[sigurpol]

Kusama/Polkadot use U128CurrencyToVote which scales stake based on total issuance: vote = stake / max(issuance/u64::MAX, 1). If total issuance exceeds u64::MAX, scaling factor > 1.

Previously the tool used stake.active directly, causing stake weight discrepancies on Kusama compared to on-chain snapshots.

[sigurpol]

an attempt to fix the issue, @martinmsb @filippoweb3

[filippoweb3]

@sigurpol thank you for the suggestion. It seems these changes do not change the simulations, i.e. built snapshot shows the tiny differences in stake compared to the fetched one, and predictions using the fetched snapshot are still inaccurate

[sigurpol]

Yeah I got it wrong, double checked and

• Relay chain: type CurrencyToVote = U128CurrencyToVote (scales by issuance) => the old legacy staking
• Asset Hub (parachain): type CurrencyToVote = SaturatingCurrencyToVote (ignores issuance, just saturates to u64)
  fn to_vote(value: B, _: B) -> u64 {
  value.unique_saturated_into() // ignores issuance!
  }

Dropping this PR