets had leakage lol those backtests were honestly mouthwatering tho but now I can smell the bullshit a mile away 
Your effort is the win
I just made 600 telling that guy to play slots
I have been running my bot for a week and it's down about 5 grand. Good stuff. I'm already in the process of building some new models and testing out new theories.
I was thinking about trying to setup a bot to be a kalshi / polymarket market maker, recently trading the shipping bets for kalshi on the lower volume bets made me realize how absurdly thin the books are so it seems like a nice way to just spread a bunch of scalping orders across the thousands of lines to try to come up a few cents every order.
Anyway, once I have finished this last round and launched it on the cloud as a stable build I think I am going to either start applying for work again and getting 0 call backs, or just work on my kalshi bot because I am going to just work on what I want if I can't get paid to do it.
1 Like
I believe in you ironstove!
your partner is over there saying she can’t play ff14 with us for a sub cost and you’re running social media sales algorithms? bro
please right this wrong
If you supported women, you would have paid it for her.
1 Like
if i were trying to game a trading system and marketing algorithms i would have paid for it already
Translation she cannot afford to, but sees evidence that you can smdh.
I'd buy my girlfriend subscriptions to good games only.
2 Likes
The trading bot is back even again, traded itself out of a hole, this is pretty entertaining, it is making all of these moves only to end up exactly where it started.
I did several more experiments to try to find some better models, I added 30 new calendar spread features surprisingly did almost nothing for the edge, so I removed it because adding all of the calendar spread API calls for each symbol was going to be a pain in the ass with apparently 0 added benefit.
So right now I am kind of tapped out of ideas on what to change, but since all of my backtests are showing positive ROI over the long term of 3-6 months, with a high frequency of trades all I can do now is just stop trying to touch it and let the program cook because it will prove once and for all if this thing works.
I think signals can be weak, but what has been a bigger game changer between models producing similar signals between a positive and negative return is the execution engine, logic like deciding where to set the TP/SL orders, trailing stop triggers and tiers, and deciding how many consecutive signals to wait before firing an order, or how long to sit out of the market when you get stopped out of a model, these are all decisions that can be modified and tweaked and each symbol/regime has different values which trigger different returns and behaviors over long periods and holdouts.
Surprisingly this is almost like half ML and half building a disciplined trader as well, who is constantly placing and tracking orders. I know I get lazy sometimes and don't immediately place a stop loss when I was trading or a take profit, the bot systematically does it all for me now, and we will see over the next few months if this stable version turns out to be better than me, if it doesn't lose money it's already better than me XD
The goal of course is to try to get a stable version running with some finalized models which I will then plug back into my cloud deployment to run headless so I don't need to monitor it locally anymore and can move onto my next models which will be kalshi/polymarket market making (sounds like that shit is rigged though from my research, we will have to see how much information I can get for free from kalshi/polymarket, if you get level 1/2 order books you might be able to potentially build a system that can try to detect 'suspicious' insider trading moves as well to pull out all spread orders to avoid getting sharked.
What ever it was doing to get back out of the red make it repeat?
literally just gambling with extra steps
1 Like