Looking at postseason stats for any player is basically cherry-picking, though. They'd need to have 10+ years in MLB and an inordinate number of postseason appearances to generate a decent sample size to begin with, and any player who does so is not going to be anything like the same player over those 10+ years anyway. Then you start looking at starter vs reliever, home vs road, etc. and it's just a jumbled mess of conflicting info. IMO the most important data by far was the fact that he'd struck out the difficult-to-strike-out Adam Eaton on three pitches 15 minutes earlier to get his team out of a jam in the 7th. He looked like a world-beater.
I do think Roberts still should have gone batter by batter for the 8th, using a righty for Rendon and then Kolarek for Soto. Those were clearly the two biggest remaining outs and the best time to empty the chamber. But it wasn't an indefensible decision. Roberts made more than one decision last night worse than trusting Kershaw there.
Alsdo the pitch to Rendon was a good one too- a breaking ball down and out of the zone. An MVP-caliber hitter just went down and pulled it into the first row. That was a "just gotta tip your hat" moment. The pitch to Soto was bad and Kolarek definitely should've been in the game once Rendon made in a one-run game, but