what is the ASR doing wrong?

[usagi]

seriously, stop proclaiming yourself as if you are always right and just form an opinion that doesn't belittle others.

You're absolutely right that I can be wrong, and I very well may be wrong. It would also be nice if you explained to me why I was wrong. What's the rationale behind the 2-1 ratio? Do you think it makes any sense at all to change it? I'm curious. I think it would be interesting to take a look at it. From a logical standpoint it is probably not the most efficient number to use. That's all. The rationale can be that "it's my league" or "this is how we do it here". It very well may be that. I'm just asking. No one is going to think less of you for saying that.

The point of the current structure is not only to give an advantage to ppl who win league games but (and this i find very important) is to prevent strong lazy players.

Thank you for clarifying that. At any rate if I was ticking you off earlier I'm sorry. My intent here is to try and look at things numerically. I'm a bit of a stats freak. I just want to know if you've run the numbers and if you feel that 2:1 is "close enough" that it's not a concern of yours right now. For example I tried to open-mindedly examine what would happen if we went to a 1:1 win-loss point system. It would promote activity. I'm wondering if that is something that might interest you. I'm just tossing out ideas. I think that should be made clear if it isn't apparent. I am not flip flopping, saying 4:1 in one post and 1:1 in another. I'm trying to brainstorm.

A player who is 4d+ should never go thinking ''oh i will churn out a quick 10 games and win easily, because no one can touch that''. Instead they should aim to play a lot and win that way. The league is based on points, and activity is key to getting those, changing it in favor of a winner too much creates laziness (see the poll i had on the ASR site, many players use the league as a way to make themselves play games!)

playing games is vital to your growth in go and it should be motivated to play games, not be held back(a reason i like points over winning%, but thats a whole other topic)

i know that a lot of you have busy lives and jobs but then accept you will not be in alpha or beta but be content that the league is there to make you play at least some games, maybe you wouldnt have played if you weren't in the league.

These are all interesting and valid points, thank you for raising them. I actually agree with everything you have said. In my heart I just wish there was some way to make it so that the 4d+ player doesn't say that because alpha is filled with 4d+ players. Yes I know, it's wishful thinking to pretend we can wave a magic wand and make that happen. And hence the current system. In the end, the ASR is a great place stalkor, don't let a few poorly phrased comments from me piss you off. If I didn't agree with how you ran things I wouldn't be in the league.

[ez4u]

Blah, blah, blah...

Blah, blah, blah...

Anyone who has read the ASR threads has to realize that you two guys have a history. But guess what? It does not help!

The current discussion is about improving the experience of hundreds of real people who are members of the league. Sorry, but we really do not need more sniping in the usagi/stalkor feud. Actually, unsupported opinions about what could be different do not help. Please be more specific! Why would a specific change be better? What would the impact be, based on the actual results in the league in previous months? If the scoring changed, who would have won in alpha in October, and why would that have been better? Who would have been promoted from gamma in September, and why would that have been better? Who would have been demoted from beta in August, and why would that have been better? Until we have some concrete expectations that can be discussed and presented to the members, we really do not have anything. Both of you have done a lot for the ASR over the years, but it makes my teeth hurt when you square off against each other... again!

[stalkor]

I just got my first delta resignation from the league because of the new format.

this player played actively last month but did not get promoted with a high number of games. he finds the new format demotivating for any player who is serious about playing but is not able to play great amounts of games. As you can see from the current results the top 3 already played over 20 games and the league is only a week underway! if this continues they will have played 80 games over the month, accumulating about 120 points or more (untouchable)!
I think (this is an estimated GUESS) that a top 40 spot will need at least 25 points. This doesn't mean that you gotta play competitively but you gotta play play play play play to get in gamma. Is this what we want? Is this a quantity over quality thing now?

my opinion now is this:

yes, the new players can easily find games in the new format. maybe its a bit TOO easy now. I have seen a couple of games and i feel that there are less reviews, less bonding moments(how do you bond with a group of 140+ ppl anyway?) and as result less serious games, because a mindset is created where 5 games is better then 1 serious game.
another worry is when they enter gamma in the "real" league format. the players who played 50+ games will get demotivated in a day probably because one moment the could play a possible 300 games, and in gamma a possible 40.
bottom line, i'm not sure this huge delta class will help the league overall, it feels like you're putting out a forest fire with a glass of water.

[ez4u]

Sad to see that at this stage of the game. Hopefully most people will stay (or least check back) to see how it all comes out.

