The numbers, live.

Name: Pick1 AI sports prediction public ledger
Creator: Pick1
License: https://creativecommons.org/licenses/by/4.0/

Straight from Pick1's public ledger. Updates as new picks are graded. Tier breakdown shows whether the model's confidence calibration is real.

Total picks

601

Wins

365

Losses

236

Overall win rate

60.7%

By league

League	Picks	Wins	Losses	Win rate	High conf	Med conf	Recent
ATP	180	123	57	68.3%	100.0%	84.9%	—
UFC	133	59	74	44.4%	54.8%	40.0%	—
NFL	64	39	25	60.9%	71.4%	66.7%	—
T20I	55	35	20	63.6%	100.0%	66.7%	—
WNBA	42	30	12	71.4%	100.0%	81.8%	—
NHL	41	26	15	63.4%	—	66.7%	—
MLB	30	17	13	56.7%	—	50.0%	—
EPL	14	8	6	57.1%	—	66.7%	—
NBA	9	3	6	33.3%	—	—	—
KHL	9	5	4	55.6%	—	—	—
TOP14	8	8	0	100.0%	—	100.0%	—
NCAAF	6	6	0	100.0%	100.0%	100.0%	—
KBO	4	2	2	50.0%	—	—	—
F1	3	1	2	33.3%	—	—	—
GOLF	1	1	0	100.0%	100.0%	—	—
AFL	1	1	0	100.0%	—	—	—
LALIGA	1	1	0	100.0%	—	100.0%	—

How to read these numbers

Win rate is wins / (wins + losses), pushes excluded. High conf = picks published with ≥80% AI confidence. Med conf = 65–79%. Recent = trailing 30 days. A well-calibrated model wins more often at higher confidence — if those columns are flat, the model is mis-priced.