MIN 0-0 NYY
TB 0-0 HOU
STL 1-0 ATL
WSH 0-1 LA
TB 1-0 OAK
MIL 0-1 WSH
The broadcasters for the St. Louis - Atlanta game sent me down a bit of a rabbit hole, by simply quoting a stat. As the game neared the end, they stated that the winner of game 1 of the Division Series wins 73% of the time. Specifically, they are 73-27 all time. That's a nice round 100 series. Since the Division Series began in 1995 (so 24 seasons before this one), I can only assume they were also counting the 1981 strike-shortened season's "Division Series" as well. In some sense, there are 3 or 4 different eras there, one with the first half and second half winners, one with a single wildcard and a 2-3 format, one with a single wildcard and a 2-2-1 format, and one with 2 wildcards (and 2-2-1). But in another sense, they are all 5 game series, so you could also include the League Championship Series from 1969 to 1984 in doing some analysis. But let's just take the 73-for-100 number.
In today's projections, I've got St. Louis at 69% likely to win their Division Series, and Los Angeles at 80%. Those are both reasonably close to the cited number. In a perfectly matched series with no homefield advantage, the winner of game 1 would be expected to win 68.75% of the time. If you attach homefield advantage to that (in all of MLB, home teams won 53% of games in 2019, 1286 of 2429 games), the number rises slightly to 68.79% in a 2-2-1 series, and falls to 66.54% in a 2-3 series. The remainder of the difference to reach 73% can probably be explained by the fact that teams are not perfectly balanced, and if the teams are not perfectly balanced, the better team is more likely to win game 1.
Oct 1: Los Angeles over Houston
Oct 2: Los Angeles over Houston
Oct 3: Los Angeles over Houston