There is more than just Delta that is interesting at this point in time though. Look at Alpha. The top player has already played 24 games and everyone in Alpha has already played. The top players in Alpha are going to run out of opponents soon and we are just finishing the first week.

It is true that the most active players in Delta are setting a high bar for promotion. However, the promotion zone has forty players. Right now the bottom of the green zone is 7 points in Delta. But look at the bottom of the green in the other classes:
Beta I - 14 points (!)
Beta II - 6 points
Gamma I - 3.5 points
Gamma II - 10 points
Gamma III - 7 points
Gamma IV - 6 points
Not to mention that the top three in Alpha currently fighting for the top prize have 23, 22, and 19 points already.

At this point in time Delta is not distinguished by any special results. What is different this month is that the active players in the other classes will be forced to stop soon by a lack of opponents but that will not necessarily happen in Delta. That brings us back to a discussion of the league's priorities. Is the emphasis on activity, winning, or some other result (e.g. quality reviews)?

Last month 54% of the players (170 out of 312) were in Delta and they played 47% of the games (600 out of 1272). As of today, 51% of the players are in Delta (149 out of 293 but registration is still open) and they have played 52% of the games (272 out of 522). The 522 games played by the league as a whole in the first week is equal to 41% of the total played in October. If we take into consideration the smaller number of members this month (so far), the number is the equivalent of 44% of last month's games (41%*312/293). It would be interesting to know how representative a figure that is for the first week's play. To what extent do the current class sizes and game limits restrict the ability of members to play as many games as they would like? It is too early to see the extent of the effect in November, but it would be interesting to be able to compare whatever happens to the experience in prior months as well.

Just some thoughts after week one of the grand experiment. Don't despair stalkor... yet!

[usagi]

Right, don't despair, just see how it works out. 1 out of 400 isn't something to worry about.

Although being blunt for which I am known, I find it incomprehensible how you can raise the question of quality over quantity on the subject of activity rewards after all this time ^^

I've always said that you should make more decisions and take stronger action on that. But you have done that and now isn't it time to just wait and see whence the cards may lay?

That being said how about something completely different. Prizes. Taking a look at last month's prizes, I noticed something odd. I noticed that it appears as if the "strongest" players are receiving prizes (although there are also prizes for the most active) but that there are no prizes for people like tabemasu who donate teaching games, and no prizes for the guy who gave the most reviews, etc.

It might be something to look into, Stalkor. Giving rewards to people who donate their time, hmm, maybe there would be some way to honor someone who donates reviews or teaching games? Perhaps having a monthly spot like "Teacher of the Month" with a small interview on the site page, you know what I mean? That would help the teachers, too. It could be a win-win.

I am sure you know I also suggested that you auto-link all reviews which were given on the results page, and/or feature a "review of the week" article on the site page. I think it would really improve the community here which seems to have (somewhat) degraded over the past year. For example as a member of alpha less than 40% of my losses are reviewed, so it might be a good idea for people like me if you call attention to good reviews given by others to players of similar strength, and so on.

Anyways, over and out.

p.s. please feel free to ignore me, of course I know you know that, just saying.

[jts]

Don't underestimate the value of experiments and option-value. If the big Delta is a tremendous failure, it's a tremendous failure for thirty days, and then we go back to the old system. If it's a huge success, it's a huge success for for the rest of the life of the league. If we hadn't tried it, we would never know.

Maybe part of the solution is for people lurking in the ASR between games to slip in at the end of Delta games and take over the review.

[ez4u]

I want to post a little update on the situation as I see it so far this month.

I used "Last update was 10-11-2011 22:59:01 GMT" to gather some data a few hours ago. There were 302 people in the league and 725 games had been played. Note that the figure represents 72.5 games per day so far in November. This compares to 41 games per day in all of October. However, there were only 165 games played in the last three days, 55 per day. Does that represent a slowing after the beginning of the month rush, the midweek blues, or what (no idea here)? If things in Alpha - Gamma slow down because the most active players run out of games to play (the most active Alpha has only six more possible games remaining), I'll be interested to discuss that at the end of the month.

