This is a segment from the Lightspeed newsletter. To read full editions, subscribe. When Anza’s lead economist Max Resnick came on Lightspeed recently, I asked him a question that frequently gets brought up in conversations about Solana’s economics: What should be done about the sandwich attack problem? His answer surprised me. “The sandwiching rate [is] way down,” Resnick said. Solana’s speed makes it an ideal venue for timing games at the expense of unsophisticated users. A sandwich attack happens when a toxic market participant both frontruns and backruns a trade to profit at the expense of the trader. Sandwiching has historically been a big concern for Solana and its developers. In early 2024, Solana infrastructure shop Jito shut down its mempool over sandwiching concerns. A couple months later, the Solana Foundation cut suspected sandwichers from its stake delegation program. Jito’s DAO also voted to blacklist malicious validators from its stake pool. None of these fixes turned out to be panaceas, however, and the sandwich attacks continued. But the story Resnick tells is that sandwich attacking was blunted not by blacklisting bad actors — who may be able to re-emerge with a different online identity — but by improving Solana’s code. Basically, sandwich attacks are numbers games for the attackers. In 2024, when landing transactions on Solana was much more difficult, transactions were sprayed widely to validators in hopes of getting user transactions to the blockchain. This meant toxic validators saw a lot of transactions and had more chances to frontrun them. As Solana’s transaction “targeting” has improved, fewer validators see each transaction on average, so sandwichers have fewer opportunities to mess with transactions, Resnick said.