StrandedKnight wrote: ↑24 September 2025, 13:08
There are three main problems with the Arena system:
1. Only the end state matters.
2. The very high K factor of 40 leads to big swings.
3. There is no way to filter by Elo (as pointed out, the Elite filter is kind of pointless since even players with the minimum rating of 100 can reach Elite).
Because of this, there is a lot of randomness in who ends up winning the season even after three months of play and 100k games. Remove any of those three factors, and it would greatly increase the chance of the best player winning in the end.
Bogeyun has arguably been the best player this season. He could have chosen to strategically sit on his Arena rating when he was in a massive lead, but didn't, and the system unfairly punishes him for it. Just not a very good system.
I think everyone is making good points in this thread.
Some thoughts:
1) Arena is voluntary. I very much enjoy the dozen games I play without playing any of them in arena. Actually, I play "Earth" in Arena but only because I cannot seem to get a game otherwise.
2) I think most of us care about our Elo, and if we play Arena, we care about our Arena ranking. So, is this system really unfair? Was it unfair what happened to Bogeyun? As stated, Bogeyun could have sat out the remainder of the season and won the Arena. Well, if the purpose of Arena is to win, then why not do that? Sitting our Arena does not mean giving up on one's favorite game. A player can still post open invite games (though maybe like my experience with "Earth", no one would join?) The other way I start games is to send rematch offers to similarly ranked players I have played in the past. So, there are two different ways to get games without using Arena.
3) Yes, the Arena K-factor of 40 makes it swingy. But the regular Elo K-factor of 20 seems low to me. Most gaming sites I've played on in the past use a 32-point system. That's how it worked on [Other gaming platform with the miserable bugs and timers and animations]. So, the 40 K-factor for Arena makes it difficult to stay on top, but it also makes it easier for a player like Bogeyun to get there in the first place. And once a player falls down from the top, the high K-factor makes it easier to get back there, right? Someone like Bogeyun, good enough to get there once, can then get there again.
4) Part of this results from the high variance of TFM. Apparently a high-rated player lost to a player whose Elo was 600 points lower. Fortunately, the K-factor of 20 limits an Elo loss to 20. But, that's the nature of high variance games. I play several low-variance games with near-zero luck. I will never lose one of those games to someone ranked 300 or 400 Elo below me. It's just impossible. TFM, or Earth, or Wingspan, or any number of other games with random card draws, is subject to the occasional unlucky game. Oh well. Nature of the game.
5) As noted, only the end-state matters in Arena. I'd like a different system, which I saw suggested elsewhere or maybe earlier in this thread: Keep the k-factor of 40. Update each player's Arena tokens after each game. But the leaderboard would show each player's current number of tokens, AND that player's highest token reached. At the conclusion of the season, the winner of the person who reached the highest token level, not whoever has the highest token level at the moment Arena season ends. Then, everyone has incentive to keep playing.
So, what is the tactic? to play as much games as possible, to try your luck with the cards, and hope some other good player will be less lucky than you in 90 days? 
