A sim and the 10 vs. 10 Part 1: North America
Book a cheap flight?
Thanks for subscribing! If you haven’t subscribed yet, you can hit the button below to do so.
One of the biggest topics of discussion last week was the proposed Street Fighter V 10 vs. 10 by Punk, who said he would fly to Europe and do battle.
That got everyone’s attention. Even I was kind of intrigued by the prospects of such a duel. Usually these types of events are settled at a hotel room on late Saturday night of an event just because someone was on tilt, went 0-2 in brackets, or both.
This is one of those posts that is constructed like a recipe you find online. If you only care about cool pictures and graphs, scroll to The Projections for Team North America to have a look.
Leading up to showing those pictures, I want to explain what got me interested in doing a deeper dive into it.
This is a unique fight proposition, where there’s no neutral site. This is the challenger intentionally being the away team. More than being away, this is a fairly lengthy trip, and who knows what may come out of it.
I thought of the nine players that should accompany Punk on the flight, but then I saw the tweet of the players Punk would bring with him. It’s a list of 10 because of course no FGC list is exactly the number it needs to be.
And then I saw the list that made up team Europe, and I said, “Book cheap flights.” After a pause, I then said, “I don’t think this is Europe’s best team and THEY STILL might not break a sweat until the end.” It might be the pessimist in me. This was before I looked at any stats.
I was confused at the players that were not on the lists on both sides. Perhaps it’s a scheduling issue. Perhaps some don’t have passports or can’t obtain visas. Maybe it’s another issue. But if this is the list of 11 that will be 10, let’s take a look.
There isn’t enough matchup history between all 22 players to do a direct matchup simulation of what could happen. Then there’s also the potential order which obviously would not be revealed until the match began. And the roster might change.
There are so many variables, but it is possible to still do some simulation to get an idea of what to expect.
Here was the plan I executed: run a simulation based on the player vs. the expected characters to be used in the 10 vs. 10 match. Instead of saying that player X would defeat the character a certain percentage of the time, split the rates into potential outcomes.
The amount of games required to eliminate a player was not revealed, so I chose 5 as the target. If a player has a 50 percent chance to defeat a character based on the simulation, how does that 50 percent split into the winning outcomes of 5-0, 5-1, 5-2, 5-3 and 5-4? And then how does the remaining 50 percent split into the losing outcomes of those same scores? This gives us a window into how strong a player could be against a certain matchup.
If that player’s 50 percent chance goes all to 5-0 on the winning side and 0-5 on the losing side, how much faith do you put in that player making an impact on the match?
The calculation for the sim
As I’ve done in previous simulations, I used a two-factor process to create the percentages.
I started with each player’s winning percentage based on the four possible round scores within a game (0-0, 1-0, 0-1 and 1-1).
The stats used were based on the following leagues and events, a total of 26,200 rounds used:
Capcom Pro Tour Top 8s 2016-21
The Olympics (Intel World Open) 2021
Gfinity Elite Series 2017-18
Red Bull Kumite 2016-21
Street Fighter League 2018-21
As a second factor, I used their strength of schedule based on the opponents they have faced. A player who goes 100-0 in Round 1 against 20 opponents might seem insane. But if those 20 opponents collectively have gone 0-2000, that means the player has faced cupcakes.
In this simulation, the player strength of schedule was a player-to-player comparison, the character strength of schedule was a character-to-character comparison.
After calculating both factors together, I ran a 9-game series (First to 5) where the series ends when one player wins 5 — in the past I would have run all 9 games regardless and chalked anything at 5 or greater a win.
The process is repeated 10,000 times. After that, there is a tally of the 10 possible outcomes a player picked up against the character.
This is repeated for the 11 characters a player could face.
Here’s an example of one run of the simulation, of Punk facing a Menat user.
Overall, Punk, based on 10,000 simulated matches played, would win about 70.8 percent of the time. Of that 70.8 percent, this is the expected breakdown of what the score could be.
