Merge pull request #1153 from gaffney2010/master

Implemented Leyvraz, K86R from Axelrod's second

Author: drvinceknight

axelrod/strategies/_strategies.py

@@ -10,7 +10,7 @@
 from .axelrod_second import (
     Champion, Eatherley, Tester, Gladstein, Tranquilizer, MoreGrofman,
     Kluepfel, Borufsen, Cave, WmAdams, GraaskampKatzen, Weiner, Harrington,
-    MoreTidemanAndChieruzzi, Getzler)
+    MoreTidemanAndChieruzzi, Getzler, Leyvraz)
 from .backstabber import BackStabber, DoubleCrosser
 from .better_and_better import BetterAndBetter
 from .bush_mosteller import BushMosteller
@@ -197,6 +197,7 @@
     Kluepfel,
     KnowledgeableWorseAndWorse,
     LevelPunisher,
+    Leyvraz,
     LimitedRetaliate,
     LimitedRetaliate2,
     LimitedRetaliate3,

axelrod/strategies/axelrod_second.py

@@ -1300,11 +1300,11 @@ def strategy(self, opponent: Player) -> Action:
         if self.burned or random.random() > self.prob:
             # Tit-for-Tat with probability 1-`prob`
             return self.try_return(opponent.history[-1], inc_parity=True)
-        else:
-            # Otherwise Defect, Cooperate, Cooperate, and increase `prob`
-            self.prob += 0.05
-            self.more_coop, self.last_generous_n_turns_ago = 2, 1
-            return self.try_return(D, lower_flags=False)
+        
+        # Otherwise Defect, Cooperate, Cooperate, and increase `prob`
+        self.prob += 0.05
+        self.more_coop, self.last_generous_n_turns_ago = 2, 1
+        return self.try_return(D, lower_flags=False)
 
 
 class MoreTidemanAndChieruzzi(Player):
@@ -1454,3 +1454,57 @@ def strategy(self, opponent: Player) -> Action:
         self.flack *= 0.5  # Defections have half-life of one round
 
         return random_choice(1.0 - self.flack)
+
+
+class Leyvraz(Player):
+    """
+    Strategy submitted to Axelrod's second tournament by Fransois Leyvraz
+    (K68R) and came in eleventh in that tournament.
+
+    The strategy uses the opponent's last three moves to decide on an action
+    based on the following ordered rules.
+
+    1. If opponent Defected last two turns, then Defect with prob 75%.
+    2. If opponent Defected three turns ago, then Cooperate.
+    3. If opponent Defected two turns ago, then Defect.
+    4. If opponent Defected last turn, then Defect with prob 50%.
+    5. Otherwise (all Cooperations), then Cooperate.
+
+    Names:
+
+    - Leyvraz: [Axelrod1980b]_
+    """
+
+    name = 'Leyvraz'
+    classifier = {
+        'memory_depth': 3,
+        'stochastic': True,
+        'makes_use_of': set(),
+        'long_run_time': False,
+        'inspects_source': False,
+        'manipulates_source': False,
+        'manipulates_state': False
+    }
+
+    def __init__(self) -> None:
+        super().__init__()
+        self.prob_coop = {
+            (C, C, C): 1.0,
+            (C, C, D): 0.5,  # Rule 4
+            (C, D, C): 0.0,  # Rule 3
+            (C, D, D): 0.25,  # Rule 1
+            (D, C, C): 1.0,  # Rule 2
+            (D, C, D): 1.0,  # Rule 2
+            (D, D, C): 1.0,  # Rule 2
+            (D, D, D): 0.25  # Rule 1
+        }
+
+    def strategy(self, opponent: Player) -> Action:
+        recent_history = [C, C, C]  # Default to C.
+        for go_back in range(1, 4):
+            if len(opponent.history) >= go_back:
+                recent_history[-go_back] = opponent.history[-go_back]
+
+        return random_choice(self.prob_coop[(recent_history[-3],
+                                             recent_history[-2],
+                                             recent_history[-1])])

axelrod/tests/strategies/test_axelrod_second.py

@@ -930,3 +930,33 @@ def test_strategy(self):
         actions = [(C, C), (C, D), (C, C), (C, D), (D, C), (C, D)]
         self.versus_test(axelrod.Alternator(), expected_actions=actions, seed=4, attrs={"flack": 5./16.})
 
+
+class TestLeyvraz(TestPlayer):
+    name = 'Leyvraz'
+    player = axelrod.Leyvraz
+    expected_classifier = {
+        'memory_depth': 3,
+        'stochastic': True,
+        'makes_use_of': set(),
+        'long_run_time': False,
+        'inspects_source': False,
+        'manipulates_source': False,
+        'manipulates_state': False
+    }
+
+    def test_strategy(self):
+        actions = [(C, C)] * 100
+        self.versus_test(axelrod.Cooperator(), expected_actions=actions)
+
+        actions = [(C, D), (C, D), (D, D), (D, D)]
+        self.versus_test(axelrod.Defector(), expected_actions=actions, seed=1)
+        actions = [(C, D), (D, D), (D, D), (C, D)]
+        self.versus_test(axelrod.Defector(), expected_actions=actions, seed=2)
+
+        actions = [(C, D), (C, C), (D, C), (C, D), (D, C), (D, D), (C, D), (D, C), (C, D)]
+        self.versus_test(axelrod.SuspiciousTitForTat(), expected_actions=actions, seed=1)
+
+        actions = [(C, C), (C, D), (D, C)] + [(D, D), (C, C)] * 3
+        self.versus_test(axelrod.Alternator(), expected_actions=actions, seed=2)
+        actions = [(C, C), (C, D), (C, C)] + [(D, D), (C, C)] * 3
+        self.versus_test(axelrod.Alternator(), expected_actions=actions, seed=3)

docs/reference/overview_of_strategies.rst

@@ -136,7 +136,7 @@ repository.
    "K65R_", "Mark F Batell", "Not Implemented"
    "K66R_", "Ray Mikkelson", "Not Implemented"
    "K67R_", "Craig Feathers", ":class:`Tranquilizer <axelrod.strategies.axelrod_second.Tranquilizer>`"
-   "K68R_", "Fransois Leyvraz", "Not Implemented"
+   "K68R_", "Fransois Leyvraz", ":class:`Leyvraz <axelrod.strategies.axelrod_second.Leyvraz>`"
    "K69R_", "Johann Joss", "Not Implemented"
    "K70R_", "Robert Pebly", "Not Implemented"
    "K71R_", "James E Hall", "Not Implemented"

docs/tutorials/advanced/classification_of_strategies.rst

@@ -47,7 +47,7 @@ strategies::
     ... }
     >>> strategies = axl.filtered_strategies(filterset)
     >>> len(strategies)
-    80
+    81
 
 Or, to find out how many strategies only use 1 turn worth of memory to
 make a decision::
@@ -69,7 +69,7 @@ range of memory_depth values, we can use the 'min_memory_depth' and
     ... }
     >>> strategies = axl.filtered_strategies(filterset)
     >>> len(strategies)
-    56
+    57
 
 We can also identify strategies that make use of particular properties of the
 tournament. For example, here is the number of strategies that  make use of the