In 2025 I created an Arbitrage Bot for Kalshi <> Polymarket <> Opinion
The results were actually quite surprising
I will share how much I've been able to make using the bot at the end of the thread!
But first, how does it even work?
1. The Strategy
A lot of the viral trading bots on X for Prediction Markets are trying to trade just within Polymarket, mostly on highly volatile markets. Buying quickly / selling quickly, this works great if you’re a VC with an infinite budget for development and the fastest infrastructure positioned right next to Polymarkets servers, but doesn't work great for regular traders.
I knew building a bot here is like trying to out sprint Usain Bolt wearing flip flops
So instead if i wanted to make a bot it would have to be doing something I haven't seen anyone else do
So after engaging 100% of my brain I remembered how I'd seen many of the same markets when browsing Kalshi that I'd previously seen on Polymarket. If the odds were different I'd actually be able to arbitrage the two platforms.
After checking 20 or so pairs I found a few examples where the odds were different, proving my theory. So i started building
2. The Bot
I began crafting an algorithm that collects all the Polymarket, Kalshi & Opinion markets and identifies any that are the same. In total that's over 6,000,000 markets, billions of potential combinations, from that i found around 6,000 that match.
A lot of markets have the same title, but slight differences in the rules that means an arbitrage is not possible. For example ‘Will Trump meet XYZ person before ABC date’ - many of these exist between Poly <> Kalshi but Kalshi defines ‘meeting’ as online, or in real life, whilst Polymarket defines it strictly as in real life.
To verify the matches I first used an algorithm that sees if the titles / rules are similar, then I feed every one of those potential pairs into an LLM that is prompted to approve / reject the potential match. This part sounds like it should be easy, but the LLMs were all so inconsistent. I’d pass the same market 10 times and it would reject it twice and approve it eight times. After a lot of prompt engineering I finally got it to a stage where it's at least 99% accurate.
The bot then recreates locally all the orderbooks for the matched markets and looks for any pairs with YES + NO values of < $1 & calculate how much you can profitably buy of each side ( your EV )
3. An Example
‘SpaceX IPO Closing Market Cap? 900B to 1T'
This market exists on both Poly & Opinion, but right now the odds are different.
YES / NO
4% / 97% - Polymarket
10% / 92% - Opinion
Lets say you told the bot you wanted to trade $100
You’d split it
$3.85 into Polymarket YES for 104.743 Shares
$96.15 into Opinion NO for 104.743 Shares
You spent $100, and you got 104.7 shares of each side
And because 1 YES + 1 NO = $1 ( when the markets are identical )
Your expected profit is $4.7, because:
value ( 104.7 shares * $1 ) - cost ( $100 ) = profit ( 4.7 Shares / $4.7 )
Then once the odds rebalance, usually a few days to a week, you can sell both sides and use the principal + the profit to take another arbitrage
$100 isn't much, but imagine if you did $1000, or $10,000 every few days.
Since creating the bot it has been finding around $350,000 in arbitrages from over 5,000 matching market pairs every month.
That's more than I could ever take my self
Assuming an average of 4% per arb (often more), It would take over $8,750,000 deployed to capture them all simultaneously
That's why I initially shared the bot with some sharp sportsbook bettors for commission on their trades, eventually making it a public version that's on AlertPilot for anyone to use.
4. The Longevity
Will this work forever?
No, eventually a VC will come and eat our free lunch, it will become more competitive and the spreads wont be as high as they are now (+10% at times).
My mantra here is just to make the most of it whilst I can, and release consistent updates so that it remains competitive against the competition. I already have a head start in that regard.
I will say over the last 3 months of using the bot myself, plus adding traders / syndicates to the platform to trade as well, the amount of Arbitrages found has only increased. In large part no doubt to the huge increase in volume on PM’s, especially Opinion where many of the more profitable arbitrages have been.
5. The Profit
So how much did I make?
I started with a bankroll of $1,000 as I was skeptical initially how well the bot would work, first week I was able to close two 5% arbs. I then increased my bank roll and started sharing the bot with other traders. Since then I've been averaging five figures every month from my own trades + commissions.
Anywhere between 12% to 20% per month in closed arbs, personally.
To anyone reading this that wants to arbitraging these markets, i've released a version of the bot on AlertPilot!
And for any VCs or betting syndicate that wants to trade these markets, my DMs are open.
Aspiring trading bot makers, there is money out there if you can be creative in your strategies
Have lots of plans for AlertPilot + the Arbitrage bot going into 2026. Happy new year everybody!
programmers are making $10k-200k monthly on polymarket
here's how they actually do it
///
1. arbitrage bots
buy YES + NO when combined price < $1
example: YES at 48¢ + NO at 49¢ = 97¢ total
you lock $0.03 profit per $1 no matter who wins
trader "distinct-baguette" made $242k in 1.5 months doing this
targets 15-min crypto markets where prices move fast
python script polls API every 1-3 seconds, executes when sum < 99¢
///
2. statistical arbitrage
find correlated markets that drift apart
"trump wins" vs "GOP senate control" should move together
when spread hits 4-7%, short expensive one, long cheap one
close when they converge
trader "sharky6999" made $480k scanning 100+ markets per minute
///
3. AI probability models
train ML models to estimate real odds from news/social data
if your model says 60% YES but market at 50¢, buy
trader "ilovecircle" made $2.2M in 2 months with 74% accuracy
uses ensemble of 10 AI models, retrains weekly
///
4. spread farming
buy at bid (5¢), sell at ask (6¢), repeat
or hedge across platforms (short polymarket, long binance)
trader "cry.eth2" made $194k with 1M trades
high-frequency loop via CLOB API
///
5. copy-trading automation
mirror successful whale traders automatically
scan profiles, execute proportional trades
one bot made $80k in 2 weeks copying near-resolved markets
///
tech stack:
python + requests library for API calls
web3-py for blockchain interactions deploy on VPS for 24/7 operation
polymarket has REST APIs for everything:
gamma markets API (prices/volumes)
CLOB API (place orders)
data API (track positions)
///
starting point:
> build simple arbitrage bot first
> fund with $100-1k for testing target high-volume markets (politics/crypto)
> expect 50-70% win rate but focus on positive EV
+$133k earned on crypto launches without participating in them on Polymarket
FilthyBera made 37 predictions across multiple crypto token launches, and bet on NO in most cases
It allows him to have 100% win rate, and it seems he's not going to stop
His profile:
https://t.co/PMm2HvUq7y