The most likely outcome would be 5-2 in favor of Punk, at an 17.4 percent rate. But it could also be 5-3, which is 17.3 percent. It’s possible he could 5-0 a Menat user, that’s a 7.2 percent chance. It’s also possible he loses 5-0, a 1.2 percent chance.
If the line in the 10 outcomes was straight, that obviously means a player has a 10 percent chance of any one of the outcomes to happen.
Here’s how to read the lines as you go through them below:
The scores are at the bottom, the left side scores are the player winning, the right side scores are the player losing. Look for the apex and the sides of the bell curve. Left is great, right is bad.
Heavy left: That player should get a first-class flight paid for.
Left-center: A good advantage but some potential for an upset. That player should get priority boarding.
Center: The player might have an advantage, but it is very shaky, and the littlest of things might turn a win into a loss. Get a normal seat.
Right-center: Any advantage they have is extremely fragile.
Heavy right: Umm … just don’t be late to the match?
The projections for Team North America
The characters North America are expected to face are Akuma, Cammy, Dhalsim, M. Bison, Falke, Karin, Menat, R Mika, Rashid, Seth and Zeku.
The 10 proposed players to be on the team with Punk are Brian F, Chris CCH, Idom, JB, MenaRD, Nephew, Rob TV, Shine, Smug and Samurai.
I’m not going to make a hard line on who should be cut. My initial idea of this post was to look at who should be on the team, but I’m not going to do that. Rather, if a player is on the team, the others need to know how to maximize everyone to get wins.
The initial analysis is that it does not look good for Shine. He’s projected to be an underdog against the characters that would be used by Team Europe.
This is where team North America needs to find a line and play to that one character. That line looks to be either Cammy or Dhalsim. He’s a 44.8 percent chance of winning against Cammy, 43.1 percent against Dhalsim.
Splitting those lines further, he’s about a 12.5 percent chance of winning against either Cammy or Dhalsim by a 5-3 score, then a 13.5 percent chance of it being 5-4.
If we think of this like baseball where a team has to get 27 outs among its pitchers, team North America is looking to get 1 of its 10 outs from Shine somewhere. That opportunity comes against Hurricane (Cammy) or Monsieur Crimson (Dhalsim) if they end up being on the team — remember that this computer simulation is against any user of that character and team Europe might want to change their roster. So when Team Europe deploys either player, Shine is the one on deck.
If Shine nails the 1-in-4 opportunity as the computer says is possible, that’s a heavy advantage toward team North America.
Brian F is leaning this way as well. There are some must-stay-away matchups but there are some where the coin flip could go his way. Again, team North America has to find a line and play to that.
The line for Brian F appears to be R Mika, which would be Luffy if he’s on the team.
Previous stats don’t look good for Brian F against an R Mika, he’s 0-2 in two previous CPT Top 8 matchups back in 2017.
In this simulation, the computer says Brian F is at 49.2 odds to beat an R Mika user. This line is deep in the center:
5-2 win: 10.9%
5-3 win: 13.8%
5-4 win: 13.9%
5-4 loss: 14.0%
5-3 loss: 13.8%
5-2 loss: 11.8%
That’s very close. As we saw in Red Bull Kumite, Brian F just needed to have a good day to get to the playoff round, and it happened. In this 10 vs. 10 if he’s having at least a decent day, he just needs to get that 1 crucial win for Team North America. Line him up against a potential R Mika and hope the 50-50s go his way.
Even if Brian F loses whatever match he does get put in against an R Mika, the projection says the match is expected to be a deep one, perhaps all 9 games if each series is a first to 5 wins. Brian F making his opponent spend a ton of mental energy on his matchup could mean a quick following win for his teammate who follows.
The Smug projection surprised me. He’s just as likely to 5-0 someone as he is to lose 5-0. It could be an epic outing or a quick one. This lead me to believe the idea spot in the lineup is first instead of someone like Chris CCH or Samurai. A potential plan is to align Smug and JB back to back to start unless the aforementioned scenario of Hurricane or Crimson go up in which Shine gets the nod.
In the next issue, I’ll go over the prospects for Team Europe.