# Week 18 and 19 In these week I refactored our original BAL work and also ran multiclient kurtosis. ## Refactoring BAL We are now using Dragan approach for bal building.I went through his implementation and made the necessary changes in alloy/evm. We are still passing the bal through `BlockExecutionResult` since it seemed to be the most suitable to me. Integrating Dragan's work was somewhat easy since he included footnote(comment) in blockchain test in revm. We initially found it will be sufficient to get access via with_bal_builder method and field. But our initial run of tests failed. I added some traces back in evm and found the account changes is not getting sorted. Fixing that all the tests passed. The now bal-devnet-0 is built using this. Meanwhile another approach which included having seperate DB to handle BAL related stuff was proposed in revm. We seperately worked on it as its a very different approach and extremely breaking. In terms of refactoring our work it was not that easy. I had to perform lot of state expansion in evm, which was taken over by my fren in reth. We had it all working just fine with some tests failing ( they will be fixed by the eest now eels in the next release). ## Multiclient kurtosis I spinned up kurtosis with all the clients till 30 epochs. The results were fine for reth but after about 500 blocks, nethermind missed all the slots and got seperated from the network. Rimeeeeee also worked on client groups ( reth-geth and reth-besu ) and informed it was fine. ## Next weeks plan * Run kurtosis again * Rebase branches # Resources The new reth [pr](https://github.com/paradigmxyz/reth/pull/19263)