After an exciting round robin phase, we have our top 4 for the playoffs. Here is the Mid Season Brawl Bracket, from the BlizzHeroes twitch stream. I number each game. These numbers are used in the tables below for odds of each team making it to and winning each game. Stop by this Reddit thread and share your predictions!
This is a Phoenix Bracket. In the above, the upper bracket is the “winner’s” bracket. A team that loses in the upper bracket falls to the lower bracket. The Loser of Game #1 goes to Game #6. The Loser of Game #2 goes to Game #7. The Loser of game #3 falls to Game #9. The winner of Game #9 then goes on to the championship Game of #10.
All series are best of 5 except for the championship game which is a Best of 7. However, in the championship game, the winner’s bracket champion (winner of Game #3) starts with a 1 game advantage, and will only need to win 3 maps to win the championship. The winner of Game #9 will need to win 4 maps.
It is theoretically possible for any team to play any team. As such, I can calculate the odds of one team beating another using the Elo scores. Here are the odds therefore of each team beating each other team. In this table, teams are sorted by their current Elo.
"Example, the odds of Nomia beating MVP Black in a single game are about 10%. Note that this is for a SINGLE game. In a Best of 5, these numbers trend away from 50%. The odds of one team beating another in a Best of 5 are below:
As you can see, while Nomia may have a 10.1% chance to beat MVP Black in a single game, in a best of 5, those odds discrete significantly.
Using the above odds, as well as this bracket:
I can then predict the odds a team makes it to a particular game, and wins it, using Monte Carlo simulation.
Here are the odds a team plays in each game (not wins, just makes it to that game). Again, teams are sorted by current Elo (noting that L-5, despite being 3rd, is in the lower bracket).
The grayed out cells are games where it is impossible for a team to participate it by the structure of the bracket. For example, no team in the lower bracket can play in Game 3. Note this does not distinguish HOW teams get there. For example, in Game #10, a team can get their EITHER by winning Game #3 or Game #9. Either way, they made it there.
Analysis wise, we can see that the teams in the winner’s bracket have a significant advantage, which isn’t surprising, since they can afford to lose once. More surprising is just how far seperated the bottom 3 are from the top 5 in terms of chances to get to Game 9.
Of course, getting to a game and winning it are two different things.
Here are the odds for each team winning each best of 5.
The most interesting column here is the column for Game #10. This is, effectively, the odds (according to Elo) of each team winning the championship. MVP Black has an astonishing 65.3% chance, according to simulations. Fnatic is a distant second at 20.4%. From there, Dignitas, E-Star, and L-5 are sub 10%. In fewer than .1% of all simulations do any of Roll 20, Tempo Storm, or Nomia win the championship.
The odds of each team winning a given match are somewhat of a weird metric to explain. In order for a team to win a match, it has to go to that match. For example, MVP Black only has an 11.3% chance to win Game #7, but that doesn’t mean they have an 88.7% chance of losing Game #7. Rather, MVP Black only goes to Game #7 13.5% of the time. Therefore, it’s more interesting to look at not the odds that a team wins the match, but they odds they win the match IF they go to it.
The odds a team wins a match if they participate in it are stated in the table below. However, if a team has a less than 1% chance of going to a match (example, Nomia going to the championship game), that number is not included. This is because this number is unstable due to being such a rare occurrence.
This is where you can see where teams are likely to drop out. For example, Tempo Storm, even if they win Game #4 and advance to Game #6, have only a 6.5% chance of winning Game #6. Nomia has a less than 1% chance of even winning their first series against L-5.
The “chalk” bracket is the bracket where all favorites win (in this case, the Elo favorite). Here’s what that would look like:
- Game #1 - MVP Black beats Team Dignitas
- Game #2 - Fnatic beats E-Star
- Game #3 - MVP Black beats Fnatic
- Game #4 - Roll 20 beats Tempo Storm
- Game #5 - L-5 beats Nomia
- Game #6 - E-Star beats Roll 20
- Game #7 - L-5 beats Team Dignitas (note that while this is less likely in the chart, this is due to the preconditional odds. If this game happens, despite the 0-3 record in the round robin, L5-would still be favored by Elo.)
- Game #8 - L-5 beats E-Star
- Game #9 - Fnatic beats L-5
- Game #10 - MVP Black beats Fnatic
Follow us on Twitter (https://twitter.com/gfesports) and on our website (https://gfe.gg/) to stay up to date on everything GFE. Also, give us a like on Facebook at (https://www.facebook.com/galeforceesports) and subscribe to our YouTube channel at (https://www.youtube.com/c/galeforceesports).