Simulating Evolution without Natural Selection

Music, models, and other licenses for the assets used can be found at https://pastebin.com/LTYpvMka

Discord: discord.com/invite/At5dqKpD7w
Patreon:   / professorghost  

Selection video:    • Simulating Natural Selection  

The simulation you watched at 2:15 was handpicked by me. It was the shortest simulation which had a nice curve, of both replicators rising to the carrying capacity first and then one of them falling down. You watched it at 2x speed (or else it would've dragged on too long), meaning that the "average" times until fixation which I gave you would be halved if we take that speed as the reference one.

Most simulations I left running at night, but I did sit and watch many many of them (at 32x speed). It's interesting the insights you get when you watch many different outcomes come out of the same random process.

The simulations with 4x the size (5:24) actually took on average LESS than 4x what the normal one took (the normal-sized took 18min, the 4x took 53min). I was expecting the opposite: for it to take way longer, given that there is plenty of time for a dwindling replicator to swing the numbers back in its favour.
Potential causes: The simulation was initialized with 4x the starting number of replicators, although it had only around 3.3x the pellet spawning rate (due to how it was coded, it was limited to spawning 1 pellet per frame). Other than that, it might've been a statistical fluke. I only ran 150 simulations at 4x size to calculate the average, and didn't think to calculate the standard deviation before deleting the data.

If you'd like to get in contact with me, I can be found at [email protected]