Source code for maverick.players.archetypes.gto
from typing import TYPE_CHECKING
from ...player import Player
from ...enums import ActionType
from ...playeraction import PlayerAction
from ...utils import estimate_holding_strength
if TYPE_CHECKING: # pragma: no cover
from ...game import Game
__all__ = ["GTOBot"]
[docs]
class GTOBot(Player):
"""A bot with strategy driven by game-theory optimal solutions.
Uses hand strength evaluation to implement balanced, theoretically sound play.
Makes decisions based on hand equity with consistent bet sizing and balanced ranges.
Aims to be unexploitable by mixing actions appropriately.
- **Key Traits:** Balanced ranges, mixed strategies, uses hand equity for optimal play.
- **Strengths:** Extremely difficult to exploit, mathematically sound.
- **Weaknesses:** May underperform in soft, highly exploitative games.
- **Common At:** Mid-to-high stakes online.
"""
cls_uid = "ef7bc5f2eca24a09b25b1daeb0056353"
[docs]
def decide_action(
self,
*,
game: "Game",
valid_actions: list[ActionType],
min_raise_amount: int,
call_amount: int,
min_bet_amount: int,
) -> PlayerAction:
"""Play balanced, theoretically sound poker using hand strength evaluation."""
# Evaluate hand strength
private_cards = self.state.holding.cards
community_cards = game.state.community_cards
# Get hand equity
if community_cards:
hand_equity = estimate_holding_strength(
private_cards,
community_cards=community_cards,
n_private=game.rules.showdown.hole_cards_required,
n_simulations=1000,
n_players=len(game.state.get_players_in_hand()),
)
else:
# Pre-flop estimation
hand_equity = estimate_holding_strength(
private_cards,
n_private=game.rules.showdown.hole_cards_required,
n_simulations=400,
n_players=len(game.state.get_players_in_hand()),
)
# GTO thresholds for balanced play
strong_hand = hand_equity > 0.65
medium_hand = hand_equity > 0.45
# Standard GTO bet sizing - typically 50-75% pot
pot_bet = int(game.state.pot * 0.66)
# Bet with balanced sizing when strong
if ActionType.BET in valid_actions and strong_hand:
bet_amount = min(max(pot_bet, min_bet_amount), self.state.stack)
return PlayerAction(
player_uid=self.uid, action_type=ActionType.BET, amount=bet_amount
)
# Raise with proper sizing when strong
if ActionType.RAISE in valid_actions and strong_hand:
# GTO raises are typically 2.5-3x the current bet
# min_raise_amount is the minimum raise-by increment
# Calculate raise-to target (2x current bet), then convert to raise-by increment
raise_to_target = game.state.current_bet * 2
raise_by_amount = raise_to_target - self.state.current_bet
# Ensure we meet minimum raise requirement
raise_by_amount = max(raise_by_amount, min_raise_amount)
# Cap at stack
raise_by_amount = min(raise_by_amount, self.state.stack)
return PlayerAction(
player_uid=self.uid,
action_type=ActionType.RAISE,
amount=raise_by_amount,
)
# Call with proper odds and medium+ hands
if ActionType.CALL in valid_actions and medium_hand:
# GTO calling requires proper pot odds
if call_amount <= self.state.stack and call_amount <= game.state.pot:
return PlayerAction(player_uid=self.uid, action_type=ActionType.CALL)
# Check in balanced way
if ActionType.CHECK in valid_actions:
return PlayerAction(player_uid=self.uid, action_type=ActionType.CHECK)
# Fold when no good option
return PlayerAction(player_uid=self.uid, action_type=ActionType.FOLD)