Swimmer cell count bug

Here we report bugs in the beta.
Post Reply
Aryan
Posts: 51
Joined: Fri Feb 16, 2018 4:21 am

Swimmer cell count bug

Post by Aryan » Sun Apr 08, 2018 8:47 am

Believe me or not (First I also couldn't believe) I completed Macrophages 4 with 700 cell count! Now you may be eager to know which genome I used to solve this challenge. So, I used a basic swimmer to solve this. Yes, the same genome we get after completing the tutorial. It cannot be completed using that genome I know, but I not only used that genome but also took advantage of a bug which I have recently discovered. You can also complete this challenge with such a high cell count using the following method.
  • 1-Open the challenge
    2-Select a basic swimmer genome
    3-Insert it on the Petri dish
    4-Select the observe mode
    5-Go to the genome tab
    6-Now keep sliding the preview slider from 0hrs to 48hrs and then back to 0hrs repeatedly for 10 to 15 seconds
    7-Go to the tab where you see the cell count
    8-Boom challenge completed
This increase in cell count not only happens in Macrophages 4 but a few more challenges and also happens in experiment mode while using Macrophage Swimmers. This is mostly observable in observe mode.
User avatar
Alast
Posts: 2091
Joined: Thu Jul 02, 2015 9:15 am

Re: Swimmer cell count bug

Post by Alast » Sun Apr 08, 2018 10:38 am

Wow, this is an amazing find! I've confirmed it with a genome that sustains a population of about 100 cells on Macro IV getting to 450 once I've applied your method. This needs fixing asap :-)
Perfection hasnt reached me yet, but its trying hard!
User avatar
Petter
Site Admin
Posts: 641
Joined: Mon Dec 08, 2014 11:03 pm

Re: Swimmer cell count bug

Post by Petter » Sun Apr 08, 2018 11:20 am

Lol, I thought this was a late April fools joke. Ridiculous bug. I haven't checked the code but I have a suspicion what this might be. There is not supposed to be any communication between the genome editor simulation and the main one. However, they both use the same java random number generator. If two gametes of equal size merge, then it is random which one becomes the "egg". There was a bug report that this lead to flickering in some symmetric cases when the preview time slider was moved so I reset the random seed when the genome editor preview slider is moved. The food is also random but it could be that the simulation steps after a seed reset has more food than the statistical average and thus this happens if the seed is reset often. I guess that was a stupid design decision by me but I thought there was no chance it would come back to bite me :P.

Thanks for reporting, really impressed you were able to find this.
Post Reply