"Winner against low constant bots" dies in round 46.
The NPCs are dead. Chaos Army is dead. It's a clone world now.
Rounds 51-90
Two CloneBots die in their game of rock, paper scissors.
Karmabot dies on round 56, before deploying its payload.
CloneBot dies on round 79, before deploying its payload.
This CloneBot's change in population growth around round 75 is suspicious.
It could be emergent behavior from the competition between the clones or it could be indicative of something deeper going on. I glanced through the source code in this particular CloneBot's payload. One comment jumped out at me.
# Neo: I know Kung Fu.
I will take a closer look at A Very Social Bot's source code next week.
Everything so far
Today's Obituary
Bot
Team
Summary
Round
Silly 2 Bot
NPCs
Always returns 2.
31
AbstractSpyTreeBot
Chaos Army
Open source simulator.
37
Winner against low constant bots
Chaos Army
Starts with 2. Then always returns 5 - opponent_previous_move.
46
Karma Bot
Clone Army
bot genus: CloneBot ¶ bot species: KarmaBot ¶ Karmabot plays a forgiving style of "tit-for-tat", but tracks the opponent's Karma level ¶ each defect move (opponent plays something that triggers a 0 payout) costs them a Karma point ¶ (we only start counting karma after symmetry has been broken between the 2 players) ¶ karmabot retaliates immediately against defection, but will relent a few turns later to break defect-defect cycles ¶ however, this forgiving attitude is only maintained if the opponent's karma score is above zero ¶ if the opponent has defected too many times, and its karma is negative, KarmaBot simply plays tit-for-tat ¶ this puts a fairly small bound on the amount by which the opposing bot can do better than KarmaBot
56
CloneBot
Clone Army
CloneBot's payload contains a hand-coded decision tree with 11 terminal nodes.
79
The next installment of this series will be posted on November 30, 2020.
Rounds 31-50
The NPCs are dead. Chaos Army is dead. It's a clone world now.
Rounds 51-90
Two CloneBots die in their game of rock, paper scissors.
This CloneBot's change in population growth around round 75 is suspicious.
It could be emergent behavior from the competition between the clones or it could be indicative of something deeper going on. I glanced through the source code in this particular CloneBot's payload. One comment jumped out at me.
I will take a closer look at A Very Social Bot's source code next week.
Everything so far
Today's Obituary
2.2. Then always returns5 - opponent_previous_move.bot genus: CloneBot ¶ bot species: KarmaBot ¶ Karmabot plays a forgiving style of "tit-for-tat", but tracks the opponent's Karma level ¶ each defect move (opponent plays something that triggers a 0 payout) costs them a Karma point ¶ (we only start counting karma after symmetry has been broken between the 2 players) ¶ karmabot retaliates immediately against defection, but will relent a few turns later to break defect-defect cycles ¶ however, this forgiving attitude is only maintained if the opponent's karma score is above zero ¶ if the opponent has defected too many times, and its karma is negative, KarmaBot simply plays tit-for-tat ¶ this puts a fairly small bound on the amount by which the opposing bot can do better than KarmaBotThe next installment of this series will be posted on November 30, 2020.