By the way, I believe there is no need for rosters. It should be possible to sub in a player from the same country between rounds at any time. The roster artificially adds a planning requirement which goes against Rento's point that people just want to sign up and play games. What was the original intention behind the roster requirement?

Probably necessary for game creation not to be complete chaos I would venture to guess.

Wouldn't a rule like only people who sign up on the clot in advance can play do the job?

Actually I believe the CLOT wasn't used in Nation Cup, as it wasn't a requirement to be signed up to be part of the rosters and there is no NC tournament when you go on the CLOT website. My guess is game creation went through the game's API directly, and having a roster is not completely necessary but it does formalise things, and you're going to need a list of players at some point. It's not a game mechanics constraint, it's an organisational ones.

This answer relies on many assumptions on subjects I know close to nothing about.

A wise man once said: "Sometimes it's better to say nothing." An even wiser man refrained from saying that.

Si tacuisses, philosophus mansisses.

Matchup uncertainty due to unfinished games
- The easy fix here is to decrease the banked time, 6 days was unnecessary

The absence of any elimination round that makes it possible to become the winning team without having played against the runner-up team
- I like your format well enough for this; the swiss format also means that as soon as you lose a matchup you're essentially eliminated and people lose interest. Fighting for top 2 in a group helps this a little bit.

Matchup randomness within teammates (the best player can be wasted against the weakest players of the opponent team and lead to a different result)
- I think using MTL ratings is flawed due to the fact that not everyone plays it, but also some players have a rating and have never pushed for/played consistently enough for a real peak rating (I'm in this group). I get the idea, but due to those two factors, I don't think it's significantly better than random. I don't think there's a great solution to this other than random or making lineups that are lightly screened.

Thanks Bonsai! I agree that removing the banked time would be an improvement, although you could still have one of the best teams in a 20+ turns game in which one of the players (let's call him T*m) takes three days every turn. We don't want to wait for two months for the next round if this is the decisive game, right? :)

Yes, that's fair. Probably reducing banked time eliminates most of the problems with slow play, but not legitimately long games.

What boot time should be used instead?
< 1 day too quick (higher chance of boot because of schedules)
seems like your implying >= 3 days too slow

I don't think a deviation from 3 days makes sense, it's the accepted standard and having individual games with different boot times from that can be annoying as you have to adapt your turn taking schedule and get a lot bound to the game. That ruins the fun for me personally, that's also why I don't play Fizzer's new ladders. I'm generally a fast MD player, but having the option of 3 days for individual turns is super important to me on days when I can't open the game.

My proposed system would solve this concern irrespectively of the boot time anyways, because matchups during the group phase will not depend on game completion.

Some good ideas here.

I think you have to scrap the MTL board requirement though, and really just put it up to organizer discretion to shut down blatant gamification. Not enough players have a reliable enough MTL rating, and currently it's down indefinitely anyways, but even if it came back it would still be problematic because it's just not played enough.

Warzone has an issue that has been made worse by the new ladders, that there's no good way to objectively rank players. You have to cobble together information from multiple sources to get an idea of how good a player is.
Also template strength is a real factor.

I think your last sentence is very important and explains my stance well. MTL is the only measure of skill that has a high degree of reliability. Also, my experience suggests that in most cases, when it's argued that people didn't play MTL enough to get an adequate rating, the reason is usually that they believe to be better than they are. I often see people try again later and get to the same ratings they had. I was like that as well, years ago. The only people I see misrated are those that haven't played MTL for years, not those that haven't played many games. A couple dozen games are more than enough for a pretty precise MTL rating for most players, maybe except for ultra elite ones.

However, I agree that scrapping that rule and putting it at the organizer's discretion would work too, probably it's the easier way to go about it :)

I think you misinterpreted my stance. It's not that I think MTL is unreliable at ranking players, it's that I think that for many players they either never played it or tried it once at some point in their warzone career and never came back to it. For anyone that plays MTL regularly, or has had a recent run on it of say 30+ games, I trust the rating.

Add team games to nations cup.

Honestly the more I think about group and knockout stages the more I like the idea.

And maybe instead of having a different template every round, we would make it more simililar to Clan League's slots? 6 players per team, 6 different templates, the same ones each round. Each team assigns a player to each template. No problems with randomisation.

That would work too, of course!

while player to template slots solves any matchmaking issues, my concern at that point is that it's too similar to CL.