The different classes vary widely in their experience so far. Below is a table of the situation by class. It shows for example that in beta I (B1) out of 20 people, 11 have qualified already, 4 are active but not yet qualified, while 5 are still inactive. There have been a total of 65 games played in B1. Down near the bottom there is a subtotal row for the classes except delta. Across all of alpha, beta, and gamma 44% of the members have qualified, 35% are active but not yet qualified, and 21% are still inactive. Meanwhile in delta 41% of the members are qualified, 30% are active but not yet qualified, and 29% are still inactive. Finally, delta represents 52% of the members, 50% of the qualified members, 48% of the active but not yet qualified members, and 60% of the not yet active members. Meanwhile 54% of all games have been played in delta.

Up to now the promotion zones have continued to vary significantly without delta standing out as unusual:
Alpha (highest score within the demotion zone) 12.5 points
B1 (lowest score to promote) 17 points
B2 11 points
G1 8 points
G2 15 points
G3 7 points
G4 6 points
Delta 10 points

The activity is very different comparing across classes. The most active member of Gamma II has played 21 games. Six other people in Gamma II have played ten or more games. No one in Gamma I, III, or IV has reached ten games yet. The two Beta's are similarly distinct with six people in Beta 1 over ten games (top = 23) versus only one person in Beta II. What do we think accounts for these types of differences at the same level? Are people in the less active classes dissatisfied or satisfied with the situation? Since the class membership is shuffled each month (as I understand it), it seems likely that people will switch from more to less active classes and vice versa fairly frequently. Has there been any consistent view, complaint, discussion on this phenomenon?

[ez4u]

...

That being said how about something completely different. Prizes. Taking a look at last month's prizes, I noticed something odd. I noticed that it appears as if the "strongest" players are receiving prizes (although there are also prizes for the most active) but that there are no prizes for people like tabemasu who donate teaching games, and no prizes for the guy who gave the most reviews, etc.

It might be something to look into, Stalkor. Giving rewards to people who donate their time, hmm, maybe there would be some way to honor someone who donates reviews or teaching games? Perhaps having a monthly spot like "Teacher of the Month" with a small interview on the site page, you know what I mean? That would help the teachers, too. It could be a win-win.

I am sure you know I also suggested that you auto-link all reviews which were given on the results page, and/or feature a "review of the week" article on the site page. I think it would really improve the community here which seems to have (somewhat) degraded over the past year. For example as a member of alpha less than 40% of my losses are reviewed, so it might be a good idea for people like me if you call attention to good reviews given by others to players of similar strength, and so on.
...

What is the story on reviews? First, are they logged automatically? For example, I play a game in delta. It has "ASR league" in the chat and is picked up by the ASR script. A link to it is embedded in the results page (I assume all automatically). At the end of the game we review. KGS prompts me and I save the record at the end of the review. That also has "ASR league" in the chat so is picked up by the script? Is anything done currently with the reviews? Would it be difficult to create a "Reviews" page that posted links to new reviews as they appeared? Of course if it's reasonably easy, the immediate next question would be whether it is also relatively easy for people to vote for their favorite reviews?
Is this potentially a platform for the above-mentioned prizes for best reviews, reviewers, etc.? Of course just the review page and social recognition might be sufficient to draw some stronger players into doing/joining reviews for other games. How much along such lines has been done in the past?

[jts]

On reviews:

In general, if the rank difference is =<2 stones, the loser doesn't want to review. This is a fairly general pattern and I don't think we necessarily want to be pressuring people who are close in rank to review with one another.

Letting people vote on good reviewers sounds like a brilliant idea, though. Probably would work better to let each person nominate one good reviewer.

In divisions with really high activity, the engine often seems to be one or two players who are on constantly and enable all of their opponents to log two games quickly.

[daal]

As someone who feels that the game reviews are the best part of the ASR, I think that the emphasis should be less on creating an archive - The Go Teaching Ladder already does a great job of this - than on increasing the percentage of reviewed games.

If we compare the reviews on the GTL to those on ASR we see that the GTL reviewers are outside observers who have taken time to think through their comments as opposed to the ASR reviewers who have just played the game and offer mostly their spontaneous observations. This is not to say that the ASR reviews are not valuable, it's just that due to their nature, they are often one-to-one conversations which for third parties make them less interesting and harder to follow than a GTL review.

The ASR reviews however have several advantages over the GTL.

Immediacy: The reviews take place when the game is fresh in the player's minds.
Regularity: They encourage the good habit of re-examining one's ideas, and make learning from a loss more probable.
Diversity: The players get to hear a wide variety of opinions.
Reciprocity: Players learn both by giving and receiving reviews. Those who don't have the confidence to give a GTL type review get to practice reviewing.
Camaraderie: The review period is where the players have an opportunity to get to know one another.
Quantity: The more the better, no?

