ok

algo/epsilon_greedy.go

@@ -2,9 +2,12 @@ package algo
 
 import (
 	"lukechampine.com/frand"
+	"tuxpa.in/a/gambit"
 	"tuxpa.in/a/gambit/helper"
 )
 
+var _ gambit.Bandit = (*EpsilonGreedy)(nil)
+
 type EpsilonGreedy struct {
 	Epsilon float64
 	cr      helper.CountReward

example/simple/main.go

@@ -17,13 +17,12 @@ func main() {
 
 	n := 100
 	for i := 0; i < n; i++ {
-		b.Update(
+		g.Observe(
 			// select a random arm
-			b.Select(frand.Float64()),
+			frand.Float64(),
 			// and supply a random score
 			float64(frand.Intn(4)),
 		)
 	}
-
 	log.Println(g.AllocateSolution())
 }

gang.go

@@ -15,3 +15,8 @@ func (g *Gang) AllocateSolution() ([]int, []float64) {
 	g.b.Reward(b)
 	return a, b
 }
+
+func (g *Gang) Observe(score float64, reward float64) {
+	g.b.Update(g.b.Select(score), reward)
+
+}