Activity / Points for Losses

[teach]

Give a win bonus of 1/10 or 1/8 points for every win.

Strong players get trumped by weak players who "cheat" the system by piling up meaningless losses.

_______Current-->Fixed

0-13 = 13 points --> 13 points
2-11 = 15 points --> 15.2 points
5-8 = 18 points --> 18.5 points
6-4 = 16 points --> 16.6 points
6-0 = 12 points --> 12.6 points
4-0 = 8 points --> 8.4 points

_________________________
November Alpha "controversy"
_________________________
cut-line 'outrage' -->1/10 point bonus fix

25/3 = 36.5 --> 39
23/1 = 34 --> 36.3
17/6 = 27.5 --> 29.2
16/3 = 24.75 --> 26.35
11/11 = 24.25 --> 25.35
3/28 = 23.25 --> 23.55 # these 3 gamed the system
4/26 = 23.25 --> 23.65 # now only one succeeds
4/28 = 22.75 --> 23.15 #
_cut-line____
11/9 = 22.75 --> 23.85 ## these two got cheated
13/4 = 22.5 --> 23.8 __## now rewarded
10/9 = 22 --> 23 ____### at 1/8 bonus this guy may ALSO MAKE IT
5/11 = 16 --> 16.5
4/4 = 10.5 --> 10.9
1/4 = 5.5 --> 5.6

edited again...

[topazg]

Give a win bonus of 1/10 or 1/8 points for every win.

Strong players get trumped by weak players who "cheat" the system by piling up meaningless losses.

Current Fixed
_______ _____

4-0 3.5 3.8
0-8 3.5 3.5

13-0

I failed to follow your scoring point example at the bottom, is there any chance you could explain it?

I suspect the 1/10 or 1/8 is unlikely to be popular, as the larger value of 1/4 (still half the current value) got only 10 out of 48 votes in this poll. However, you do have a point about "meaningless" losses. There are some games that feel particularly like someone has simply played a game without much care just to get their points, and this is rather against the spirit of the league. I do however suspect that this is in the minority - An 8k that plays a 3d but knows they have no chance of winning can still give the game their all, use their time, and be involved in the post game review. If so, the loss is not at all meaningless, and exactly the sort of game the ASR is trying to promote. If that person gets a 0-18 win/loss record we still want to have a system in place that really rewards their effort to get games in, provided that games are valuable games.

Interestingly, I share part of your feeling, and this is why I put together proposal B above. I have found out that many kyu players participate because they really want to get promoted and think they can do so simply by "playing lots". I find this slightly uncomfortable because a) I think it encourages people to play guaranteed losses and just quickly get them over with so they can start the next one, which isn't useful to anyone, and b) it seems to misunderstand the basic value of promotion. Being in Gamma isn't really "cooler" than being in Epsilon, and you'll still have plenty of stronger people to bash your head against in Epsilon. If your goal is truly to learn and get stronger, surely being in any tier is fine if you have a 0-18 W/L record, as clearly you have plenty to learn from people in your class - why would you want to get promoted further if you weren't winning the majority of your games? I mean, what's the point?

So, I still prefer the idea of having some form of cap - if you aren't getting points for losses after the first 5 until you win some, at least you will take on the loss games because the game itself has value, rather than just the points you can earn. However, I think people feel that part of the motivation is gone for kyu players if they do not feel they can promote - I feel this is sad, as it doesn't seem to be a major point of the ASR, but for some people it is, and that is important.

Either way, we do want the higher tiers to be populated with people who play fairly actively, and any dan player who gets in 15 to 20 games per month will promote up to Beta in consecutive months. Getting into Alpha is hard, but frankly a KGS 1d is at the weak end of the Alpha class, so considering all the players are both strong and active at that end of the league, getting into Alpha is supposed to be hard

[teach]

My "proposition C" has the advantages of;

Simple to understand for players.
Easy to code for.
Fixes the more obvious score manipulating shenanigans.

Should keep dan guys happy if they are rewarded for teaching kyu guys in lower classes. They get a fair shot at top classes without worry of "tweakers" holding them down.

If ~4d guys have to play ~7k on even all the time there is no way many of them will stick around. It gets old really fast.

