🌿ELO System Updates

[RazorOz]

One thing is the ELO reducer is way too sudden, if you are going to have this, it's needs to be more gradual. A 100% reduction for 6 wins is ridiculous, it should be at least double that.

Playing Heat: Pedal for the Metal it's pretty common to get up winning streaks vs players, even with a large player pool, and you end up wanting to avoid players just because your have a 4-5 game winning streak against them, and have little to gain. It makes it hard to gain ELO after a certain point, as if you play lots, you will keep playing the same players.

6 consecutive wins isn't even a lot at many games, and there has to be a better way of doing this than actively discouraging people to play.

[GeoPaladin]

I would like to thank you all for your feedback on this subject.

I continue to regularly read your messages.
It goes without saying that any regression that occurs may need to be examined and dealt with.

But in reference to the primary remarks:

• Power leveling: we continue to monitor the status and it seems unfair not to be able to win ELO and only suffer the loss, even for players that keep on losing against the very same opponent.
• ELO Decay: based on all of your comments, it appears clear that this was a bad idea and that the leaderboard and arena currently in place allow for the prevention of the mentioned behavior. Consequently, we have chosen not to apply any ELO decay as originally planned.
• Red Thumbs: an attractive suggestion surfaced regarding the potential trade-off of having decaying red thumbs.

Thank you for responding. It's very much appreciated.

Regarding power levelling, I think the issue is broader than summarized. As a competitive player trying to climb, it's important to me to get full elo for a win. It is unfair to be penalized when I am not power gaming, so the penalty should only be applied very selectively.

The current system is very quick to stack to 100% and extremely slow to leave. While I would rather see the system removed entirely, here are a few suggestions to make it less punishing for legit players:

1) The penalty for consecutive wins should be reduced/eliminated by playing games against different opponents.

As an example, I got 0 elo against an opponent in Obsession after my 6th consecutive win against him. However, I had played 20 games against other people in that time frame. It's clear I wasn't targeting the player nor exploiting them so I shouldn't have received any penalty.

I would suggest that 5-10 games should be enough to fully reset the penalty.

2) The penalty for consecutive games should be reduced by time between games.

In my example above, my 5th win against the player in question occurred 4 days ago, prior to the penalty. If a player has a significant amount of time between games with the same player, they're less likely to be power gaming.

I would suggest reducing the penalty by a tier every 1-4 days. It's more important to play against new people IMO but I would expect power gaming is likely to occur in shorter intervals since you get relatively little elo each time.

3) The penalty should not apply to Arena games, which have blind matchmaking and are much harder to power game.

Again, using my example above, I was playing an Arena game and was penalized for 100% even though I had no way of selecting this player myself. It's just a natural result of being fairly high on the leaderboards in a small to medium Arena community.

4) Six games is a very small window for the penalty to reach 100%

While this isn't as much of a problem if you add in ways for the penalty to decay quickly by playing other opponents, it's still very harsh. My understanding is that currently the penalty lasts for a full month. It's not power gaming to play the same opponent 6 times a month. It's not even close.

Any changes to this depend on the above, but at least in the current system it's far too punishing.

[orgle]

I think ELO in coop games is still useful in that it gives a feel for how much someone has played. Players play different clues in Just One if someone is well experienced vs a beginner. Same in Hanabi (which I've not played) going from recent comments here. So what if someone has a lot of ELO amassed in coop games? Who cares? Who exactly does it affect? It just means the player has spent a lot of time playing that particular game. Good for them. Who is so upset about it that they feel the need to prevent this? What is the point?

The problem they're trying to fix is that they feel Hanabi players are too focused on ELO, and are aggressive towards players who make mistakes. Unfortunately, making ELO invisible is aggravating the problem, rather then improving it.

The proper solution is to give us more tools to indicate ALL the conventions used at the table, in addition to allowing us to see the ELO of players so we can give skill appropriate clues. There is a massive difference in knowledge between a player with 700 ELO and one with 1100 ELO, but right now we're all grouped together and you can't tell if someone is a new Master (and will need coaching and/or less complicated clues) or an old Master (who consistently understands focus, nuance & hand management).

Invisible ELO is terrible for the Hanabi community.

[Stroom]

I think ELO in coop games is still useful in that it gives a feel for how much someone has played. Players play different clues in Just One if someone is well experienced vs a beginner. Same in Hanabi (which I've not played) going from recent comments here. So what if someone has a lot of ELO amassed in coop games? Who cares? Who exactly does it affect? It just means the player has spent a lot of time playing that particular game. Good for them. Who is so upset about it that they feel the need to prevent this? What is the point?

The problem they're trying to fix is that they feel Hanabi players are too focused on ELO, and are aggressive towards players who make mistakes. Unfortunately, making ELO invisible is aggravating the problem, rather then improving it.

The proper solution is to give us more tools to indicate ALL the conventions used at the table, in addition to allowing us to see the ELO of players so we can give skill appropriate clues. There is a massive difference in knowledge between a player with 700 ELO and one with 1100 ELO, but right now we're all grouped together and you can't tell if someone is a new Master (and will need coaching and/or less complicated clues) or an old Master (who consistently understands focus, nuance & hand management).

Invisible ELO is terrible for the Hanabi community.

I have written personal notes about many players I play with. What kind of typical mistakes they make. Also a rudimentary counter of how "bad" I think they are. This often worked even better than elo score because some 900s would also not play very consistently.

[Silene]

The proper solution is to give us more tools to indicate ALL the conventions used at the table, in addition to allowing us to see the ELO of players so we can give skill appropriate clues. There is a massive difference in knowledge between a player with 700 ELO and one with 1100 ELO, but right now we're all grouped together and you can't tell if someone is a new Master (and will need coaching and/or less complicated clues) or an old Master (who consistently understands focus, nuance & hand management).

Can you make a suggestion how this list of all conventions would look like?

Because I find that very hard! I'm meeting someone following an aspect of convention that is new to me every other day. Although I don't consider many of those legitimate or good, I respect when people play that way (although I might mention to them, that I prefer a different approach for convention).

Even among very strong players, there are still different conventions used. So it would be absolutely impossible to me to make a list reflecting all of them. The convention set a singular player follows can be a collection of many different pieces and some of them overlap with others and others don't.

Invisible ELO is terrible for the Hanabi community.

My ELO is currently alternating between high Expert and low Master. I consider myself being capable to understand the most complex and high-level clues. But I blunder sometimes. Also in most of my games I play with the most difficult settings, which many players who you would consider the best players do not (i.e. using flamboyants although that is making the game very easy).

I hate the idea of other players judging me for a low-Master-ELO and dumbing down their clues for me because of that. So I am fine with only ELO-tier being shown.

If you really want to know, you can pause a second on game-accept-screen to note everyone's ELO if that is so important for you

[FaintBlueGlow]

RE: Hanabi

While I'm here, it would be very helpful to have a distinction between 2 player and 3+ player games. The conventions used in 3+ player game are often diametrically opposed to those used in 2p games (eg. early saves). A 2p-only Master would be completely lost in a high level 5p Master game & vice versa. Separate ELO ranking for 2p games would amazing. For instance, I'm a Master who generally plays 5p games, but I would probably be a Novice in a 2p Hanabi system, and that accurately reflects my understanding of & experience with 2p conventions. I shouldn't be playing 2p games with 2p Masters, but I'm often restricted from lower ranked 2p games because of my experience with 5p Hanabi. This makes it incredibly challenging to practice 2p Hanabi, as 2p Masters obviously have much higher expectations due to my Master rank.

...

Thank you for such a great site, and for listening to player feedback.

This! I play exclusively 2p Hanabi, and I would be terrified to try 5p at Master level. My Master rank at 2p should not transfer to 3+p games-- it's a completely different skill. Please separate the rankings!

[RazorOz]

The other thing with the ELO reducer for consecutive wins. Is it's designed to stop cheating right, wouldn't it be better if this only applied to 2 player games? As the chances of someone cheating this way in a 4-5 player game is very low. And the more players you are playing with, the more likely you are to run into the same players. It's just so frustrating to be playing players ranked like 300 in 5 player games, but get 0 from beating them because you have a win streak against them, and what's the alternative? Refuse to play with that player, which seems silly, or not play at all. And the win streaks to get to that point are pretty common, as it really isn't that uncommon at all to go on 5-6 game win streaks vs others.

[h_illes]

The other thing with the ELO reducer for consecutive wins. Is it's designed to stop cheating right, wouldn't it be better if this only applied to 2 player games? As the chances of someone cheating this way in a 4-5 player game is very low. And the more players you are playing with, the more likely you are to run into the same players. It's just so frustrating to be playing players ranked like 300 in 5 player games, but get 0 from beating them because you have a win streak against them, and what's the alternative? Refuse to play with that player, which seems silly, or not play at all. And the win streaks to get to that point are pretty common, as it really isn't that uncommon at all to go on 5-6 game win streaks vs others.

It's just as bad in 2-player. Top ranked players will have a win streak against most of the field. Before, the wins could balance out the occasional losses. Now they don't. I have 96 wins in my last 100 games of Century: Spice Road. Normally, with a streak like that, I would get to over 900 rating. But now I am below 870 due to the win streak penalties. Also, most of these games are in Arena, where I have limited control over the opponent. But even outside the Arena, I have no wish to avoid playing certain players just because I have a win streak against them.

This change is just bad overall. I have no clue how it got through, but they need to get rid of it ASAP. That comment from lordalx a couple of days back is simply not sufficient if there is no follow-up. BGA, there is no point in waiting any more, you already have the answers you are looking for. Further waiting will only result in more damage. Act now!

[Ranior]

I agree that it's wild to me that the admins say they are reading and considering things in this thread, yet still are sticking with keeping the win streak penalty for the moment.

It's causing so much more harm to so many players, and it's creating bad incentives. It makes me want to stop playing games as much on the site, or it makes me want to really prune what opponents I'm facing in a weird way.

It simply does not make the elo system better, which surely was the point of the change. Why then is it still around, and not been reverted yet?

[ChiefPointThief]

The other thing with the ELO reducer for consecutive wins. Is it's designed to stop cheating right, wouldn't it be better if this only applied to 2 player games? As the chances of someone cheating this way in a 4-5 player game is very low. And the more players you are playing with, the more likely you are to run into the same players.

Actually it is more efficient to cheat in games w/ 3+ players. There is more elo to steal other than from the duplicate account which most likely isnt ranked high. In a card team game you are guaranteed to win. If the duplicate account is paired w/ the cheater they will know where all the cards are. If they are on opposing teams the dummy account can sabotage. Worse case scenario if they still don't manage to win is they can just have the dummy account quit preventing a loss for the main account. I've seen several cheating accusations posted for team games this season. In solo games the main account is already ahead of the dummy account meaning no last place. Most likely they will be able to do enough damage to affect at least one other player.

It doesn't stop anything in 1v1 games either. The elo penalty isn't significant until the 5th straight victory. By that time someone can accumulate well over 100pts. So with 2 dummy accounts easily over 200. From their they can wait a month for the penalty to reset or just create a few more accounts if they want. In games with 3p+ the climb is much faster with increased k factor for new players.

Also in most of my games I play with the most difficult settings, which many players who you would consider the best players do not (i.e. using flamboyants although that is making the game very easy).

I try explaining this to ppl.

This! I play exclusively 2p Hanabi, and I would be terrified to try 5p at Master level. My Master rank at 2p should not transfer to 3+p games-- it's a completely different skill. Please separate the rankings!

You can say this about a lot of games. Also consider some games not only have different player counts but other settings, expansions, etc. It would be like asking to separate rankings for stone age players by player count, also by anniversary edition, and by expansion.