ELO scores

[HATFIELDER]

So I play 2 player and 3 player with the same regular players and playing is more about the fun and challenge than the assigned ELO, however the system seems totally broken.

If we win (complete all 5) with the standard set up we get +0 ELO (yes actually zero, before that we were getting 0.4, 0.23, 0.04...)
If we lose (ie. get 24points!) we get -7 ELO.
If we then win the next game we get +0 not even regaining what was lost due to the previous loss.
Playing with rainbows in as a sixth suit and clued with all colours we do get some ELO (but tiny amounts)
Playing with Black and actually succeeding for the first time on this extremely hard set up (are we the only ones finding this so much more of a challenge than the rainbows) we got 0 ELO again.

What incentive is there to play out full games where you win according to the rules but not in the eyes of the website?
How are people getting scores of 1000+ when we are stuck on 700?

If you got a +ve result for any game where you won according to the rules (>21pts) then I imagine there would be more players happy to play to completion.

Also is it possible to add the Crowning Piece variant? It's so gutting to lose the game because the last card of the deck is the last 4 that you need (when it's matching pair was discarded as one of the first discards)

[Raal]

What incentive is there to play out full games where you win according to the rules but not in the eyes of the website?
How are people getting scores of 1000+ when we are stuck on 700?

Most of those players play tables of 2 player and Multicolor difficulty, and archive 30 scores.
As ELO of Hanabi is calculated from average ELO of playing players, playing with lower player gives many ELO points.
If you always play with same players, there will be ELO limit virtually.

[qwertasdfg]

What incentive is there to play out full games where you win according to the rules but not in the eyes of the website?
How are people getting scores of 1000+ when we are stuck on 700?

Most of those players play tables of 2 player and Multicolor difficulty, and archive 30 scores.
As ELO of Hanabi is calculated from average ELO of playing players, playing with lower player gives many ELO points.
If you always play with same players, there will be ELO limit virtually.

How do they achieve 30 if the maximum score with that new "version" is 25?

[Raal]

What incentive is there to play out full games where you win according to the rules but not in the eyes of the website?
How are people getting scores of 1000+ when we are stuck on 700?

Most of those players play tables of 2 player and Multicolor difficulty, and archive 30 scores.
As ELO of Hanabi is calculated from average ELO of playing players, playing with lower player gives many ELO points.
If you always play with same players, there will be ELO limit virtually.

How do they achieve 30 if the maximum score with that new "version" is 25?

It had seemed fixed, at least on new tables.

[Travis Hall]

Here’s the thing, Hatfielder... You’ve got a current rating of 750. Just going by your last game, you’re playing with friends who are at a similar rating. That’s master level. However, you’re playing the easiest form of the game.

An Elo system isn’t designed to let you just keep accumulating points forever. It’s designed for your gains to taper off, until you reach an equilibrium and just bounce up and down a little bit as you win and lose. Once there, you have to demonstrate improvement to get higher.

And currently, you’re not demonstrating that higher level of skill. Maybe you are skilled enough to deserve a 1000+ rating, but because of how those gains taper, it takes an incredible amount of time to get a significant increase. The mean score of a 750 player in the basic game with two players is going to be something in the region of 24.9. Even if your mean is 24.99, it’s still going to be very slow.

So if you want to demonstrate higher skill, faster, you’ll need to play harder forms of the game. Adding a sixth suit gives you a bit more range at the top end - you might be comparing your mean of 29.9 to something like 29.5. Avalanche of Colors makes it harder again. So do larger games - I believe that is built into the Elo system too.

I’m not sure how well Black Powder has been taken into account for difficulty yet. (It might be waiting for data to accumulate to determine how much adjustment is needed.) And I don’t know that 5 Flamboyants is actually harder at all; it may be a little easier.

But certainly if you want to see your rating increase, play with a larger group, and Avalanche of Colours. And rack up lots of 30s.

There also appears to have been a bug concerning scores for non-Black Powder games (possibly only the games started before the new version was released - so typically turn-based or untimed games). That might have been hurting you as well, but I doubt it is your full problem. I think you should start by aiming for 30, rather than 25.