In 14 man groups, 20 played games out of ~40 possible should be considered as an active player. If a guy plays 20 extra games but loses them all should he really trump the guy who put in 20 quality games?

[topazg]

My "proposition C" has the advantages of;

Simple to understand for players.
Easy to code for.
Fixes the more obvious score manipulating shenanigans.

I agree, but it seems to be not what the majority of the league wants from the voting on proposal A. EDIT: See next post, I'd misread

Should keep dan guys happy if they are rewarded for teaching kyu guys in lower classes. They get a fair shot at top classes without worry of "tweakers" holding them down.

If ~4d guys have to play ~7k on even all the time there is no way many of them will stick around. It gets old really fast.

I agree, but I haven't seen any examples of being held down by tweakers. We attempted to seed dan players into Delta this month to help with the problem, but the seeded dan players have been almost completely inactive so we won't be doing it in January.

In 14 man groups, 20 played games out of ~40 possible should be considered as an active player. If a guy plays 20 extra games but loses them all should he really trump the guy who put in 20 quality games?

I have never seen a dan player get 20 games played and not promote, except in beta where three dan players were competing for two spots. If the games are sensibly distributed (say play 2 players three times, 4 players twice and 6 players once), you get 31 points from 20 games. 39 losses (3 against each player) gives only 22.75 points, over 8 points less. So the answer is no, the guy shouldn't trump the quality player, and the guy most likely won't manage to without a good number of wins himself.

EDIT: For example, you are already storming ahead in your class, with 10 points (twice as much as second place). Realistically, with dropouts, I expect 6 people to promote from Epsilon. If you hit 20 points you will definitely promote, and 15 is probably enough, so you're virtually there already. It's not as onerous for a dan player to work their way up as you're probably thinking at the moment, I promise (having been there myself)

[topazg]

Further comment: I misread your proposal. I see now what you are asking. That's a possible idea certainly. I'll speak to my fellow admins.

The problem you are attempting to resolve is the the reason that I decided to create the poll. Spam losing is not a viable strategy for promoting, but it is a good strategy for avoiding demotion. Having a fix for this that doesn't affect our kyu players' desire to play is still a good thing to aim for.

[Gabalon]

Give a win bonus of 1/10 or 1/8 points for every win.

Strong players get trumped by weak players who "cheat" the system by piling up meaningless losses.

_______Current-->Fixed

0-13 = 13 points --> 13 points
2-11 = 15 points --> 15.2 points
5-8 = 18 points --> 18.5 points
6-4 = 16 points --> 16.6 points
6-0 = 12 points --> 12.6 points
4-0 = 8 points --> 8.4 points

_________________________
November Alpha "controversy"
_________________________
cut-line 'outrage' -->1/10 point bonus fix

25/3 = 36.5 --> 39
23/1 = 34 --> 36.3
17/6 = 27.5 --> 29.2
16/3 = 24.75 --> 26.35
11/11 = 24.25 --> 25.35
3/28 = 23.25 --> 23.55 # these 3 gamed the system
4/26 = 23.25 --> 23.65 # now only one succeeds
4/28 = 22.75 --> 23.15 #
_cut-line____
11/9 = 22.75 --> 23.85 ## these two got cheated
13/4 = 22.5 --> 23.8 __## now rewarded
10/9 = 22 --> 23 ____### at 1/8 bonus this guy may ALSO MAKE IT
5/11 = 16 --> 16.5
4/4 = 10.5 --> 10.9
1/4 = 5.5 --> 5.6

edited again...

It's interesting...
usagi wrote about the mistakes in the system, about wrong and unfair thinks, there is a poll and the result is, ppl want it this way.
now the system is wrong again, and unfair, has to be fixed (or horrible thinks will happen).
the direction is the same (and the poll made clear this direction isn't wanted) cause some weaker ppl with 10 games more then some strong guys cheated the system and were able to stay in alpha.
...interesting indeed

[Apoah]

Currently in Gamma II I am in the promotion zone with 4/14 my score is better than a 9/3 record and within one point of tying the top two places with 10/0 and 11/1 records.

