Will the Dodgers Ever Lose? Are the Braves Screwed?
Your guide to early MLB records.
I’ve written before (a lot) about the meaning of early-season MLB records, but it’s always a fun topic to revisit, because we all love to obsess over the first few weeks of results after months without baseball. And this season has given us plenty in that department already.
Most notably, the Los Angeles Dodgers (7-0) and San Diego Padres (6-0) are both undefeated, while the Atlanta Braves are winless (0-6, all against the Padres and Dodgers) and others such as the Milwaukee Brewers (1-4), Boston Red Sox (1-4) and Minnesota Twins (1-4) aren’t looking much better. With extreme cases like these, it’s understandable to already be asking what the benefit — or damage — will be to a team’s eventual record after just a few games.
We know that every win removes one more possible loss from your potential record, and vice-versa with losses and wins. But as my colleague Ben Morris wrote about with the NFL here (in one of my favorite early FiveThirtyEight posts), early-season games carry more importance than you might think, as they both add to the record that a team has “banked” away, and give us fresh information about a team that we hadn’t quite seen play together before the season began.
MLB has its own version of this double-effect. In full seasons since the 1994 strike, teams who start the season 1-0 go on to average 83 wins, while teams who start 0-1 average 79 wins. That’s a 4-win gap — twice as large as the simple 2-win swing you'd expect from one result in a 162-game season (i.e., 1 win above or below .500).
This effect is amplified further in a very short period of time; for instance, teams who start 6-0 like the Padres win 91 games on average, while 0-6 teams like the Braves win just 69 games. Here’s a heatmap of every possible season start, up to 25 wins and/or losses, along with their eventual wins for the entire season:1
The general pattern is clear, that you can put yourself on a notably good (or bad) pace surprisingly quickly, if you start with an outlier record in either direction. There are some funny anomalies in the data, however — teams that moved from 10-2 to 11-2 actually saw their expected wins drop by 2, while teams who fell from 10-2 to 10-3 improved by 4 wins, for whatever reason. (Only 16 teams in the sample ever started 10-2 anyway, which tells you about how weird and fluky early records still are.)
If we try to strip away most of those weird cases — a team’s net swing in expected wins with another win shouldn’t be negative — by raising the sample size required to appear in the chart, here are the pivot points in a season that have the most “leverage” in terms of swing in net full-season expected wins with a win in the next game, versus a loss:
Some weird outliers remain, but in general, many of the highest leverage games are concentrated surprisingly early in the schedule, among teams in need of a win because their records might otherwise begin to spiral into a true disaster. For instance, the swing between going 8-11 and 7-12 was six net wins in terms of full-season expected record — which is probably itself an outlier relative to surrounding entries, but also indicative of how a bad start needs to be stabilized quickly. As the season progresses, there are still opportunities to swing your eventual record by four or five wins by winning the next game, but they become rarer than they are early in the year.
None of this is to say that the Braves will go on to match the eventual average of 69 wins for teams who started 0-6. They have faced the most difficult early schedule of any team, and they were expected to be better than average going into the season — even though I already had questions about them, and that was before they lost Jurickson Profar to a PED suspension this week. For a more precise estimate of exactly how much starting 0-6 has cost the Braves, that’s where something like my ⚾ 2025 MLB Elo ratings and win projections📈 come in handy — and it has them on track to win about 82 games, which isn’t great but is better than 69.
Still, there’s value to looking at how the typical team with your squad’s current record ended up finishing the season — and how that might change with just a few more wins (or losses).
Filed under: Baseball
Starts that had fewer than 5 instances since 1996 were excluded. Also, we’re throwing out data from the weird 60-game season of 2020.
100% hit rate on you linking to a Benjamin Morris article and me going to read it (almost certainly for the fourth or fifth time), then having to stop myself from reading all of his stuff again