Ranking

[Gammonrider]

Hallo.

How can I found out how many players we have on FIBS and what is my Ranking in this List??

regards
Michael ( Gammonrider)

[maria]

Hallo.

How can I found out how many players we have on FIBS and what is my Ranking in this List??

regards
Michael ( Gammonrider)

When you log in to FIBS, in your command line type STAT.
There are 16287 registered users.

In the command line type  RAT Gammonrider

You are ranked 2918.

To see the top 20 players, type RAT in the command line.

[don]

New users should be aware of the extensive FIBS help system.  This is ignored by every client that I know of.  To access it, type "help" or "man" on the FIBS command line (many clients make this obscure too) or go to http://fibs.com/CommandReference/index.html.

The actual command (RAT) mentioned by maria is "ratings".  Type "help ratings" on the FIBS command line for more.  (maria is correct in her description of usage of 'RAT' as far as she takes it.)

--
don

[socksey]

I didn't know that!     Wow!  Thanks for the info, Maria, don, and PersianLord!     I learned a lot today.  For as long as I have been playing on Fibs, I have never gotten into the actual list of commands for Fibs.     I guess I always do things backwards.   

socksey



"No animal should ever jump up on the dining-room furniture unless absolutely certain that he can hold his own in the conversation."  - Fran Lebowitz

[blitzxz]

Ratings command doesn't seem to work... From to gives only my own rating and top 20 doesn't work.

Player who has rating 193 has rank 6729. So that is propably the closest number of active players on fibs. Median is somewhere between 1523 and 1525.

I was really getting interested in these so I worked some distributions also.

1901+         66

1798-1901   232

1699-1798   639   

1601-1699   1102

1499-1601   1925

1399-1499   1693

1305-1399   689

1211-1305   248

193-1211    135

or

1901+         #+

1798-1901   ####+

1699-1798   ############+   

1601-1699   ######################

1499-1601   ######################################+

1399-1499   ##################################

1305-1399   #############+

1211-1305   #####

193-1211    ##+

[playBunny]

And similarly for Dailygammon

http://www.gellie.plus.com/Stats/DG_Ratings_Freq_Dist__Nov_1st_2007_to_Feb_1st_2008.htm

Interesting how both populations have the same starting rate of 1500 but the peak is (of established players, green) is higher.

[blitzxz]

Here is fibs ratings distribution in november 2000.

http://www.bkgm.com/rgb/rgb.cgi?view+888

Very interesting small change has happened. The number of players is almost identical but now there is less players in the tails and more close to avarage. Less variance in other words. What might be the reason for that?

Also the median is couple rating points lower then in 2000.

http://groups.google.com/group/rec.games.backgammon/msg/26df7ba20b6f549a

[boomslang]

Very interesting small change has happened. The number of players is almost identical but now there is less players in the tails and more close to avarage. Less variance in other words. What might be the reason for that?

I noticed these thinner tails too. It is probably the aggregate effect of the gBOTS: they take from the rich and give to the poor.

BTW: the Gini coefficient is less than 0.1; no need for digital Robin Hoods I'd say!

[dorbel]

To some extent the rating system is distorted by a number of factors, which are well known and fairly obvious to all. Fibs ratings are only a very general guide and a median figure between a player's highest and lowest rating would be much more interesting. Would it be too hard to show that I wonder?
A rating system undistorted by droppers, bots, careful selection of opponents and limitations on match lengths is of course available to all those fibsters who play in Tomawaky's excellent Fibsleague. This is IMO a very accurate system and less liable to great fluctuations, as one can only play 70 or 80 matches a year.

[blitzxz]

Fibs ratings are only a very general guide and a median figure between a player's highest and lowest rating would be much more interesting. Would it be too hard to show that I wonder?

I already posted median. It's somewhere between 1523 and 1525.

[dorbel]

You misunderstood me. I meant the median figure for each player. The rating as displayed is a poor guide, because we can't know where the player is in his range. If we were given the median figure as well as the current rating, that would probably be a more accurate reflection of ability. Even better might be an average of the last 100 ratings that have been displayed for that player. Certainly there must be some better way of ranking players, that isn't so volatile.

[blitzxz]

You misunderstood me. I meant the median figure for each player. The rating as displayed is a poor guide, because we can't know where the player is in his range. If we were given the median figure as well as the current rating, that would probably be a more accurate reflection of ability. Even better might be an average of the last 100 ratings that have been displayed for that player. Certainly there must be some better way of ranking players, that isn't so volatile.

Ou, you meant that. Max rating is also good and simple one if the avarage or "real" rating is clearly over 1500 and you have high experience. I would like if fibs would track also the max ratings.

And talking about volatility and luck in backgammon...

"This is IMO a very accurate system and less liable to great fluctuations, as one can only play 70 or 80 matches a year."

Low number of games in year means that flunctuations are small or more accurately slow but they last ages... Good or bad streak could last several years if you are only playing 70 or 80 matches in a year. This is even more clear if you play only couple times a year in high profile live tournaments (lots of players and big prizes). You're bad or good streak could easily last longer then your life time. Luck will cancel out in long run but the run just might be too long.

[stiefnu]

You're bad or good streak could easily last longer then your life time.

LOL!  Now I know what people mean when they say they were dead lucky! 

[lewscannon]

You misunderstood me. I meant the median figure for each player. The rating as displayed is a poor guide, because we can't know where the player is in his range. If we were given the median figure as well as the current rating, that would probably be a more accurate reflection of ability. Even better might be an average of the last 100 ratings that have been displayed for that player. Certainly there must be some better way of ranking players, that isn't so volatile.

Yeah, I've gone to a low in the upper 1600s to a high of 1924 (coincidently, the year that mookie was born). I'm sure my true rating is very close to that top number, perhaps even a bit above it, but it would be interesting to see where I average out.

[PersianLord]

Ou, you meant that. Max rating is also good and simple one if the avarage or "real" rating is clearly over 1500 and you have high experience. I would like if fibs would track also the max ratings.

And talking about volatility and luck in backgammon...

"This is IMO a very accurate system and less liable to great fluctuations, as one can only play 70 or 80 matches a year."

Low number of games in year means that flunctuations are small or more accurately slow but they last ages... Good or bad streak could last several years if you are only playing 70 or 80 matches in a year. This is even more clear if you play only couple times a year in high profile live tournaments (lots of players and big prizes). You're bad or good streak could easily last longer then your life time. Luck will cancel out in long run but the run just might be too long.

Exactly 

Dorbel's proposal to judge players based on their FLG rating is nonsense, because statistical data will be far from enough to make any liable judgement on. In each FLG session, a player will play at most 12 matches (if he's enough lucky to get all his opponents to play) in a 2-month period, while regular fibsters play about 15 matches EVERY DAY.

Regards

[blitzxz]

because statistical data will be far from enough to make any liable judgement on

I didn't mean that... I only meant that changes are slow and last long with low number of games. It doesn't have to mean that it is also more inaccurate. FLG ratings seem correlate with fibs ratings and those correlate with analyzed snowie error rate. So I think they're both working fine.

[socksey]

A rating system undistorted by droppers, bots, careful selection of opponents and limitations on match lengths is of course available to all those fibsters who play in Tomawaky's excellent Fibsleague. This is IMO a very accurate system and less liable to great fluctuations, as one can only play 70 or 80 matches a year.

I don't agree!  My rating in Tomawaky's system is at 1467.72 with 2197 exp.  I've never been that low in any other system, and this is taking into account that i won the Master A league once.   

socksey



"Some people are born on third base and go through life thinking they hit a triple." - Barry Switzer

[blitzxz]

I don't agree!  My rating in Tomawaky's system is at 1467.72 with 2197 exp.  I've never been that low in any other system, and this is taking into account that i won the Master A league once.   

You can't compare the actual numbers because players in flg are on avarage better then in fibs. There seem to be almost no players with under 1500 fibs rating. So most have lower ratings there then in fibs. And there always is those +-100 fluctuations.

