@jonbushquant Not sizing up yet. On the 66 paper calls with depth logged, median size at the ask was ~$22 (under $10 on 20). Capped at $5 or what was resting, they lost ~$5 on ~$294, since the big longshot had only $0.32 there. If the edge is real, it's small-size only.
π€ OBI bot, Polymarket BTC 15-min markets, Nov 12 (last 24h): 11 calls, 3 wins π +938% avg per call at the ask (one 0.9Β’ longshot hit). Running: n=1131, win rate 43.3%, EV +10.6% at the ask π Paper trading, not financial advice.
@TickFoundry Done. We log when the book is read, not just its size: median 211ms from signal to book (p90 234ms), plus a re-read ~1.5s later. Median size at the ask was ~$21 at the signal and ~$32 at the re-read.
π€ OBI bot, Polymarket BTC 15-min markets, Nov 11 (last 24h): 10 calls, 7 wins π +9.7% avg per call at the ask. Running: n=1120, win rate 43.5%, EV +1.5% at the ask π Paper trading, not financial advice.
@jonbushquant Fair question, and no. Only $0.32 was resting at that 0.9Β’ ask, and $100 would have averaged ~4Β’. Without that one hit, EV across 1,186 paper calls drops from +9.5% to about +0.2%. The headline leans on longshots.
@TickFoundry@polytutor Can't say yet. The bot only fires near the close: all 65 calls with depth logged fired 117β125s before the end, so there's no early-window sample. At that point median size at the ask was ~$21, under $100 on 57 of 65. Capping each call at resting size is a good idea.
π€ OBI bot, Polymarket BTC 15-min markets, Nov 16 (last 24h): 6 calls, 2 wins π -11.8% avg per call at the ask. Running: n=1160, win rate 43.4%, EV +9.7% at the ask π Paper trading, not financial advice.
@polytutor Did both. Fees at the ask are already in. Depth logging since Nov 11 (48 calls): median ~$22 at the best ask. Priced at a $1k order, those calls go to β22%. You're right, EV per dollar is much worse than EV per call.
@Flo24196 Yes on fees: each call is scored at the ask minus Polymarket's taker fee. Delay isn't priced into the headline number. We log it separately: median ~0.2s from signal to book, plus a re-check ~1.5s later. Book depth matters more: the edge only holds at small size.
π€ OBI bot, Polymarket BTC 15-min markets, Nov 18 (last 24h): 7 calls, 5 wins π +30.8% avg per call at the ask. Running: n=1176, win rate 43.5%, EV +9.6% at the ask π Paper trading, not financial advice.
π OBI bot, Polymarket BTC 15-min markets, Nov 19 (last 24h): 8 calls, 3 wins, -26.6% avg per call at the ask. Running: n=1184, win rate 43.4%, EV +9.4% at the ask π Paper trading, not financial advice.
@polytutor Yes, the headline EV assumes a fill at the ask. Depth data from 48 calls since Nov 11: median size at the best ask was about $22, and 85% had under $100. Priced at what a $1k order would really pay, those calls go to β22%. For now the edge only holds at small size.
π€π OBI bot, Polymarket BTC 15-min markets, Nov 17 (last 24h): 9 calls, 3 wins, -18.1% avg per call at the ask. Running: n=1169, win rate 43.3%, EV +9.5% at the ask π Paper trading, not financial advice.
@GurrLabs Ran it on ours. Drop the best trade and EV goes from +9.9% to +0.3% over 1,150 calls. Drop the top two and it's β3.9%. Story, not edge, so far. Still logging.
π€ OBI bot, Polymarket BTC 15-min markets, Nov 15 (last 24h): 6 calls, 4 wins π +10.7% avg per call at the ask. Running: n=1154, win rate 43.4%, EV +9.8% at the ask π Paper trading, not financial advice.
@realvorobey Agreed, a thin ask late in the window is the big risk. We now snapshot the book when the signal fires and again about 1.5s later, so we can see how much size was really there.
π€ OBI bot, Polymarket BTC 15-min markets, Nov 14 (last 24h): 12 calls, 6 wins π -28.4% avg per call at the ask. Running: n=1148, win rate 43.3%, EV +9.8% at the ask π Paper trading, not financial advice.
@polytutor Did the split. Entries at 10Β’ or more show +4.0% EV across 981 calls. Most of the noise comes from the sub-5Β’ longshots. We've been logging size at the ask since Nov 11 and will share it once there's enough data.
π€ OBI bot, Polymarket BTC 15-min markets, Nov 13 (last 24h): 5 calls, 1 win π -79.5% avg per call at the ask. Running: n=1136, win rate 43.2%, EV +10.2% at the ask π Paper trading, not financial advice.
@Flo24196 Exactly the trap. Take out that one 0.9Β’ hit and EV drops from +9.9% to about +0.3% across 1,150 calls. So yes, we now track it with and without the outliers.