my quant friend explained how polymarket AI bots actually work btc/eth/sol/xrp 5-min markets now live they use black-scholes for binary options C_binary = e^(−rT) · N(d₂) for 5-min contracts: model says 50% probability (coin flip) but bots print hundreds of thousands how: binance confirms 0.15% move at minute 3... true probability jumps to 92% polymarket still shows 65% bot captures 27% edge before odds update the pipeline: ▸ deribit implied volatility ▸ monte carlo 200k simulations ▸ detect 3%+ model divergence ▸ execute on lag they're not predicting direction they're arbitraging information delay polymarket launching 1-min markets signal-to-noise drops 2.24x only sub-100ms systems survive when fair value updates in milliseconds who controls the orderbook? forget about losing with Poly Bot Copy ⭣show more

do you ever feel bad for the retards you scam selling copy trading bots for the short term crypto markets? you are a bug person

people do whatever they want got a solid product, i'm pushing it

if it was a solid product you'd just print. you're selling it for pennies to retards because it doesn't work. use your profits to purchase rope

"monte carlo 200k simulations" Most people give up their imaginary friends in childhood.

@NotthemPR my good friend

none of this matters . u either know this && are scammy or don't know this && are smoothbrained . which is it ?

why does it not matter bro?

what u outlined is totally trivial alpha-less slop ... the most potent edge to be had is in infra :: speed / low to zero latency data feeds && lightning quick processing && decision making algos .

AI slop, polymarket 5m markets dont use Binance prices, they use chainlink (Binance + other exchanges). Dont fall for this crap

anyone who mms this and blindly plugs in deribit vols to n(d2) is a moron and is guaranteed to lose money. (i am one of the larger mms for these products.) you’re a POS for pushing slop like this

As if this edge is still there

definitely exists

that's a pretty simplistic model for a complex market, using black-scholes for binary options in such a short timeframe seems like it would introduce a lot of unnecessary friction. the fact that bots are still able to print hundreds of thousands suggests there are some underly...

yeah you're right, it's super simplified

Fake. No bot catches a 27% edge before updates. Updates are Immediate, at most 500ms and thats SLOW.

The real edge isn't the model—it's the volatility smile. Retail prices 50/50 at 50¢, bots arbitrage the skew. Same formula, different data feed.

kinda yeah / partly true

this is the post that tell me you don’t trade micro-timeframes without telling me.

what interest rate are you getting over 5 minutes?

Quant friend just exposed the whole game

nice to have a quant friend

Fucking prediction markets messing even bitcoin price action. Incredible.

they're the ones running this shit

too complicated bros but cool...

tough but interesting

Practically N(d2) is just random shit which will make your bot worse

why?

Hit me on telegram im working on bots rn have tons og those some of them are actually working The inly problem is model I have secured latency issues and stuff What im confused atm is buying both legs for arb

coming from someone at v7 of their bot and probably all your deam features implimented lol what i need to know is arb behavuor for bots and we roll are they buying both legs at a same time or delayed thats it cuz N(d2) hits here

C_binary = e^(−rT) · N(d₂). black-scholes for binary options ▸ detect 3%+ model divergence ▸ execute on lag binance confirms 0.15% move at minute 3 true probability jumps to 92% polymarket still shows 65% bot captures 27% edge before odds update (while it takes 300,000 ms for a single 5-minute candle to complete and a new one to form, the visual data on that candle is updated continuously, often every 500 to 1,000 ms ) Bot is cashing out in miliseconds beforehand whatever the result of the candle. 🫳 💰

well, I can say one thing: this bot has a cool sound

Ur quant friend must be good at maths

otherwise he wouldn't be a quant

Models fail all the time. Black Scholes doesn't work in real life, not for a short time frame of 5 min. 'true probability jumps to 92%' Is all make believe. People are making money, but not by fake models which have not much to do with reality-good in Finance departments only.

buzz word buzz word buzzword ref link

honestly i like this angle i like this angle for me it feels like using black-scholes for 5-min contracts is a pretty bold move

that's definitely the case

When pricing lags real volatility, tight models and faster execution turn milliseconds into edge.

curious about the monte carlo piece tho, 200k sims on a 5-min window means you need full run under ~500ms to still have edge, they doing GPU-accelerated or pruning the sim tree somehow?

use polybacktest for backtesting as well. it has a month of historical data and sub second snapshots. no need to risk capital testing your bot

Horrible compute usage. Monte carlo makes no sense if you already have Deribit’s implied vol surface. Running 200k simulations to detect a 3% mispricing is wildly inefficient. Not to mention black-scholes assumes geometric Brownian motion both of which do not apply to 5-m markets

@K1ngChrisCrypto

I built something like that actually planning to opensource it, the main problem usually is not the algo itself, but rather the price data as every delay changes how the bot works rapidly with some optimisations I think it is promising.

Feels like market microstructure 101

absolute display of how black scholes is eating the polymarket 5m trenches. the information delay is the ultimate alpha in 2026 and if u r not under 100ms u r just exit liquidity for the monte carlo quants no cap.

have you tried it yourself?

No they don’t do that you are full of shit farming engagement

@shinytogetick i didn't mean to offend anyone bro

just like this bot:

Interesting.