Games played has a HUGE impact on rank, limiting your maximum ranking points severely up to about 28 games played at which point it reaches 95% of the max possible. After 42 games played it has a very small impact.
playing fewer games is rewarded as opposed to playing more games
That's simply not true. Take a look at the ranking formula.
it's very unclear what would happen if they continued to play and if they would be able to maintain their win-rate in the face of more diverse opponents.
That's not true either.
There was no change in frequency of complaints in S11 that I saw.
As for people complaining, that's not the criteria for better matchmaking. The criteria (in accordance with the principles Microsoft used) is "creates more even matches" and TV+ undeniably does exactly that.
I don't understand why it's considered necessary to handicap good coaches
To create more even matches, and therefore better matchmaking. If we had many, many coaches in the game it wouldn't be needed, but there aren't, so it is. It's worth remembering that BB already has a form of handicapping in the form of inducements.
Prize money is €950 per season.
There are lots of community leagues, so going automatically into an official league is not a good solution. The Cabalvision Official League is the first one in the list, though.
No idea if there will be cross-platform, ever.
No, they aren't. In terms of official leagues there is CCL, which is the main "competitive" league with playoffs and a full reset at the end of every 6 week season and there is prize money; there is COL which is an eternal league with no playoffs, but the rankings reset every 6 weeks like CCL; and there is Anarchy, which is another eternal league which also allows them mixed teams. They serve different purposes and CCL and COL both have pretty active communities on PC. I know it's slower on consoles (PC gets about 4 times the traffic as both consoles combined) , and that's why cross-platform would be great.
I don't see how speaking to the people behind TrueSkill makes TV+ any more appropriate or not. It might have been interesting but it isn't relevant to Bloodbowl.
I talked to them about matchmaking concepts in general, not TrueSkill alone - TrueSkill is not pertinent for Blood Bowl since it is a 2-player game, and Glicko would serve the same purpose. Making fair matches is a central concept to non-random matchmaking.
If fair matches were the only thing that made games fun and interesting then you could reduce the game to a coin toss and that would be more fun and interesting yet that isn't something people find particularly fun.
Oh come on, you're above that sort of reductio ad absurdum. There's more to "interest and fun" than just fairness, but if we take the same basic game of Blood Bowl and can offer either a fair match or an unfair match it's not unreasonable to say that more people would want the fair match. Or would you want the unfair match (bearing in mind you have no idea which side the advantage will go to)?
It's more complicated than that, and involves understanding the way the game is designed and works. You've made a change that results in matches being less predictable and you've made an effort to verify that the change has that affect which is fine. It can't be used as an argument for why that was a good thing in the first place.
The in-game design has some relevance to the method of matching, but it's not central when the most influential metrics are known. That's one of the reasons straight Elo is a good systemic foundation for a number of symmetrical games, and why Elo needs to be adjusted for asymmetrical games. That's what Mike and I have been looking at.
I don't think data on its own can judge whether a greater number of people are having fun or not. Looking at concessions is flawed because people are heavily punished for conceding.
There is no way to tell if more people are having fun or not (certainly not how loudly a few individuals protest), but the blind test we ran, which included people complaining about the matchmaking on stream while TV matching was happening, showed that people simply couldn't tell the difference between TV and TV+ matching. That puts the lie to the "it makes the game less fun" brigade, because they can't tell.
Data can tell you whether more people are getting more balanced matches or not, and unless you are going to claim that unbalanced matches are generally more desirable than balanced matches (which would be contrary to the basic concept of non-random matchmaking and contradicts all the calls we've seen to make the game "more balanced") then that's precisely what the data does tell us that TVPlus does.
But wouldn't that show itself in higher concession rates (which seem comparable)? Or is it just a very small number (compared to the total) of these repeat offenders which seem to be unaware that they're in the wrong league?
The concession rates are higher. There's a 30% increase in concession rates compared with your average season. There are also far more people conceding over the 5-match limit: 250 so far this season (which we are a little over halfway through) compared with 190 in the whole of last season (for reference, 30 of those 190 were in the last week).
Are those who exceed their allotted 5 concessions by far banned longer or are they treated just the same as those just above the limit?
They're treated the same.
Our main issue this season is that CCL has been put as the default league (an error - we tried it before and it doesn't work). When people have to actively select it the concession rate is far lower and it's easier for the admins to maintain.
Are there reasons other than lack of TV+ for S11 having an unprecedented successful end of season, with the flattest drop off rate seen?
Yes, the Christmas break. The same was seen last year at the end of season 2 where there was a marked uptick in games played in that last couple of weeks in comparison with the previous 2 weeks.