Arena rating decay for inactivity

[msclark]

My question to the developers would be - what does the past data show you? How much do people stop playing and at what level in the standings does that become a problem?

This is the correct comment, in my opinion. Instead of posting a poll asking people what their gut feelings are, do some science and present us with some facts.

What's the true scale of the problem? What proportion of top players don't play in the last N days of a season? Is there a correlation between score/position and activity?

Then ask how your proposal would have affected things in past seasons. You should be able to work this out. Would it have changed many results? Perhaps compare the Elos of the people involved and decide if that change is "fair". For example, if weaker players were winning seasons currently and this proposal meant that stronger and more active players won them more frequently, then perhaps this is a good idea.

I've voted no, but might change my mind depending on the data. I feel like people stop playing for all sorts of reasons and while you might argue that those who do it to park their position at the top deserve to be penalised - and I'm not convinced about this - it does seem like you're throwing the baby out with the bathwater by penalising everybody else too.

[RebaSheba]

I'd be glad to volunteer to do some analysis to address some of the questions that are being raised. Admin, please feel free to contact me if there is a way I might be able to help.

[frogstar_A]

Peak score sounds like the best alternative (if any).

Peak score is fairer than rating decay, and it does remove the incentive to “park” (though that doesn’t necessarily mean that those players will actually carry on playing once they have enough points for a trophy!) however it also could be somewhat unfairly influenced by the order in which turn based games finish.

For example, I play Draftosaurus and games can take up to 2 weeks. I often start 10-20 games at once and then a few weeks later they all finish. My arena season score currently will be based on all 20, which seems fair, but my peak score might be much higher just because all the games I won happened to finish first. So you reward or penalise players for the random order in which their games finish instead of on the whole season!

It would be like saying the winner of a soccer match was whoever led by the most goals at any point in the match, rather than using the final score. Peak scoreis saying being 2-0 at half time means you “win” even if the final score is 2-3 against you!

[euklid314]

I am not a huge fan of all these incentives of board game sites to bind players to their site. Playing games is addictive enough on its own - and this is a problem for more players than you might think.

Negative incentives for non-regular play are even worse, even if they apply only at the competitive level (i.e., Arena).

And another aspect. Stopping play in Arena after a lucky streak is just the correct competitive idea for heavily luck-based games. Even a (slight) point decay will not change this tactics at all.

In a strategic game the best players have a strong incentive to play on, even if they reach a top spot. Because they can expect to win by their superiority. In luck based games this is not the case - at least not in this extend.

Thus, one would need to discuss the rating decay separately for different games.

[Peyrol]

(I assume my earlier post was not posted because I was using a VPN.)

I don't use the Arena much, but IMO the more a rating system is based on sound statistical analysis instead of politics, the more value it has as an accurate predictor of the results of a game between two players. Deciding how many points to deduct after how much inactivity based on the results of a poll does not seem likely to result in a great improvement in accuracy. But hey, show me wrong, post a link.

That said, I approve of the idea of increased accuracy in the ratings. Personally I would like to see some kind of Glicko ratings in place, for all 2P games, not just Arena. Glicko uses two numbers for each player: The rating itself, and a ratings deviation value (RD.) On average, the longer you are inactive, the less accurate your rating becomes. Glicko would not change your rating, but would increase your RD value over time. This would affect the amount your rating would increase or decrease depending on the result of the game and the two values your opponent has.

But I recognize BGA is trying to please as many users as possible. So everyone gets a say on what happens to "their" number. Good luck with that.

[KongKing123]

Peak score sounds like the best alternative (if any).

Peak score is fairer than rating decay, and it does remove the incentive to “park” (though that doesn’t necessarily mean that those players will actually carry on playing once they have enough points for a trophy!) however it also could be somewhat unfairly influenced by the order in which turn based games finish.

For example, I play Draftosaurus and games can take up to 2 weeks. I often start 10-20 games at once and then a few weeks later they all finish. My arena season score currently will be based on all 20, which seems fair, but my peak score might be much higher just because all the games I won happened to finish first. So you reward or penalise players for the random order in which their games finish instead of on the whole season!

It's really no different from what happens now. If you lose 10 games in a row, tanking your rating, and then win the next 10, your final rating will be different than had it been the other way around (also depending on the opponent's ratings): https://stackoverflow.com/questions/239 ... ame-played

[frogstar_A]

Peak score sounds like the best alternative (if any).

Peak score is fairer than rating decay, and it does remove the incentive to “park” (though that doesn’t necessarily mean that those players will actually carry on playing once they have enough points for a trophy!) however it also could be somewhat unfairly influenced by the order in which turn based games finish.

For example, I play Draftosaurus and games can take up to 2 weeks. I often start 10-20 games at once and then a few weeks later they all finish. My arena season score currently will be based on all 20, which seems fair, but my peak score might be much higher just because all the games I won happened to finish first. So you reward or penalise players for the random order in which their games finish instead of on the whole season!

It's really no different from what happens now. If you lose 10 games in a row, tanking your rating, and then win the next 10, your final rating will be different than had it been the other way around (also depending on the opponent's ratings): https://stackoverflow.com/questions/239 ... ame-played

True the order of games has an impact on your final EAS but if you analyse possible ways the same 20 results could finish in your peak EAS will have more variability than your final EAS.

[dschingis27]

I am also opposed to the rating decay. I don't think that such a penalty will motivate in the long run, maybe only in the short run. Arena already feels quite chore-like, now you would need to plan ahead to play e.g. Agricola every 5 days? Imagine a player grinds a lot in first half of the season but has not much time in 2nd half. Will they be more or less motivated by the rating decay to come back to Arena in the next season? I would say less as their grinding did not count in the end. Same if you just don't play for 15 days and you see your rating decreased. Are you now more or less motivated to come back again? Maybe the first few times, you are more motivated. But over time, you might get frustrated by it and don't start Arena in the first place. Or you better postpone your grinding to the end of the season.

One thing not mentioned yet, I could imagine that this system would bound players more to one or two specific games during a season rather than taking a shot at more different games.

Instead of penalties, I suggest to make the Arena experience for Elite players more enjoyable.
- Provide an option to filter opponents by elite arena score
- The K factors are ridiculously high for some games. Why not have 3 categories (e.g. the known family games, casual games, "pro" games) with 3 different K factors like 20, 40, 60? You don't need to define a specific K factor for each game but just categorize the games properly.
- If there could be some mechanism to make the most frustrating part of Arena less severe, that would be great. In Arena, hours of grinding and building up your score can be destroyed by one mouse slip. The weight of single recent games is just way too high to motivate me to play regularly. Every game feels like a finale and not like a league game. I have no clear solution yet but maybe something in the direction to make games between players of similar score/skill more important than the games between players with a large score difference.

To summarize, I would say that Arena is already a bit too tense, and the addition of rating decay would do that even more.

[Alexey_LV]

This would incentivize players not to play until the end of the season. So we'd have very few players participating in the first two months and then a frenzy at the end.

Against.

[DoctorPecker]

So many people against this lol. Hmmm, I wonder why? The lack of competitive integrity from some of these players just makes me sad.