do overall ponts mean nothing now??
im num 3 in overall standings and #7PO , random who is 1st in overall standings is #8 on PO
why :S
overall points are there to measure a player's skill level and to calculate points for your opponents.
season points are there to measure a player's performance in that one season.
so no, your own overall points do nothing to help your own season performance.
after the first full season under this new system, we see that season rating isn't reaching overall rating in only 80 games.
maybe this is ok, maybe it isn't. it's always up for discussion if you want to.
one thing I can think of if this really bothers a lot of people is to increase the cap on maximum points won per game..
but I'm not 100% sure if this even solves the problem. just a thought off the top of my head.
I wanna increase the minimum points you can win, 10 at least (for seasonal).
well it isn't ok :S it's forcing me to play many many games to keep up with players who play 200+ (even 500 :O) games a season.. I barely play 200 games in tel/tus/trl alltogether
so more active players will reach playoffs instead more skilled players
it does look like a slight advantage to active players at the moment.. but not that much. think the point system does work quite fairly but maybe a slight adjustment to the formula?
To be honest, congratulations for the active players.
I don't see why they should be critisized for playing more with or without more available time than others, that's their choice to run their life that way, good for them.
We should know by now to be successful in a WA League, at least modernly speaking, you have to be very active, so, no offense to anyone but, if you don't have the free time to play, then either don't, or try your best to give yourself more time to play.
This is the best we have, we can't get every player playing every other player a certain amount of games that would give optimum standings like professional sports or something, not to mention the fact we have new players signing up and getting involved with the League every week almost everyday.
Doesn't matter what the system is, I think the most active players (obviously with average skill with at least 2/3 schemes) are going to be the victors...
What's so bad about that? If it means that much to you, hijack your own life and turn full-on WA-Nerd! xD
Quote from: franz on April 27, 2011, 10:27 PM
one thing I can think of if this really bothers a lot of people is to increase the cap on maximum points won per game..
but I'm not 100% sure if this even solves the problem. just a thought off the top of my head.
That does help indeed, but the points at the start of the season will be quite outrageous. (although... as far as I know the max score of points isn't scored often.. it's only possible for something like a 500 beating a 2000 or not?)
Nail's idea is the opposite.. it makes it even easier for active people, and for noob bashers.
I can see Almog's problem, although I do think, with just 24 games played over the minimum amount of games needed for Po's and a #7 spot he can't really complain about the activity needed. The only one that should perhaps is Angus, but if he plays a couple more games he can approach the top 8 as well.
I don't really know (off the top of my head) any quick fixes to this. Besides, I kinda think that people like Phanton and Campz who play over 250 games each season (with a winning % of at least 65) have all the right to be up that high and deserve a PO ticket.
Btw, it's kinda messed though that the players with the best winning % are likely to face each other in PO's (Zippo vs Random :P).
It would perhaps be an idea to seat those first with the highest winning pct (although the strength of the opponents isnt considered then) or maybe with the highest overall rating or something like that.
Random00 has just got 8th in the standings after a short break maybe? Anyway in the remaining days should be able to lift himself up more.
Also note that we can't have the standings table exactly as we want with just a formula. Even if the current formula works the best this season, there's no guarantee that the next season we get our desired table as well.
lets stick with it i reckon.
see how it goes, i think it looks fair enough, u have active players in there (who have had good seasons, all with nearly 70% winning ration plus). Also u have not so active top overall players like random00 and almog.
activity is always going to help i just in old point based leagues we had a game limit. but this is kinda like a ladder but still point based..
we couldnt really have a max of 200 games a season could we?
Quote from: Chicken23 on April 28, 2011, 12:14 PM
we couldnt really have a max of 200 games a season could we?
its rly good idea to make requirement for all players to play determined number of games.
every man should play "X" games in each seazon to be classified for playing PO, no more and no less,
if someone playing for example only 50- 100 games, and other playing 300+ its not fair when we compare both of them ( when both got the same points)
:(
im not sure if introducing a maxinum number of games would be visable, it would put more focus on your performance per games as you couldnt just be the most active and reach playoffs but because different players are earning different points unlike previous point system leagues i dont think u can cap the max games allowed and move away from unlimited.
franz, hhc, your thoughts?
I'm not a big fan of that. Primarily because it hurts activity as well. If Phanton is only allowed to play 200 games, TUS will miss out on the other 391 (lol..).
I prefer a system in which active and slightly-active players have almost equal chance. I think the current system is ideal for 3 months-seasons. 2 seasons is a little short to get the ratings to settle at a certain point.
However, 3 months is rather long. Even if it's slightly beneficial to superactive players, I'm preferring just 2 months.
Same applies to the clanner league, even more so I guess.
Quote from: Komito on April 28, 2011, 10:36 AM...we can't get every player playing every other player a certain amount of games that would give optimum standings like professional sports or something...
Actually, we can. That will be the new league I will code and it will launch around 2013. ;D But seriously, it's a whole new system and it will work in a similar way. ;) But that's far away from here.
Before you start hating me alredy, MI knows about it, hopefully he will be involved as well. It's not a challenge or a rivalry on any level against TUS or anything like that. It's a different prospective that will hopefully attract people, a new way of competing, a new level of competition.
