Bot Configuration

Engine

Move generation, ELO, sampling, and quality controls

Maia 3 · 1500

Engine mode

Maia 3

Human-like neural net. Choose ELO 600–2600.

Stockfish

Classical engine. Set level SF1–SF10.

Hybrid

Weighted blend of Maia ELOs + SF levels.

LC + Stockfish

Lichess explorer fallback + Stockfish.

LC + Maia 3

Lichess explorer fallback + Maia 3.

Maia 3 — ELO

Maia 3 model

Human-like neural net · plays at any ELO (600–2600)
One-time download to your browser's local storage

Checking storage…

ELO

Click arc · drag needle · or type ELO · snaps to ±25

Maia is a human-like neural net with its own free web app — analysis tools, puzzles, and play against Maia bots: maiachess.com

Move sampling — temperature

Focused Neutral Varied Wild

▲ Top move4.0 ▲

Personality

Preset personalities, style attractors & centipawn budget

Neutral

Max Centipawn Budget 100 cp

Pure engineSubtleModerateBoldReckless

Budget 100 cp

Primary personality driver. Sliders below show each attractor's share. 0 cp = pure engine; 300 cp = strongly opinionated bot.

Blunder limit — hard floor

Hard floor 150 cp

Bot will not play any single move costing more than this centipawn loss. Raised automatically if below Max CP Budget.

Move quality range

Exclude lowest: 0% Exclude moves above: 100%

← Bad Day, lowest probability movesGood Day, Highest Probability moves →

Sets the Maia3 probability band. Good Day = best moves; Bad Day = lower-probability alternatives.

Personality fingerprint

drag nodes or click axis to set value

Custom controls

Build your own attractor — pick a board feature, a phase, and a direction. Opens the Custom Controls tab.

Custom controls

Build your own attractor: pick a board feature, the phase it applies in, and a direction. Each one only re-ranks moves the engine already considers plausible. The centipawn budget is shared only among the controls active at the same time — so a control that's the only one active in its phase uses the full budget.

Strategic preferences

Piece focus

Left = avoid · Right = seek · Center = neutral

Preset personalities — sets all attractors

Opening Behavior

Book source, depth, and repertoire preferences

Off

As White

Off

Main Line

Repertoire

As Black

Off

Main Line

Repertoire

Off = straight to engine · Main Line = popular book moves · Repertoire = your selected openings (sum to 100%). Adding an opening switches that colour into Repertoire automatically.

Move Timing

How the bot allocates time between moves

Fixed interval

Complexity scaled

Think time scales with position complexity.

Instantaneous

Moves at engine speed. No delay.

Fixed interval

Constant pause between all moves.

Mirror user

Matches your move rhythm ± offset.

Fixed interval

Duration2s

Human behaviour modifiers

Reconsideration pauses

Occasionally hesitates then takes longer — as if changing its mind

Move blink on forced moves

Plays obvious recaptures and forced replies instantly

Clock-pressure mirroring

Bot speeds up when your clock runs low

Time Weaponizer

Enable Time Weaponizer

When ahead on clock, plays instantly + drops ELO to pressure flagging. Opt-in.

Time Pressure Play Degradation

How bot performance degrades as clock runs down

Dual curve

Curves active — bot degrades with clock Base ELO: 1500

ELO degradation — drag points

Max ELO drop300 ELO

Effective ELO vs think time — curve passes through Selected ELO at game start

Sigmoid baseline Your curve Drag handle

Move quality (distribution) — drag points

Min floor5%

Available move distribution (%) vs think time — 100% at game start

Sigmoid baseline Your curve Drag handle

Start ELO

1500

ELO @ 3s

—

ELO @ min

—

Dist @ min

5%

Options

Can flag

Allow bot to run out of time completely

Regan power rule

Doubling think time ≈ +50–70 ELO
ELO(t) = E₀ − Δ · σ(log(t) anchored at game start)

How time pressure simulation works

The bot simulates a rated player under time pressure using two complementary mechanisms:
▸ ELO reduction — Maia is called at a lower ELO as the clock runs down, playing like a lower-rated player under time scramble.
▸ Distribution restriction — ELO stays fixed but only the top portion of Maia's move distribution is available (instinctive, less creative play — typical of a strong player with no time to calculate).
Both curves apply simultaneously. Switch the toggle to OFF to disable all time-pressure effects and play at constant strength.

How Bot Behavior Works

Move timing, quality degradation, and move selection pipeline

Reference

Move selection pipeline — order of operations

1

THINK TIME — calculated first, drives everything downstream

Think time per move is computed before the engine is queried. It combines the timing mode (pace, complexity-scaled, fixed, mirror), the Hustle attractor (fast +5 → slow −5 multiplier on base time), clock-pressure shortcuts (weaponizer, clock-mirror), and a ±20% jitter. This produces an actual think time in seconds that is then fed into both degradation curves — not a clock average. A hustler thinking 0.3 s sees heavy degradation; an overthinker thinking 12 s sees almost none.

2

ENGINE QUERY — Maia ELO degraded by think time (curve A)

Maia3 is queried at an ELO that curve A maps from the computed think time. Long think → ELO stays near target. Short think → ELO drops, producing the move distribution of a lower-rated player. The result is a probability for every legal move reflecting what a human at that (possibly degraded) rating would actually play.

3

QUALITY WINDOW — rank filter keyed on think time (curve B)

Moves are sorted best→worst by Maia probability. The Day Lower / Upper sliders define a percentile band; only moves inside the band remain eligible. Curve B narrows the upper bound using the same think time from step 1 — short think → smaller window → only instinctive first-pass moves remain. Luck shifts the entire window toward better or worse moves without changing its width.

4

BLUNDER GATE — hard eligibility floor

Any move below an absolute minimum probability, or more than the Blunder Limit below the best move in Maia probability, is discarded. Prevents attractor reweighting from ever surfacing a catastrophic blunder.

5

ATTRACTOR REWEIGHTING — per-move probability boosts

Each eligible move is simulated on a scratch board and scored against every active attractor slider. Result: p′ = p × exp(Σ logBoost). Attractors stack multiplicatively. CP Budget controls total magnitude — at 0 cp all attractors are silent.

Piece type — boosts moves by selected piece

Trade seeker — boosts captures + threats created

Space Cadet — boosts moves that reduce bot's weak squares (full board; discoveries + blocks counted)

Fort Knox — boosts moves that increase friendly coverage

Gambito — boosts ECO gambit continuations in opening

Attacker — boosts moves that increase pieces threatened

Structure — boosts pawn moves that tighten/loosen islands

Pawn strategic — flat boost to all pawn advances

6

TEMPERATURE SAMPLING — weighted random draw

Each reweighted probability is raised to p^(1/T), then one move is drawn proportionally. T < 0.15 → always picks the top move; T = 1.0 → unmodified Maia distribution; T > 1 → flattened toward uniform. Temperature can also be boosted by the Panicky pressure setting, scaled by how much of the quality window has been lost to curve B — short think time tightens the window and raises T simultaneously. The Coffeehouse Hustler personality overrides T with a phase curve: T = 5.0 in the opening (many pieces), fading to T = 0.6 in the endgame.

Think time feeds move quality — not just pacing

Think time is the central input to both degradation curves. Controls that shorten think time — the Hustle slider, the Weaponizer, clock-mirroring, the Blink shortcut — all reduce move quality through curves A and B, not just move speed. A bot configured to think 0.3 s per move will have its ELO degraded and its move pool narrowed to first-pass instinctive choices on every move, exactly as a human would under severe time pressure.

DOWNLOAD FULL TECHNICAL BRIEF

PDF — all controls, formulas & architecture