While I’m having some second thoughts about SAMP‘s ability to predict matches, because it’s based on past, full-season performance, it’s an excellent way to judge how likely it was that something happened during the season. I’ll be doing a few pieces like this in the offseason, but I thought I’d start this blog off with a pretty simple scenario.

The 2018 Shanghai Dragons‘ SAMP was 0.1626. Against an average team (SAMP = 0.5), that means they would have had a 16.2% chance of winning a map and 3.3% chance of winning a full game and a 96.7% chance of losing a game. The chance of losing 40 games, then, is simply .967 to the 40^th power.

With a 40-game season against all average teams, the Dragons would have had a 25.9% chance of a winless season.

But Shanghai didn’t play a season against all average opponents. Its strength of schedule, according to SRS, was 0.27 – the highest in the first season of the Overwatch League. That’s due in part to the fact that the Dragons didn’t get to play against the worst team in the league (themselves) and also due to the relatively balanced schedule in the Overwatch League, which tends to balance the SOS out in inverse proportion to teams’ win-loss record. In a later post, I might go into why this happens in more detail.

Using SAMP, I can estimate Shanghai’s map and game winning percentage against every team on its schedule. You probably won’t be surprised to learn that this gave New York a 99.7% chance of beating Shanghai, which some of you might still feel is too low. The Dragons’ best odds, naturally, are against Florida, which only beats Shanghai 81% of the time.

Multiplying everything together, Shanghai comes up with an 18.1% chance of going 0-40 for the season. That’s low, but not so low that an 0-40 run seemed improbable.

So, who would the Dragons beat? For that, I simulated the season 10 times, using the win probabilities I calculated and a random number generator. Or, to be more precise, I’m about to simulate that; I haven’t actually done it all on my spreadsheet as of this writing. Before I do so, understand that if they played a million games against each other, Shanghai probably would beat New York at least occasionally, and they certainly would have a decent chance of squeaking out some wins against Florida or Dallas. So if it happens, or even if Shanghai wins, say, three games, in one of my 10 simulations, that doesn’t mean everything is wrong. It’s just how things turned out in one possible “alternate universe.” In the “real” universe – an 11^th possibility, if you will – the Dragons went 0-40.

All right, the numbers have been entered, and here’s how it went:

The first thing I looked at was that final “Wins” row. In 2 of the 10 simulations, Shanghai performed “as expected,” racking up zero victories. That’s as close as we can get to the 18.1% prediction. In three other seasons, the Dragons won a single game, so they totaled zero or one win half the time, with a max of four wins in a magical Season #6.

Taking a closer look at #6 reveals some interesting details. The Dragons beat the Valiant(!), Dynasty, Fusion, and Mayhem. The last one is to be expected; of the Dragons’ 18 total wins across all simulations, the Mayhem accounted for 8 of them. The next-worst team in the league, the Dallas Fuel, provided Shanghai with 5 wins. That leaves just 5 more wins that didn’t come against one of those two teams, and 3 of them came in Season #6.

One of those wins came against the Fusion, which, for some odd reason, lost 3 times to the Dragons in our simulations! Admittedly, the Fusion are probably the weakest of the 2018 playoff teams, with an SRS of just 0.27 and SAMP of 0.5307, both eighth in the league. Even so, our simulated Dragons never beat the Shock – compiling an eerily familiar 0-40 against San Francisco across all 10 simulations – which was a worse team than Philadelphia.

The Fusion accounted the only victory for the Dragons in Season #4, the Fuel the only win in Season #2, and the Mayhem the only win in Season #9 – which came in the 37^th game after an 0-36 start. In three seasons, the Dragons beat the Mayhem in their 4^th game of the season, which meant that we wouldn’t have watching their 0-fer for very long.

The real Dragons took teams to a fifth map four times during the season, so with a little more skill or luck, they might have won one or two of those. Alas, that wasn’t to be, but maybe 0-40 wasn’t as inevitable as it seemed.