The two things that should be rewarded in league play:

1) Activity
2) Winning

I have played 18 games, making me the 2nd most active player in Gamma II this month. The most active player has 19 games with a record of 0/19. He is on the bubble for demotion. If this player were to play the rest of the games with the people below him, he would most likely fall into the demotion range if his record continues.

The current system would allow the best players to move forward if everyone played all of their games. Even though it doesn't seem right that a 10/0 record could be only 1 point ahead of a 4/14 record. It is important to note that I have played 8 more games and therefore been almost twice as active in the league.

I think that high activity in the league should be rewarded. Gamma II currently shows how this can work. Currently there are two people below me (one in the red zone) who can jump past me with only 2 more 1st game wins (which they should easily be able to find) and then all will be right with the world. The only real way that I can see to "game the system" would be to somehow hold out and not play strategic people, but unless there were 2 or 3 players working together in one group in one month, this would not work either.

Just one observation. Thoughts?

[topazg]

Currently in Gamma II I am in the promotion zone with 4/14 my score is better than a 9/3 record and within one point of tying the top two places with 10/0 and 11/1 records.

The two things that should be rewarded in league play:

1) Activity
2) Winning

I have played 18 games, making me the 2nd most active player in Gamma II this month. The most active player has 19 games with a record of 0/19. He is on the bubble for demotion. If this player were to play the rest of the games with the people below him, he would most likely fall into the demotion range if his record continues.

The current system would allow the best players to move forward if everyone played all of their games. Even though it doesn't seem right that a 10/0 record could be only 1 point ahead of a 4/14 record. It is important to note that I have played 8 more games and therefore been almost twice as active in the league.

I think that high activity in the league should be rewarded. Gamma II currently shows how this can work. Currently there are two people below me (one in the red zone) who can jump past me with only 2 more 1st game wins (which they should easily be able to find) and then all will be right with the world. The only real way that I can see to "game the system" would be to somehow hold out and not play strategic people, but unless there were 2 or 3 players working together in one group in one month, this would not work either.

Just one observation. Thoughts?

I agree completely with every word

[stalkor]

its funny how people always seem to blame the players that "game the system" e.g. weak players getting in many games. you can turn that back around and say we can blame all other players who did not play enough games! Yes its true that you need less wins to stay ahead of the one who loses a lot but you should still make an effort to stay ahead. i mainly think we have a lot of "lazy" strong players who want to just play a few get in first place and be done with it for the rest of the month but thats not what the league strives for. It's a competition in which you need to prove you are what it takes to be on top of your class!

So we could say we reward activity, but also that we punish INactivity.

[teach]

I think we need to agree to a definition of active.
To me, 20 game played out of 39 possible is "active".
To expect strong players to play any more "teaching" games in a month is unrealistic.

If "teachers" with a ~10/10 record gets cut down by ~3/33 "students", the teachers won't be coming back.

Right now the league is short on dan players.
Does the league want strong players?
How much does the league expect them to give?

Should the same kyu players really be squatting alpha for months on end?
The top class should be a reward for good play, not a safe haven for weak playing system mechanics/tweakers.

We just need to look at recent numbers and close the larger(obvious) loopholes. At least shrink them down a little.

Maybe my 1/8 win bonus proposal is not the right call for all activity ratio.
Maybe just upping the class churn with more promotions/demotions numbers can achieve the same goal.

[Gabalon]

with 50% winratio you're not a "teacher" but normal member of the league...

a "real teacher" with 100% wins can get more points then anyone without wins by playing (and winning) against every member of his class once. so he gets a better result with 1/3 the games a weaker person with 100% losses gets.
getting a 3/33 result is worth less then winning against everyone once too, with 36 instead of 13 games.
if someone can beat 8 of his 13 league mates he can get a better result then the 3/33-guy with 22 games.

activity is rewarded, i won't deny it, but if the stronger ppl aren't MUCH less active the weaker and very active ppl have no chance.


ps:
someone who joined the league one week ago shouldn't tell us the system is wrong, ppl abuse it to cheat and how it has to be fixed...
everyone can start a discussion or suggest changes, but not everything you don't like is wrong and if you make it look that way ppl won't like your ideas no matter what they say, just due to the way they are presented.
the league runs for over 12 months now, thinks were changed, others not and for the activity reward...maybe it's not the only reason, but alpha and beta got much more active the last months and i think weaker ppl (like me) who fight to stay are one part of it.
there was a month with 25 games played in alpha, now we have more then 20 games per person...


