Use WEIGHT_WIDTH to index full weight range in RNG #94
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The RNG was using the POSITION_BITS to determine how many total bits can be used in a pool index. This limited the RNG's range to the total number of positions in the pool, rather than the total weight the pool could manage. In previous usage, this was sufficient for total weight as well, but since the system has since moved to handling ETH fully backed pools and denominate KEEP stakes by individual tokens instead of minstakes, the ~2M position limit is no longer sufficient to represent the magnitude of total weights that need to be handled. The RNG now bases the total weight it can index into on the total weight that the underlying sortition tree can manage, which is significantly higher (~4B vs ~2M), represented by the WEIGHT_WIDTH constant.
Implemented two tests for BondedSortitionPool and two tests for FullyBackedSortitionPool distribution of selectSetGroup results. For BondedSortitionPool we are making sure that selectSetGroup result distribution: - is similar between operators with the same stake, - promotes operators with higher stake, For FullyBackedSortitionPool we are making sure that selectSetGroup result distribution: - is similar between operators with the same unbonded value, - promotes operators with higher unbonded value.
|
Disclaimer: the selection result distribution tests are taking a lot of time to execute and we could consider moving them to a separate job. Given that we don't have a lot of active development in this repo right now, I think it's fine to accept the fact the tests are taking longer to execute and revisit it when we'll move |
pdyraga
requested review from
nkuba,
Shadowfiend and
eth-r
and removed request for
nkuba
nkuba
approved these changes
Sep 14, 2020
There was a problem hiding this comment.
The change has been verified by @Shadowfiend @eth-r and @pdyraga. All tests are passing, I'm merging this thing!
nkuba
added a commit
that referenced
this pull request
Sep 14, 2020
Bumping up version to 1.1.2 for a new release The new release will contain two changes: - Use WEIGHT_WIDTH to index full weight range in RNG #94 - Babe the Blue Ox: Add rudimentary buidler setup #93 See https://github.com/keep-network/sortition-pools/milestone/5?closed=1
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The
RNGwas using thePOSITION_BITSto determine how many total bits can be used in a pool index. This limited the RNG's range to the total number of positions in the pool, rather than the total weight the pool could manage. In previous usage, this was sufficient for total weight as well, but since the system has since moved to handling ETH fully backed pools and denominate KEEP stakes by individual tokens instead of minstakes, the ~2M position limit is no longer sufficient to represent the magnitude of total weights that need to be handled.In addition to this fix, we are adding two tests for
BondedSortitionPooland two tests forFullyBackedSortitionPoolcovering the distribution ofselectSetGroupresults.For
BondedSortitionPoolwe are making sure thatselectSetGroupresult distribution:For
FullyBackedSortitionPoolwe are making sure thatselectSetGroupresult distribution:Here are example stats from 2500 selection and 100 operators in a
BondedSortitionPool, where: