Devourer Bias

[unionruler]

I just wanted to post that 'r' that danieldubois couldn't understand what is is actually the probability of getting a certain outcome. Not sure which it is in this case, i.e., I'm not sure if r is actually the probability of taking aether quanta or light quanta.

131! is too big and I can't be bothered at the moment to find out where to calculate it.

Ah well it doesn't matter one look at those numbers and you know it'd fail a z-test with reasonable significance levels.

[Hobnob5000]

The Bias definitely exists, my RoL/ Fractal Vs decay. About 20 Aether quanta, and 17 light. Each time, it drained my light fully before even touching my aether.

[smuglapse]

Made a video showing the Aether bias.

(

)
Messed up in trying to add it to the Elements Youtube group.  Aether is not the last element chosen every time, but it clearly is not preferred by the Pests.

Ok, added another video.  This one shows  vs .  I might add it up later, but the video does a good job of illustrating the issue.

(

)


EDIT:  Removed videos since zanz said the problem will be fixed.

[Hobnob5000]

I'm guessing the variable for the randomness of the devourer is broken. Where the rest has a 0.083 chance of being eaten up, perhaps aether has a 0.0083. That would make sense.

[Demagog]

Did you pull those numbers out your rump or are they based on something? I'm not arguing, I'm just wondering.

Has anyone tried this with something other than aether? Like a fire/light duo, or anything really. I'll try to remember to self test this (duel myself on two separate accounts) sometime soon.

[smuglapse]

I'm guessing the variable for the randomness of the devourer is broken. Where the rest has a 0.083 chance of being eaten up, perhaps aether has a 0.0083. That would make sense.

That makes a lot of sense.

Did you pull those numbers out your rump or are they based on something? I'm not arguing, I'm just wondering.

.083 is just 1/12 and the other number is a decimal place off.  It would explain why the Aether is not completely rejected.

Has anyone tried this with something other than aether? Like a fire/light duo, or anything really. I'll try to remember to self test this (duel myself on two separate accounts) sometime soon.

I tested against other duos and they seem to work fine.  Its real easy to see by testing the AI in the trainer.
You can use this or something similar:

Code: [Select]

6rm 6rm 6rm 6rm 6rm 6rm 6rn 6rn 6rn 6rn 6rn 6rn 6u7 6u7 6u7 6u7 6u7 6u7 77k 77k 77k 77k 77k 77k 7ai 7ai 7ai 7ai 7ai 7ai 7js 7js 7js 7js 7js 7js 7t6 7t6 7t6 7t6 7t6 7t6 808 808 808 808 808 808 808 808 808 808 808 808 80i 80i 80i 80i 80i 80i

[Hobnob5000]

Yeah, 0.083 is 1/12, so maybe aether has a 1/120 chance. (in a rainbow)

I don't see the bug with any other quanta, as when I played my rainbow (before I saw the awesomeness of RoL/Hope), aether was always last, the rest was eaten up equally (more or less)

[Clathius]

I've notice a similar bias as well.   It becomes very obvious when you play duo-decks.   Specifically when I play my graboid deck and the pest just keeps taking the  turn after turn.   It is not uncommon to see them steal the same quantum 5-6 times in a row.  Often the game ends before the pest switches to the other quantum.   The record I have seen is 11 of the same quantum in a row.

I am glad someone posted about this.   I've given up complaining about these types of things in free games because people just rush in screaming cognitive bias and flame the OP with out even taking the time to investigate if the OP might be correct.   Too many people think because they took a stats class in college that online games are immune to bias.   

[zanzarino]

A bug is currently causing Devourers to pick aether only with a 1/144 probability. Will be fixed in ver. 1.24

[Hobnob5000]

G'day Zanz, and nice work on spotting it. Now it will make facing Decay with my RoL / Hope deck about 12 times harder

[Demagog]

It's nice to know that it really wasn't just all in our heads. Thanks for looking into it, Zanz, and getting it fixed.

[smuglapse]

It's nice to know that it really wasn't just all in our heads. Thanks for looking into it, Zanz, and getting it fixed.

Seconded