edit:
2 things i forgot:

more promotions/demotions...
there were 2/4 promotions/demotions, now it's 3/6 for all classes and demoting more then half the class seems to much, so i think 3/6 won't be changed to 4/8 or so (1 person staying xD)

league short in dans, what does it offer...
the league offers seriouse games with relative long time settings and if someone is strong and active he'll find equal and stronger ppl in the higher classes.
he has to start at a low class and make his way up to beta and maybe alpha, some ppl might not want to do it, but it can't be helped, seeding strong player in higher classes was tested and failed.
there are dans in the league, also high dans, so i don't think the league is "short in dans". the more join the better ofc, but i think on kgs there are more sdk then dans, so i'd expecd the league to look similar and often strong players don't play much, so they might be not interested in the league...
but anyway, the league doesn't "need" ppl of a specific rank, whoever joins is welcome and the active and strong ppl will be in gigher classes after a while.

[topazg]

I think we need to agree to a definition of active.
To me, 20 game played out of 39 possible is "active".
To expect strong players to play any more "teaching" games in a month is unrealistic.

20 games out of 39 if you are a 3d+ will get you to promote all the way to Beta, and probably Alpha. To play more is not required.

If "teachers" with a ~10/10 record gets cut down by ~3/33 "students", the teachers won't be coming back.
Right now the league is short on dan players.
Does the league want strong players?
How much does the league expect them to give?

If the teachers want to just teach, that's fine. If they want to promote, 10 games a month is not enough. We want strong players, but we want active strong players. The league has plenty of dan players. In the top 3 tiers, 33 out of 98 players are dan players (I didn't check them all, just those that had good enough records for me to check or people I already knew were dan players). Ok, that's a top heavy selection of players because they are the top tiers, but out of the remaining 14 classes I suspect we'll have another 20 or 30. This is obviously excluding the superclass which is all mid dan plus players.

We don't expect them to give much, other than play, and enjoy. Enjoy teaching, or play enough to get promoted to the level where the games are hard (I re-iterate, as a 4 dan, this will not be many games. Either is fine. 15/0 is a good enough record to get promoted in pretty much every class, and that shouldn't be too hard to achieve for a dan strength player in the lower tiers - in fact, 10/0 should be enough to get to Gamma as a 4d.

Should the same kyu players really be squatting alpha for months on end?
The top class should be a reward for good play, not a safe haven for weak playing system mechanics/tweakers.

We just need to look at recent numbers and close the larger(obvious) loopholes. At least shrink them down a little.

Maybe my 1/8 win bonus proposal is not the right call for all activity ratio.
Maybe just upping the class churn with more promotions/demotions numbers can achieve the same goal.

I agree with this, and at the moment 4 of the 5 kyu players in Alpha are in the demotion zone. 4 of the promotees from Beta at the moment are dan players. This imbalance will work out with time. Remember the class churn has been increased in the way you suggest this month already, we want to see what the next 3 or 4 months do before making further changes to this, particularly as the league is growing almost exponentially at the moment, making it hard for the tiers to settle down properly.

On the note of activity, congratulations on your 17-0 record. You have gone way beyond the activity required for you to promote this month

[Apoah]

A few random thoughts:

1) It seems to me that if the ASR league stays around long enough, then some of these very active sdk players might one day turn into the high activity teachers in the league.

2) I think ASR has the potential to 'grow' their own strong players "in house".

3) It is a good point that it only takes about half as many games for an undefeated opponent to promote than for a losing record to promote. If a weaker player overtakes a promotion spot by sheer activity then it is usually by less than 1 full point. With one game, the stronger player can achieve as much as the weaker player can in 5 games. This balances itself out.

4) As an SDK I am very excited for the chance to play stronger players in an even game. Every strong player I play does not review and they do not all review fully. Some reviews last an hour and others 5 minutes.


5) I originally voted for one of the scoring changes, but I may have talked myself out of it now.