The current structure of the league places value on the quantity -awarding points for losses- the quality -slow games, reviews encouraged- and the winning -more points, prizes- of games. This makes sense because the league's goals (I assume) are to better the learning experience. What I would like to see, is for the ASR to encourage reviewing by somehow placing similar value on the review process.

Regarding quantity, one way would be to include the review in the point system - in other words, games that get reviewed could be worth more points than un-reviewed games. As for quality, it is difficult to put a number on the quality of a review, but it would be nice for the reviewee to be able to publicly express their appreciation of a particularly good review or reviewer. This might take place in the form of likes or stars on the website, and also by letting people vote on nominated reviews for an award such as usagi's "review of the week". As to winning, we know that one shouldn't try to "win the review," but if the reviews help us win games, friends and status then they should be better integrated into the structure of the ASR.

[stalkor]

today after 15 days of the league (exactly halfway) we broke 1000 games!!

thats 66.66666666 games per day which is 20 more on average per day.

comparing the stats:
delta played 546 games (54%) >< alpha to gamma played 456 (46%)
delta has 78 active (4+) players >< alpha to gamma have 91 active players
delta has 37 players with 0 games >< alpha to gamma has 24 with 0 games

[daal]

delta played 546 games (54%) >< alpha to gamma played 456 (46%)
delta has 78 active (4+) players >< alpha to gamma have 91 active players
delta has 37 players with 0 games >< alpha to gamma has 24 with 0 games

And: alpha - gamma 146 players, delta 170 players

[stalkor]

average points delta promotee in october: 22,61
average points delta promotee in november(17th): 28,54

although there seems to be a lot more activity in the top 40, the number of games is now at 584 (thats 3.39 per player) and imo a bit on the low side

I would like to ask here what you think are the milestones for the giant delta class be named "succesfull"

I feel the following should be weighed:

- Compare number of dropouts in delta to the previous months. (oct:69, sep: 77, aug: 50. avg->65)
- Competitiveness of delta, it shouldnt be a meaningless game. It should be a game that makes you want to play as well as you can because you really want that win.
- availability. even though delta now has over 170 players a player named baconade was unable to find a game yesterday (16th) while others have played over 50 games now.
- Number of games played. should at least average out at 4. this has been about the normal for the league in the previous months, and with such a big class it should not be a problem
- The experience. although highly subjective i think i should not simply do math and conclude from there, but have the actual players who play in the league have a voice. Although i feel that most ppl who post will either be on either side of the extremes id still like to hear it.

[stalkor]

poll is up on the site about delta here. if there is something i should add, contact me

[ez4u]

I have pulled together the statistics after three weeks. First is a table comparable to post #112 above. The currently 324 members have played 1,230 games to date: 544 in alpha, beta, and gamma combined (ABG), and 686 in delta. The breakdown by class is shown. For reference in October 312 members played 1,262 games: 662 in ABG and 600 in delta. The current delta has exceeded the October full-month game total by close to 15% after 3 weeks. Meanwhile, ABG are about on track to equal last month's result.

More interesting perhaps is some analysis of what has happened week by week. The second table shows that in alpha 74 games were played from November 1 to November 7 (week 1), 43 games were played from November 8 to November 14 (week 2), and 22 were played from November 15 to November 21 (week 3), and so on. The 74 games in alpha in week 1 represented 10.7 games per day and 1.17 games per active player (i.e. those players who had played at least one game by the last day of the relevant week; so the number of active players changes week to week). There are subtotals that combine the two beta classes and the four gamma classes. These are in addition to the subtotal for ABG.

We can readily see that the activity declines significantly from week to week. This has reached rather extreme situations with gamma 1 and 3 where only 6 and 2 games respectively were played in week three. Currently the chances of finding a game in gamma 3 are essentially nil. Among the smaller classes, only alpha is maintaining a level of activity ('per active' basis) that exceeds that of delta. Beta has a level of activity that is similar to delta's but lags a bit. Gamma on the other hand is trailing by quite a bit. Gamma 2 has kept up but that is only one out of four gamma classes. Although not shown explicitly in the tables above, the current level of inactive players is the same in gamma (combined basis) as in delta = 20%. While the qualitative results from the delta experiment have yet to be gathered, we should perhaps also be looking at the other classes and asking whether the experience there is sufficiently appealing to sustain the league or not.