You are not logged in.
Maintenance is currently disabled.
This thread is open.
Posted by burytobolton 2010-07-20 17:33:24 GMT
Wondering about determining how well(efficiently) a player has been playing in the last 5 months. By taking the ranking of the player in nunmber of games played and subtracting the ranking of the player in their scoring rank. I am presently #94 in games played but #96 in scoring rank so I would have a (-2) rank for the past 5 months. gische is #85 in games played and #34 in scoring (+51) mylor56 is #21 in games played and #1 in scoring (+20) beleqk is #110 is games played and #16 in scoring (+94)
this ranking tells me that beleqk has been playing really well in the few games he is playing in the last 5 months, and that I am basically treading water and that my scoring rank is superficially supported only by the amount of games I am playing.
The major flaw I see is that "friendly" games are counted in games played but not toward score. So is there a way to configure a "games played" just on "competitive" games and then use that rank in comparison to scoring rank?
I did not get other players permissions to use their statistics and I apologize if this offends anyone.
Posted by gische 2010-07-20 21:11:00 GMT
I'm only offended that I've plummeted to #34 in ranking. Stupid vacation that caused me to get kicked from 5 games...
Posted by randomuser 2010-07-21 05:53:58 GMT
I think that the number of games, while somewhat relevant to scoring, isn't that strong of an indicator.
Looking at my own stats, discounting my most recent game (which I lost), and counting backwards, my best rating would be if you only took the last 60 games, and I get within 500 of that rating by game 23.
I think a better way to tell how you are trending is to look at win % vs expected win % over the course of your last 10, 20, 50, 100, 200 (or whatever) games.
Posted by burytobolton 2010-07-26 14:59:53 GMT
Ed, that is true but my win percentage can be skewed (like all numbers I guess) by me only playing in games that have no minimum starting rank vs. me only playing people who rank higher than me. I guess with a "3rd ranking" I could see how well others have been playing. Maybe there should be a bracket for non private games, that would allow you to only play people within say 1000 points above or below your grade. So at all times you can compare how you are doing against your direct scoring peers (I suppose I could just start my own games with this criteria). But getting victories against a couple new players and getting stomped by the big boys isn't telling me much about my play.
Posted by randomuser 2010-07-27 06:20:30 GMT
Well yes there are many ways you can game the system or be gamed by the system.
Short of having an ELO type of system, where the points you get are based off the rating of your opponent, there are going to be problems (and I've heard of some problems even with ELO).
Unfortunately, I think this was discussed in the past and I don't think Hammerite has time to code it in.
Posted by gische 2010-07-28 04:42:56 GMT
I also think Brass is a strange beast when it comes to competing against players of assorted skill levels. Ed and I have discussed this several times IRL. The game is completely different when you're playing against people who have no interest in working together as opposed to players who look for ways to maximize their scores. Rookies also make moves that can seriously gift the player who happens to go behind them in any given turn. So playing in games with all rookies vs. all experienced players vs. a mix can be completely different experiences.
Posted by burytobolton 2010-07-28 17:47:18 GMT
Yes, "Lies, damned lies, and statistics". Well I've started my three game experiment 0-4500, 4000-5000, and 4400-max. I'm really interested to see how well the flow of the games go, how being the highest ranked player, versus the lowest ranked player, vs., playing amongst my peers makes a difference in my own game play, my score, and the moves I make knowing the rank of players I'm playing against. I'll check to see if either of you have written any advanced strategy notes on BGG. Gische, best of luck to you in my 4400-max game.
Posted by Hammerite 2010-07-29 21:11:51 GMT
I did read this, I haven't ignored it.
I don't really know what to think about this idea. I want to implement an ELO system at some point, that is still planned.
Posted by burytobolton 2010-07-29 23:44:35 GMT
Oh, don't worry about it. I wasn't looking for immediate response from you imparticular. Just voicing my thoughts to the community. jason and ed have been great to talk with. I just thought it would be fun to see who has been playing really well lately, versus who is just treading water and not getting better and who was initially good, but has fallen. There are some players like Daniel Corban who amaze me with their ranking and how few games they have played. That sparks me to want to look at their games to see how they work the game. I agree that an ELO would be a great addition, as it would give people more cause to "play up" against the higher echelon players.
thank you again for all you do.
Posted by Shibumi14 2010-07-30 05:19:43 GMT
I'd be interested to see a ranking system for pts. average in 4 player competitive games. I specify 4 player because average isn't terribly relevant in 2 player, and the 3 player specialists have a slight advantage in the regular system, as a win in a 3 player is worth the same as a win in a 4 player (despite having one less opponent). I agree this rank would not be as good an overall indicator of skill as the current system, or an ELO, but I believe it is a worthwhile criteria for determining skill in general, as evidenced by the finalist of last years tournament. Also it would give the 2 and 3 player specialist incentive to play more 4 player games if the wished to top both the rankings.
Click here to return to the Board Page, or here to return to the Main Page.