Toby Ord: Elo grows linearly but winning odds grow exponentially — odds are the real AI progress metric
tobyordoxford · x · 2026-11-14
Philosopher Toby Ord argues in a thread with Ramez and Ben Todd that the fundamental measure for game-playing AI is not Elo (proportional to the log of winning odds) but the odds themselves: Elo has grown linearly while underlying odds grow exponentially.
He maps this onto general capability measurement:
- Win probability ≈ benchmark score (typically logistic growth)
- Winning odds ≈ task-success odds (expected consecutive successes equals the odds; median 50% horizon becomes near-linear in odds at high odds)
- Elo ≈ ECI
This explains why logistic benchmark-score growth corresponds to linear ECI growth and exponential growth in task time horizon — odds serve as the bridging variable.
Related event: Oxford's Toby Ord: RL Compute Needs 10,000x Boost to Lift Benchmark Scores(17 posts)→
More from AGI Musings
- Interconnects: Jensen's resignation is a major setback for US open-model ecosystem — Interconnects (Nathan Lambert) · 2026-11-14
- AGI endgame: either a balanced multipolar world or a centralized panopticon — CedricMakes · 2026-11-14
- "60% of people around you have some form of AI psychosis," says Alsop III — StewartalsopIII · 2026-11-14
- AI won't make everyone equal—it will split the smart from the rest worse than money — prasenx · 2026-11-14
- Toddlers learn new verbs only from responsive partners; a call to rerun the test with AI — AnnaCiaunica · 2026-11-14
- Developer muses: in Murphy's world, billions of interdependent events make rare failures daily — bingxu_ · 2026-11-14