vdbergh
vdbergh
> The fact that green workers were purged is weird. Green workers can in principle be purged if the p value is very low. But if all workers are green...
It might be interesting to know how much more purging would be necessary for the p-value to go up.
We could. But perhaps @ppigazzini can transfer the db to DEV?
@vondele I browsed around on the RPI repository. At first sight this seems to be the source code http://sourcearchive.raspbian.org/main/c/cutechess/ It is from 2013. This is not recent enough.
It might be helpful to post the output of ``` g++ -Q -march=native --help=target ``` and ``` clang++ -E - -march=native -### ```
We _could_ specify bounds so that we get appropriate resource usage for some reasonable draw ratio & rms bias and then internally rescale the bounds depending on the actually measured...
@vondele If we use bounds of the from `[0,u]` then we could make it so that if the test passes the `p` value would be `0.05`, `0.01` or whatever (recall...
Ok I see it a bit clearer. Instead of specifying explicit bounds [l,u] it is possible to specify the p-value of a passed test and the worst case expected number...
> > it is possible to specify the p-value of a passed test and the worst case expected number of games. Internally Fishtest would then calculate the bounds based on...
@vondele wrote > The numbers and probabilities must be such that other statements are true. Like 'A gainer should be 99.8% sure to be no regression' This would be `p=0.002`....