WTC Stats

No, he's saying there might be other factors. Like how good players are.

I see. Maybe I'm not clever enough for this, but please bear with me: Why do you attribute player wins solely to player skill when faction choice having an influence on this can certainly not be ruled out? Or is that not what you are doing?

Or in other words: How do you decouple the factors of player strength and faction strength in their contribution towards a players wins?

Just for the sake of argument, say you're playing an FPS and all guns were roughly equal in power, but one gun scares the shit out of everyone because in the right circumstances it will almost always one-shot you. Say you're a pretty decent player and you know that no specific weapon will give you any actual advantage, but one of them will give you a psychological advantage because half the players will just run away from you if they see it, which do you choose?

Now, I'm not saying this is 100% applicable, but the internet echo chamber has put Cryx forward as the boogeyman of the meta, and thus, if you are already a very good player and it doesn't really matter what you play, you could use this as a psychological advantage. I know that Sascha, who played Cryx in the winning German team, also played a lot of Skorne in the past few years. I believe their Skorne player was also undefeated at the WTC, so Skorne can't be too bad, but Sascha chose to play Cryx, possibly because having a Cryx player in the team gives you a psychological edge during matchups that having an extra Skorne player doesn't give you.

Personally I played Cryx at the WTC and went 3-3, every game that I lost came down to player skill, my opponent simply played better than me and/or I made a critical mistake. Cryx may be slightly above the curve, but it's definitely beatable.

professorlust If I'm reading your last post correctly, you're saying that, while teams with Cryx players in them won more often, they often won by a smaller margin (3-2 rather than 4-1 or 5-0), or are you saying that the Cryx player actually lost often while the team took the win?

Do you have access to pre-WTC player surveys about faction selection or at the very least records of which teams chose which match ups?

I'd love more data!

I just don't understand. Most of the teams had one or more Cryx players. So this analysis should probably look at Cryx teams vs. non-Cryx teams and then is the amount of data enough for statistics?

IMO the results of the regression are not reliable because of multicollineraity. Player win rate and faction are likely to be correlated. The coefficients for these variables obtained from the regression are likely to be unreliable.

Okay so one final round of rough updates.

Firstly, in an attempt to disaggregate the effect of any given player playing cryx and look to the effect of having a cryx player on a team, I created an ordinal variable based on the number of players playing cryx on a given team (0, 1, 2). This gives an .052 R^2 p<.000 with a B of .452 for total Team wins at the WTC. When we add in Personal Total wins, the model improves to .314 R^2 p<.000 with a B of .320 for Cryx players on a team and .335 for Personal wins. What's really interesting is that when we look at Standardized Coefficients, Having more cryx players on your team has less effect on the standard deviation (.161) while Personal wins increased Standard beta to .408.
So what the above means is that in the first model we can explain 5.2% of total variance in team wins if we know the number of cryx players on a team and that we can say with mild certain that more cryx players on a team had a small but positive relationship with a teams total wins.

The second model indicates that we can explain 21.4% of variance in team wins with we know both the number of cryx players on a team and the number of matches each player earned at the WTC. The unstandardized Bs for both Number of cryx players and Personal wins is fairly low at .320 and .335 respectively. So while these are both significant, they also both had a very small impact on the number of rounds a team won at the WTC. The Standardized B however is interesting in part because it indicates that changes number of cryx players on a team (ie going from 1 to 2) had a smaller impact on teams total wins than the changes in the number of matches an individual player won (ie going from 3 to 6).


However, Team wins while determining the final victor and each teams final rankings, thats not the only way to evaluate how well a team did. The other was However, when we look at the total number of individual matches a team won at the WTC, Cryx's impact on the WTC 2017 meta, is more noticeable.Just a basic regression between number of Cryx players on a team and a teams total wins , the R^2 is .106 at p<.000 BUT the unstandardized B is 1.982! this means that though we can only explain about 10% of total variance in team total match wins by knowing the number of cryx players on a team, we can say with some confidence that each cryx player on a team lead to nearly 2 additional match wins for the team as a whole over the course of the WTC.

Next I ran a stepwise regression modell. Stepwise models differ from normal "enter" models of linear regression in that they test for the individual impact of a set of variables on the model as a whole and add them in and potentially kick them out if they aren't significant contributors to the model. I set the probability of F for entry at .01 removal at .05 which is a bit below the most extreme standard but of entry at .001 and removal at .01 but good for which I analyzed how a Teams Total number of match wins a the WTC was impacted by a) the number cryx players on a team, b) a player playing cryx compared to any other faction, and c) each players individual win totals. Again collinearity was present but with VIF of 1.0 in model 1, 1.027 in Model 2, and 1.116, 1.133, and 1.225 in Model 3, all well below accepted levels of collinearity.


In the stepwise model, individual wins total had the largest impact on a team total wins with model1 R^2 of .248 and b of 1.25, or to put it another way each win for a given player is generally associated with rest of the team team getting another .25 wins. This isn't really surprising given that it only takes 3 match wins to win a match but there are 5 matches total. so a B of 2 would indicate that everyone time anyone won an individual match, two someone else on on their team won also. But given that 5-0s and 4-1s along with 3-2s, and of course their inverses, simply winning yourself doesn't mean that the rest of your team did. Standardized Beta is .498.

In the second stepwise model, the total number of cryx players on a team is introduced and increases R^2 to.309 (p<.000) with Individual wins having a B of 1.153/Beta of 4.98 and and Number cryx players having a B of 1.526 respectively. Whereas we talk about individual wins being associated with team wins (ie each individual win was associated with another .153 wins team wins but they did not cause it), the B for number of cryx players on a team is probably best expressed as each player playing cryx added 1.5 wins to the team.

Playing cryx return in the third stepwise model, we now explain 32.7% of the variance in team match wins. More importantly playing cryx becomes significant again (p<.000) but this time its B is -1.518, while the number cryx players on a team has a B of 1.79 while Personal Match Win Totals has B of 1.253

So what this means in while Personal Match wins has a significant effect on the total number of matches a team one at the WTC, the number of cryx players on a team had a larger effect. However, simply playing cryx had a negative impact on a teams total match wins.

However before you think this seems really contraindicated, its important to know why would simply playing cryx have a negative impact while having one or more cryx player on your team would be positive. Namely that since this is a team event, focusing only on Cryx players vs everyone else ignores the strategic aspects of matchup selection.But by examining the impact on your teams wins, your wins, and the sum of all your teams individual wins of through effect having a cryx player on your team, we can maybe get a glimpse at the strategic importance of having two players playing the same "question asking" faction on the same team.