GitHub

ChessAgent 0.1

Learnt COMP30024: Artificial Intelligence, and thought to myself

Make bot better than Magnus

Match

Click Start Game to play

Advantage

Equal

Positive means human/white is ahead; negative means agent/black is ahead.

Select minimax to start plotting evaluations.

Agent stats

Loading

Games

0

Agent win rate

0%

Human win rate

0%

Draw rate

0%

Abandoned

0%

Avg moves

0

Avg duration

-

Avg final eval

-

Control how the agent works!

Search stats

Alpha-beta minimax with deterministic move ordering.

Last agent search

\(b_{root}=0\)

\(\bar{b}=0\)

\(N_{generated}=0\)

\(N_{expanded}=0\)

\(B_{pruned}=0\)

\(r_{prune}=0\%\)

\(T_{hits}=0\)

\(T_{size}=0\)

Game totals

\(A_{moves}=0\)

\(N_{generated,total}=0\)

\(N_{expanded,total}=0\)

Heuristics

\[ h(s)=1h_1(s) + 1h_2(s) + 1h_3(s) + 1h_4(s) \]

h1: Mobility

\[h_1(s)=M_{agent}(s)-M_{opp}(s)\]

h2: Piece count

\[h_2(s)=P_{agent}(s)-P_{opp}(s)\]

h3: Material

\[h_3(s)=V_{agent}(s)-V_{opp}(s)\]

Pawn 1, knight 3, bishop 3, rook 5, queen 9, king 0.

h4: Center control

\[h_4(s)=C_{agent}(s)-C_{opp}(s)\]

Counts occupying or attacking d4, e4, d5, and e5.