Eternal best model actually hurts learning

eureqa.jl

@@ -470,8 +470,6 @@ function run(
         )::Population
 
     allT = LinRange(1.0f0, 0.0f0, ncycles)
-    bestScore = Inf32
-    bestTree = Node(1f0)
     for iT in 1:size(allT)[1]
         if annealing
             pop = regEvolCycle(pop, allT[iT], annealing=true)
@@ -479,14 +477,6 @@ function run(
             pop = regEvolCycle(pop, 1.0f0, annealing=true)
         end
 
-        # Save current best
-        bestCurScoreIdx = argmin([pop.members[member].score for member=1:pop.n])
-        bestCurScore = pop.members[bestCurScoreIdx].score
-        if bestCurScore < bestScore
-            bestScore = bestCurScore
-            bestTree = copyNode(pop.members[bestCurScoreIdx].tree)
-        end
-
         if verbosity > 0 && (iT % verbosity == 0)
             bestPops = bestSubPop(pop)
             bestCurScoreIdx = argmin([bestPops.members[member].score for member=1:bestPops.n])
@@ -495,10 +485,6 @@ function run(
         end
     end
 
-    # TODO - make this the oldest member instead of 1
-    pop.members[1].tree = bestTree
-    pop.members[1].score = bestScore
-
     return pop
 end