[lewscannon]

I don't agree!  My rating in Tomawaky's system is at 1467.72 with 2197 exp.  I've never been that low in any other system, and this is taking into account that i won the Master A league once.   

socksey



"Some people are born on third base and go through life thinking they hit a triple." - Barry Switzer

And just think, you'd probably be in the 1300s if you didn't run across me from time to time. Don't know what it is about sox, but she always seems to get the upper hand on me. I think distraction has a great deal to do with it.

[socksey]

You can't compare the actual numbers because players in flg are on avarage better then in fibs. There seem to be almost no players with under 1500 fibs rating. So most have lower ratings there then in fibs. And there always is those +-100 fluctuations.

I almost never play players rated lower than I am on Fibs.  Often when I do that, I lose, thereby hurting my rating tremendously.     

I have a 43% win rate over vegasvic who is rated far higher than I.  Not bad, but he has taught me a lot.  

Some of the higher rated players on Fibs will not play with me, i.e., Niholympic, Honeygirl_II.  Too bad about that.  We could probably each learn from the other.  Now that resh_lakish has reached over 1700, he refuses to play with me anything but 2 ptrs, if that!     Yeah, I know, they are scared of me.  

Distraction is my middle name, lewssssss. 

socksey



"Trying to eliminate Saddam…would have incurred incalculable human and political costs. Apprehending him was probably impossible…We would have been forced to occupy Baghdad and, in effect, rule Iraq…there was no viable "exit strategy" we could see, violating another of our principles.  Furthermore, we had been self-consciously trying to set a pattern for handling aggression in the post-Cold War world. Going in and occupying Iraq, thus unilaterally exceeding the United Nations' mandate, would have destroyed the precedent of international response to aggression that we hoped to establish.  Had we gone the invasion route, the United States could conceivably still be an occupying power in a bitterly hostile land." - George Herbert Walker Bush, from his memoir, A World Transformed (1998)

[dorbel]

Because rating systems are closed, i.e. they have no connection with each other and because they have different parameters and even different rating formulae, comparisons between them are useless. Within each system all one can do is compare one's rating to the other players within it. The actual number of one's rating is unimportant.
Socksey has "never been as low in any other system". This doesn't mean that the system is bad, only that socksey has had a long bad run in FLG and is close to or at the lowest point in her rating range. In a year's time she may well be 100 rating points higher than she is now, but I am not sure that learning from vic and resh is the way to go!

[lewscannon]

Because rating systems are closed, i.e. they have no connection with each other and because they have different parameters and even different rating formulae, comparisons between them are useless. Within each system all one can do is compare one's rating to the other players within it. The actual number of one's rating is unimportant.
Socksey has "never been as low in any other system". This doesn't mean that the system is bad, only that socksey has had a long bad run in FLG and is close to or at the lowest point in her rating range. In a year's time she may well be 100 rating points higher than she is now, but I am not sure that learning from vic and resh is the way to go!

Well, you may not learn much about bg from them, but I'm sure there are a lot of other subjects they are well versed on.

[playBunny]

Because rating systems are closed, i.e. they have no connection with each other and because they have different parameters and even different rating formulae, comparisons between them are useless.

"Useless" is too strong. For an individual it's difficult to make sense of two disparate ratings but for a set of players you can expect a correlation even if not a n accurate mapping. High in one system tends to be high in another, likewise with the lows. DailyGammon is comparable to Fibs, to a degree and even BrainKing's whacky rating system (modified Chess formula, unequal amounts for winner and loser, no match lengths considered) is still comparable to DailyGammon's. We've been discussing the former only just recently, in fact. http://www.dailygammon.com/bg/forum2/main/read/25784

[don]

Actually, what is wrong with dorbel's (and Tomawaky's) rating system is simply the group it is assessing.  For example, if you take the "master" players' games against each other, at this level of play average play will result in a 1500 rating even tho the players are all above average in backgammon skills.

Do some math on "help ratings" to convince yourself of this.

--
don