So that selection bias seems to project that buying before a share split is actually not an educated+good decision
That being said, holding before the announcement performed the best.
There's a bit of time travel / survivor bias with this one. A company that has not beaten the market is much less likely to split its stock. In other words, if I know nothing other than that a company is splitting its stock, I can reasonably guess that it's shown good returns in recent history.