Other sports leagues that use SRS can go about it in a fairly straightforward manner. They’ll compute it for the regular season, lumping it with other regular-season statistics, such as wins, points scored, and so on, and then not keep it for the post-season. That’s how most leagues work with their stat-tracking in general. Everyone knows, for instance, that Babe Ruth hit 714 home runs … unless you count his postseason home runs, in which case he hit 729. But we don’t do that. We typically separate regular season and playoff stats.

The Overwatch League complicates that system. For the first two seasons, there were stage championships sandwiched in during the regular season, which were considered to be “playoffs” that didn’t count in the regular-season standings. The standings will say that San Francisco had a 23-5 record in 2019, but in reality, its record for the entire season — including stage and end-of-season playoffs — was 35-8.

In the attempt to keep things separated, as other leagues do, my standings have only reflected regular season performance. Thus, in 2019, San Francisco was 23-5, with 92 maps won and 28 maps lost, a +66 differential, a 2.40 SRS and .785 SAMP. It’s those last two numbers, and their related advanced stats, that make things a little tricky. San Francisco’s playoff positioning was determined by its win/loss record and differential. I think we can all agree that the team’s performance in stage playoffs shouldn’t have had an impact on that.

However, when you’re talking about advanced stats that are meant to explain a team’s true performance, then there’s really not much reason not to include in-season, and probably post-season, playoff performance. Well, there is one reason, and the one I’ve been mostly using as my reasoning for not including the playoffs: it’s hard and the math is a little counter-intuitive.

With the recent May Melee tournament making a mess of my data, as well as the general craziness of the schedule this year, I think I’ve been left with little alternative other than to finally implement full-season advanced statistics. Those are now in place for the 2020 standings in the following categories: MOV, SOS, SRS, and SAMP. I always used all of a team’s game to compute its Elo rating, so that’s unchanged, but you can think of the final five columns in the “Expanded Standings 1” chart as representing a team’s full-season statistics, if you want.

Keep in mind that these are the only places in the standings where full-season stats are used. Wins and losses, maps, differential, map type records, reverse sweeps … everything else represents just regular season totals. This leads to some of the “counter-intuitive” talk I mentioned earlier; a team with a differential of +0 should have a MOV of 0.00, but that might not be the case once playoff maps — which aren’t represented in differential — are figured in.

I might change my mind later when it comes to some of the split stats, like W/L record on Saturday or on Assault maps, if it seems more logical to include playoff stats there. The reason I’m hesitant to do so is because I feel that W/L records shouldn’t be unduly affected by schedule strength. Games against playoff teams are typically tougher than they are against regular season opponents, so if a team is 3-6 on a map type, is it the case that it’s bad or that it played against better-than-average — i.e., playoff-caliber — teams? SRS and SAMP take team strength into account, so that’s not an issue for those stats.

For now, these full-season stats have only been implemented for the 2020 season. At some point — after the season, most likely — I’ll go back and compute them for 2018 and 2019. As always, let me know if you have any opinions on the new system!