Skip to content

Crypto touch ladder

Paper bot · Retired

Prices Polymarket's daily "What price will Bitcoin hit today?

Historical record unavailable

Live record lost in the 24 Oct 2026 server move.

Simulated paper trading only — orders are not sent to Polymarket.

In plain words

TL;DR

Prices Polymarket's daily "What price will Bitcoin hit today?" markets with a touch-probability formula and makes at most 5 picky bets a day. A forward test.

How it bets

Polymarket asks daily "Will Bitcoin reach $X?" (and "dip to $X"), for Bitcoin, Ethereum, Solana and XRP. It says YES if any Binance 1-minute candle that day (ET) touches the level. The bot takes the live Binance price and the last 28 days of 15-minute moves to measure how jumpy the coin is, separately for Saturdays, Sundays and weekdays and for each part of the day (Saturdays and nights are much calmer), then nudges that up or down depending on how wild or quiet today has been so far. The reflection principle (a standard result for random walks) turns that into a fair chance of touching each level before midnight ET. Levels already touched are skipped. When YES or NO is at least 5¢ cheaper than fair after the Polymarket taker fee, it buys: $20, best edge first, at most 5 a day, held to the end. Credit: Polymarket's daily crypto touch markets, the reflection-principle barrier probability, and Binance as the data (and resolution) source.

Example

Bitcoin is $84,600 with 18 hours left. A random walk this jumpy touches $85,000 about two times in three, but YES costs 39¢ + ~1.7¢ fee → about 25¢ cheap → bet.

Why it might work

Daily touch markets are thin and price the far strikes by feel; the touch formula is the right shape for "did it ever reach X".

Why it might fail

Our research check found +5% (NO bets) and +8% (YES bets) over 183 events, in-sample after cleaning stale prices; the edge is close to the spread and the fee. The formula ignores trends and news jumps, and the market may know more than 3-day volatility.

Exact rules

No new bets since 6 Oct 2026 (retired); open bets settle as usual. The rules it traded by: Paper bot on Polymarket's daily "What price will Bitcoin / Ethereum / Solana / XRP hit on <date>?" events. Each strike market resolves YES if any Binance 1-minute candle of <coin>/USDT between 12:00 AM and 11:59 PM ET on that date has a High at or above an up-strike (↑) or a Low at or below a down-strike (↓). The bot prices that touch with the reflection principle: P(touch) = 2·(1 − Φ(|ln(H/S)| / √V)), capped at 1. S is the latest Binance 1-minute close and V is the variance of the log price over the time left in the window, taken from the last 28 days of 15-minute Binance bars: a variance-per-hour profile for Saturdays, Sundays and weekdays separately and for each 4-hour ET block, summed over the hours still to run (Saturdays and quiet overnight hours are far calmer than a flat trailing number). That remaining variance is then scaled by how today has behaved so far: the square root of (realized variance so far ÷ the profile's expected variance for those hours), clamped to 0.4–2×, once at least 8 fifteen-minute bars have elapsed. Strikes already touched today are skipped, and so is any strike where the model and the market's YES price differ by more than 25¢ (more often a model error than an edge). It buys YES or NO when the model beats the ask plus the taker fee (the market's own feeSchedule) by at least 5¢, the ask is at least 2¢, and at least 1 hour remains; at most 5 bets per UTC day, best edge first; one bet per market; $20 per bet; hold to settlement. Research backtest: in-sample, +5% (NO) and +8% (YES) over 183 events after removing stale prices; that is close to the spread and the fee, so this is a forward test, not a claim. Main risks: a driftless model ignores trends and jumps, and the market may know more than 3-day realized vol.

Source: Polymarket daily crypto touch markets · Binance data ↗

Bets

No published bets yet: this record has no settled, closed or open bets.