(Note: The quality of this article has been partially affected by flu. Apologies from the management.)
On the last nationally televised game of Overwatch League’s Stage, a worldwide audience finally got to see the expansion Washington Justice win their first match. Even more exciting was seeing Gi Hyeon Chon (Ado), a 2nd-year player, end the league’s longest active losing streak by earning his first W in 37 tries. His former team, the Shanghai Dragons, got off their own schnide two weeks prior by winning their first three maps in their match. With a now 20-team league there are winners everywhere.
Except with the Los Angeles Valiant.
After a great 2018 campaign the team has fallen off the table and is 0-7 after Stage 1. Despite all the excitement and happiness surrounding the new Overwatch season, one has to worry if the Valiant is going to be the 2nd team in OWL history to go 0-28.
The short answer is “No.” The long answer is below.
Last year we talked about a metric called Expected Wins, which was derived from a team’s win percentage on individual maps. It stood to reason that the more maps you win, the better chances you have to win a series. Lucky (or unlucky) for the Valiant, their winning percentage on maps is already light years ahead of where Shanghai was in 2018.
The Valiant Dragon Slayer
LA 2019 vs. Shanghai 2018 in Pythagorean/Expected Wins
| Team | Yr- Stage |
MP | W | L | T | Exp W/ Match |
Exp W/ MP |
| LA Valiant | 19-1 | 7 | 10 | 19 | 2 | 0.253 | 1.77¹ |
| Shanghai | 18-All | 40 | 21 | 141 | 2 | 0.043 | 1.70 |
| Shanghai | 18-1 | 10 | 6 | 36 | 0.048 | 0.48 | |
| Shanghai | 18-2 | 10 | 2 | 38 | 1 | 0.010 | 0.10 |
| Shanghai | 18-3 | 10 | 9 | 32 | 0.107 | 1.07 | |
| Shanghai | 18-4 | 10 | 4 | 36 | 0.025 | 0.25 |
¹-A new exponent was derived from the 2019 Stage 1 rankings and used for the 2019 stats in this article. If you’re interested, it’s 1.81 vs 1.67 for 2018.
In just 7 matches, LA has blown past the metrics set by Shanghai in 10 or 40 matches. At this pace, LA would be expected to win 7 matches in four stages, not enough to be in the hunt, but certainly better than who they’re looking up at in the standings.
Better Than Bad, But Not Good
The Bottom 10, Ranked by Expected Wins
| Team | W | L | T | Diff | Exp W | Actual |
| Guangzhou | 15 | 16 | 0 | -1 | 3.30 | 3 |
| Houston | 13 | 16 | 1 | -3 | 2.87 | 3 |
| Shanghai | 13 | 17 | 0 | -4 | 2.67 | 3 |
| London | 12 | 16 | 2 | -4 | 2.67 | 3 |
| Hangzhou | 12 | 16 | 1 | -4 | 2.64 | 3 |
| Chengdu | 12 | 20 | 0 | -8 | 1.99 | 3 |
| Paris | 9 | 17 | 2 | -8 | 1.80 | 3 |
| LA Valiant | 10 | 19 | 2 | -9 | 1.77 | 0 |
| Washington | 9 | 20 | 1 | -11 | 1.39 | 1 |
| Florida | 9 | 21 | 1 | -12 | 1.30 | 1 |
Bolstered by three matches going to a 5th match, LA’s -9 map differential is only 2 notches worse than the minimum to by an 0-7 squad. This puts them solidly ahead of the gawd-awful Washington and Florida, and into the group of just-bad Chengdu and Paris.
So, while they certainly aren’t as bad as Shanghai, they do still have the possibility to go 0-for-2019. But how realistic is that? I figured out a tree diagram for every potential outcome, just like how you plot out successive coin flips, modifying for the fact that Control maps can’t end in ties, and I determined that based on the Valiant’s map winning percentage², they have a 24.2% chance to win a game against an average opponent (so, Dallas). Run that into the binomial distribution and that means that LA has a 0.03% chance to lose out, or something that will happen only once every 336 times. Not only would the Valiant have to get much worse to match Shanghai’s feat, they’d also have to be incredibly unlucky.
²-For the non-Control rounds, each outcome was weighted by the frequency of that outcome in all maps. For Control, I simply assumed a draw was worth a half-win. Ideally, I should be doing probabilities for each game type, but neither the Overwatch League site or my other favorite compendium of info Liquipedia has this breakdown readily available and there’s only so much I’m willing to do myself for free.
