ELO rank now visible from everyone!

[Canis Minor]

[quote="Rajam"] Elo is a relative system; the score per se should mean nothing, the only thing that matters is how your score compares against others'. The problem is that there will always be an score that actually means something: the new-player's base score... and the problem is when some people reach scores below that. Is it possible to be worse than a complete beginner? Honestly it shouldn't, and I agree with a minimum Elo which match a new-player's base score (in this case, 0).[/quote]

Well, if Elo is a relative system (and it is), it also has capacity to reflect a situation when someone plays worse than a complete beginner. And the new system doesn't give a 0 rating (meaning the lowest anyone may have) just to complete beginners, but to all of them. Most of them will get higher, of course, but some will sink, and it should be reflected in their rating.
Why? Imagine you are a 120 elo player and you play against 10 different 0s. Your "Elo" changes as if they all were actual 0s, but in fact some of them may be -30 or -100. Therefore your Elo gets also inaccurate, and then your opponents' as well. Which leads us to the fact the whole system is kinda corrupted.

However, I agree completely with the rest you have written. The less "impassable thresholds" are introduced, the better the rating system.

[braven]

The biggest change I see is that new players now effectively start at 1300 instead of 1500 and are considered, by the system, as "the worst possible players" from a skill/rating perspective. This makes it much more risky for higher ranked players to play against them since they would now lose significantly more points if the "beginner" defeats them. This could then result in a less welcoming experience if players avoid allowing them to join their games because of the ELO risk.

I think it is incorrect to assume a new player is a "poor" player. They could be an exceptionally good player, perhaps a master at the physical board game, and just doesn't have a record here yet. Or maybe a naturally good board game player and have high ratings at all the other games they have played on the site, but still considered a very weak player by the rating system every time they try a new game. True, they will rise up the rankings quickly, but other players they defeat will unduly lose a lot of points because the player's skill was not accurately represented in the ELO rating.

To me, it seems like a better system would be to simply not lose or gain points when playing against a beginner/apprentice until their skill level has been better established. Or have some other mitigating factor in the calculation.

[Canis Minor]

[quote="braven"]To me, it seems like a better system would be to simply not lose or gain points when playing against a beginner/apprentice until their skill level has been better established.[/quote]

Aye, this is called provisional rating and is being used widely. Very reasonable solution, in fact.

[demiurgsage]

To me, it seems like a better system would be to simply not lose or gain points when playing against a beginner/apprentice until their skill level has been better established.

It's quite good solution I think. But the purpose of new system that players can't drop below start level. With your suggestion bad players will drop below start level after 10-20 games. Beginners can play with bad players or with "ELO off" option. Actually this ELO thing only makes people nervous. If you aren't in the top it's no matter how much ELO you have. You just need the possibility to choose opponents of your and similar level.

[fuonlynu]

nervous no elo on bga means nothin not a opinion I can proof this.period end of discussion.win versus losses of games shows whom is a better player period elo is a frame a small one where a player is winning that's all.i can proof this.please look at long term stats that will show u whom is good.
if u do it will show u player has best winning percentage for games played.so/so/simple.but bga wants to distort the truth.many of top elo will not come close to this winning percentags nor wins.and I believe we play to win and fun,whomkidding whom.

[Liallan]

But on the other hand, ELO is a really bad system from a game design perspective. Telling a beginner after his/her first game that he/she has a worst level (1470) than initially (1500) is really discouraging.

Why? If I lose my first game, I don't expect to go up! Are we catering to people who can't handle losing?

This could be also fulfilled by introducing a rule that any changes in Elo may occur not before some number of games have been finished (let's say, 25 or 30). A player will either improve during this number of games, switch to another game, or, if he / she's dedicated to the title, will be playing it no matter whether his / her Elo drops or not.

That's a pretty high number. I think I only have that many games for one, maybe two, games. Please keep the turn-based players in mind. Given the length of time it can take to finish a game, it can take a while to finish 25-30 games, and you don't need that many to get a feel for where someone is.

[fuonlynu]

now bga in there wisdom is punishing new players starting them at zero that isnt a solution hell even I started at either 1500 at one time.please again bga stop this.if a new player plays 5 games show his win/loss and games played we will figure it out.stop trying to do the impossible with this ELO THING.and please stop punishing the new players and me whom careless if they have a elo of 6oo or o. P.s same as karma rating by the way.wow/wow wouldn't want to be in your shoes bga instead of bringing greAt games and a greAt site to play on you have put urself in all crosshairs with a elo system that don't work and never will and on top of that your karma system.WHY////YYYYYYY.AGAIN THK U FOR UR HARD WORK.

[dudi2]

ELO is a good system, as it reduces the number of players which just try to end a game fast if they don't have a chance to win.

Normally in Hearts a very good player should have a good win rate, but also should end more often in 2nd or 3rd place as in 4th place.
Your ended in 261 games more often at last position than you won:

1st 82x
2nd 44x
3rd 41x
4th 94x

If you would try to get a good result, your ELO would be higher.

[fuonlynu]

thks for opinion dudi.winning is all that counts period.im 2nd in wins and my average on wins is probably top 5or 10 so in my opinion I'm a very//very good player and yes there are few that are better.P.S HOPE eNN DONT GET MAD FOR ME REPLYING TO WHAT DUDI WAS SAYING.THKS bga

[Liallan]

Elo has its inherent flaws, but getting rid of it and putting percentages is WAY worse. Please never do that.

I'm going to suggest ignoring fuonlynu. He's usually blabbering nonsense. I know it's tempting to respond - I'm right now trying to resist that temptation. LOL. (Not an easy thing for me.)

Now, I DON'T agree with players not being able to lose Elo between an arbitrary range (1-100 currently). If players just keep losing, they should get stuck at 0 Elo. If they win once, raise their score a little, and then keep losing again a lot, they should just return to 0 if the Elo calculations dictate it, because it means that those players are still pretty weak and still at a beginner level. A new floor (of 100) is kinda redundant with the base minimum of 0. Please just keep one floor: 0. Do you want to make distinctions between complete beginners and players that have played some games? Then putting visible the n° of victories and n° of games played should help in that regard; no need to make arbitrary floors and extra rules on the Elo system.

+1 Totally agree - this makes sense.

I think it is incorrect to assume a new player is a "poor" player. They could be an exceptionally good player, perhaps a master at the physical board game, and just doesn't have a record here yet. Or maybe a naturally good board game player and have high ratings at all the other games they have played on the site, but still considered a very weak player by the rating system every time they try a new game. True, they will rise up the rankings quickly, but other players they defeat will unduly lose a lot of points because the player's skill was not accurately represented in the ELO rating.

To me, it seems like a better system would be to simply not lose or gain points when playing against a beginner/apprentice until their skill level has been better established. Or have some other mitigating factor in the calculation.

Agreed. Someone new here doesn't mean they are new to the game, and if they are already experienced at it, they should rise quickly enough. Someone totally new to the site itself will likely start with games they already know. And I would guess the majority already know something, especially with some popular mass market stuff like Yahtzee, Checkers, etc.