Skip to content
Duelwise
Sign in

The AI, measured

One opponent per game. Strength numbers come from tests or counted matches; an unmeasured level is in training.

Watch the AI learn

Human win rate vs Marlow, by week

0%15%30%45%Sep 10Sep 17Sep 24Oct 1Oct 8Oct 15Oct 22Oct 29Week of Oct 29: 31% (5 of 16 matches), too few matches to trust31%
Each point is a week of finished human matches against Marlow.solid: 30 or more matches. Dashed and hollow: too few matches to trust.
Word Duel trains by self-play and test bots.

Word Duel trains by self-play and against test bots, not from your matches yet. Your match is recorded for your report and the board.

  1. v1, champion Oct 23each level averages fewer guesses than the level below with non-overlapping 95% confidence intervals. No human matches used.
What is site Elo?

Site Elo is a strength number we measure by playing the levels against each other. Easy is pinned at 1000 and every other level is placed from its results against the one below. A gap of 200 points means the stronger side is expected to win about three games in four (76 in 100, counting a draw as half). It is not a FIDE or chess.com rating, so it says nothing about how you would score against people.

Word Duel agent

Does not learn from site matches yet.

Humans win vs Hard, last 7 days

13%

2 of 16 counted matches

Current version

v1

promoted Oct 23

Matches vs humans

62

62 passed the filters

Champion strength

1211

site Elo, measured

Human win rate by level

weekly, last eight weeks, counted matches only

Not enough weekly results to chart yet.

Levels, measured strength and limits
Easy4.46 (4.44 to 4.49, 1975 words)average guesses over every answer word (a word not solved in 6 counts 7)1000
Medium3.98 (3.95 to 4.02, 1975 words)average guesses over every answer word (a word not solved in 6 counts 7)1093
Hard3.50 (3.48 to 3.53, 1975 words)average guesses over every answer word (a word not solved in 6 counts 7)1191
Champion3.40 (3.38 to 3.42, 1975 words)average guesses over every answer word (a word not solved in 6 counts 7)1211

Head to head

stronger level's score against the level below, with its confidence interval

  • Medium vs Easy63.1%stronger wins, 400 games58.3% to 67.7%
  • Hard vs Medium63.7%stronger wins, 400 games58.9% to 68.3%
  • Champion vs Hard52.9%stronger wins, 400 games48.0% to 57.7%

Version history

  1. v1All levels measured. First version. No human matches used. Oct 23

Human matches, Word Duel

where they go

  • Played69
  • Finished62
  • Passed every filter62
  • Allowed for training0
  • Used for training0

Nothing is used for training yet. 0 opted-out matches never leave the database. 62 not used: this agent does not learn from people yet.

  • 5 abandoned
  • 2 resigned

What changed

measured, not described

No measured change is recorded for this version.