• Ten rules for a Polymarket or Kalshi bot: five to keep, five to dropTen rules for a Polymarket or Kalshi trading bot from two months of live testing: five that saved money and five that cost it.
• Where to get Polymarket and Kalshi historical dataFree Polymarket historical data, Kalshi candlesticks, on-chain fills and your own logger, plus how to backfill official outcomes and get spot prices.
• Where to host a prediction-market bot, and why the region decides itWhich region to host a trading bot in, what latency did to a live result, which VPS providers fit, and the burstable tier to avoid.
• What a trade really costs on Polymarket and KalshiPolymarket fees, Kalshi fees and the other two exchanges compared: the formula, the breakeven win rate at every price, and why rebates do not fix it.
• Case studyThe favourite in the last 15 secondsA taker edge found by measuring the book near the close. Positive on seven coins, dead without a fast order path.
• Case studyA model with a live checkpoint that passedNine book features, one season of training, one untouched test, then live money with a pre-registered bar.
• Case studyThe maker seat on 4-hour marketsLiquidity rewards pay per market per day. On long windows that arithmetic clears the bar, then stops scaling.
• Case studyThe first live bot: 83 percent wins, zero profitBuy the favourite, hold to settlement. What broke, what the labels hid, and why a high win rate is not an edge.
• Should an AI make your bot's trading decisions?Per-decision model costs against what a $10 ticket on a thin prediction-market book can pay, and where a language model actually earns its keep.
• What happens after you find an edgeLate books emptied within a week, leaders rotated, and one exchange changed its settlement rule overnight.
• Do prediction-market bots work? Two months live, in numbersDo Polymarket and Kalshi bots make money? The base rates, which strategies cleared the fee in live testing, and where the money actually sits.