I don't like discouraging activity. Why should active people get penalized for enjoying playing lots of tus games?
A maximum games cap would force them to stop playing for a month if they reach it early.. that doesn't seem best.
I think my first suggestion wasn't quite right (or even possible), to increase max points per game.
If I understand correctly, MonkeyIsland says it's still ELO formula calculating points won per game.
Now for TEL, everything looks fine, and many players reach their overall rating after reaching minimum games.
For Classic, it's different because there are so many schemes.. it's way more games to approach overall rating in every scheme separately.
I'm wondering what the K-factor MonkeyIsland uses for the ELO calculations (75? 80?) because if I understand that math correctly, that determines the range of points won/lost. If that K-factor increases, it might amplify and intensify how fast people gain/lose rating.. though this might only be appropriate for classic league (and probably only for the season rating calculations, not overall). I have no math to back this up, and anyone can talk on this if they know better. Though MonkeyIsland might be the best resource here.
As far as I am aware, MI is using 80 franz.
Ray, sorry but I just don't see that working, how are you going to get every player to play every other player a set number of games?
Without forcing limits that are too little or too much for the players pleasure and enjoyment?
You will see. :)
... hmmm
so maybe add minimum number of games, 150 for example..
http://img26.imageshack.us/i/beztytuuguco.png/
there already is a minimum
An alternative to receiving/losing points with each game could be using league performance rating rather than having seasonal standings to decide on playoff standings.
It's a system currently used during chess tournaments as well (called tournament performance rating there for obvious reasons), though it has no effect on actual standings right now (there are other tie-break rules in play in chess that can be applied because there's a fixed number of games) and there are several ways to calculate league performance rating.
What league performance rating means, well, it's just what it says! You take a look at your performance; how many points did you score and how strong was your average opponent. If I were to score 50% of my games (30 wins in 60 games) against 60 people whose average rating was 1200, then my league performance rating is 1200. In other words: The score I got would theoretically be reached by someone with a rating of 1200.
There are various ways to calculate league performance rating, but perhaps it's easier to show this link (http://www.fide.com/fide/handbook.html?id=73&view=article) for an example. The first table shows a row of percentages next to a row of rating differences.
Let's take season 19 to practice with :) https://www.tus-wa.com/leagues/classic-stats/?s=19
Random for example has scored a percentage of about 81%. The rating difference associated with that is +251. Let's say the average rating of his opponents was 1600. This means his league performance rating is 1600 + 251 = 1851. Let's say angus' opponents averaged a rating of 1500. Angus scored 82%, which corresponds with + 262, which makes his league performance rating 1500 + 262 = 1762.
The system therefore makes use of both the score you achieved as well as the average strength of your opponent. I'm not saying this particular calculation should be used, but it's the idea that counts :)
If this were to be used instead of now, there would be a few advantages (at least in my opinion, though of course some people may disagree :)) in comparison to previous and current systems.
- Your actual rating plays no role for new seasons. HHC's system already helped in this respect, of course.
- Activity is not enough to qualify. Naturally, you'll need a minimum amount of games played to be eligible for playoffs, but it's the results that count. It doesn't matter whether you win 80% of games against 1500 average opponent with 100 games won or 1000 games won; your league performance rating will be the same. The current system allows you to gather points simply by being active, while league performance rating doesn't have this. But on the other hand, this also means that there is no need to have a maximum amount of games per season :)
- Each game counts equally. Right now, if you win a game against Random (or lose against me :)), the result matters most when it happens near the end of the season and if an important result is right at the start of the season, the effect of that game is completely lost at the end of the season.
- Noob bashing could be a thing of the past. This is where the calculation for league performance rating would have to be tweaked :) I picked the average rating of the opponents for the examples of Random and angus for a reason (I didn't check whether their actual opponents); if you play only weaker opponents, you're making your maximum attainable league performance rating lower. Of course, it's still not a problem to play a new player once in a while since only 2 games will play a relatively small role within 100 games for example.
Just a thought :)
I'm hearing a lot of "you gather points simply by being active"
or "more active players will reach playoffs instead more skilled players"
but look again: https://www.tus-wa.com/leagues/classic-stats/?s=19
sure, Phanton played 610 games, but Zippo is ahead of him by only playing 130 games.
then look at the 'active' people who didn't reach the playoffs "simply by being active":
XxXM0j0XxX 288 games
Karlos 338 games
BoNuS 272 games
Berria 330 games
True, but if you look at the people that qualified for playoffs, the #1 spot went to the person with the lowest win percentage :) Look at the winning percentage and the amount of games played and you'll people with lower percentages have more games played consistently (at least, when you're looking at similar ranks)
right, but..
in this system, win percentage doesn't really mean much anymore.
you could have 100-0 vs the worst players and still be outside the top 8.
keep in mind, this most recent classic league season was the most active ever.
if this massive activity continues, it pushes those high overall rated players to play more to stay competitive. (see Zippo)
if activity drops, high overall rated players won't have to play as many to be competitive in the top 8. (Almog/Random would likely be higher)