However, there's a potential problem with your proposal that may lead to the opposite effect. The Main Line forum would emphasize the influence of <RV> and become even more default status. It would be difficult for findings from other forums to have a chance in the voting.

Here's the method I find most useful for selecting my own vote:

1. Find the list of candidate moves (or short lines) that are considered to be our best options.

2. For each candidate, I want to be able to go to one place, read (and perhaps participate) in the discussion and arguments.

3. Eliminate those candidates where the arguments against have been reasonably convincing.

4. Select the move from the remaining candidates where the arguments for have been most convincing and best fits with my own assessment.

The original forum method (when used properly by the participants) really seemed to work well for this.

My comments:

My initial reaction is very favorable and I think your recommendations, for the most part, encompass what I endeavored to achieve with my suggestions. I would add the following caveats and changes.

1) <isemeria>'s comment about granting Rybka and <RV>'s (forgive me <RV>) undue influence is well-taken. <RV>'s Deep Rybka line is not always the *main line* as we saw in this game with <kb2ct>'s game-saving g6! If we keep the forum the same each move, the problem should work itself out and, hopefully, have the added benefit of transferring much of the main line conversation to the fora and away from the main page. In my work on a *model* forum, I have incorporated <RV>'s Deep Rybka in another way.

2) <dotsamoht>'s comment is also valid. Someone (or more likely several people) will have to take over this much-needed and sorely missed summarizer role, although, in your recommended plan with the same forum holding the main line, this job should be easier once we have defined an appropriate format, i.e. less places from which to pull information. Although, summarizing the alternates fora will be more challenging.

3) The forum hosts must be extremely active, diligent and knowledgeable...<zanshin>, <dalbertz>, <hms123>, come to mind. This is a different group than our senior analysts, such as <Tabanus>, <TofK>, <imag>, <ajile>, <whatthefat>, <kb2ct>, <kwid> and many others. My suggestion would be to assign a host, two assistants and at least two permanent senior analysts to each of your fora. At least one assistant to ensure all lines are being posted to the analysis tree by the analysts. Your past experience should be helpful in this regard.

The senior analyst position, however, is a difficult role as it will require objectivity rather than creativity...some very good analysts will shy away from this type of responsibility rather than embrace it. If the role is sufficiently defined, may be a good candidate for rotating assignment. <sentriclub> is good at defining such roles, maybe we could discuss this with him.

4) Any overlap can be dealt with in summarization.

5) Continuity of fora is the top benefit, especially if it aids in summarization and centralization of access to the latest lines.

As always, <kutztown46>, good job, the team owes you a great debt. Any change is slow and evolutionary and we will have to be patient; however, without change we stagnate. :)

You have hit the nail on the head. This pretty much summarizes the state of forums in both games. To his credit, <kutz> is always trying to think of ways to adapt the forum system to the current situation, even if it means (as it does here), that his influence is diminished. He has always been something of a reluctant leader anyway and will do what it takes if it is for the Team good.

Having been involved in forum prep, I have to add that few people realize the amount of constant work and dedication involved. Look in my forum and you will see the analysis I have had to do alone. This does not include that done separately by <kutz> and <hms>. After all the analysis, the forum coordinator then has to check the analysis tree, forums (especially <RV>'s) and main page to gauge the Team interest. After all that comes the subjective part - assigning forums that will be useful but don't appear to be leading the Team down a specified path. Finally, there is the assignment of forums to hosts. Which host wants or needs a break? Which one wants more action? So many considerations to juggle - sometimes within an uncertain time frame. (Suppose the GM moves early?) When people stopped using the forums, <hms> really had no choice but to suspend them.

<Forum A- Main Line > - Prevention of White g4 10...h5 11. Kb1 Qc7 12. Bd3 Ne5 <Kosintseva v. Atalik, 2007>

1) Either Deep Rybka primary line (20+ ply deep), consensus top choice among senior analysts or however we wish to determine (can also be accomplished by designating FC or a committee of senior analysts [<kwid>, <kb2ct>, <ajile>, <imag>, <Tabanus>, etc.)

2) Base line from <RV> establishes top candidate alternatives after the first move.

3) Analysts select a given line to analyze based on predetermined criteria to ensure all bases are covered, e.g.

...One analyst for each engine available, Rybka 3, Rybka 2.2.3, Fritz 11, Naum, Zappa, etc. (this is only an example)

......minimum 20 ply search depth each move, sliding back from last move ......minimum 24 ply search depth infinite analysis
......minimum 20 ply search depth each move, sliding forward and back ......minimum 24 ply human-aided analysis

...Analysis posted to forum in a designated format, discussed, posted to analysis tree, and new lines reassigned for additional analysis.

...12 hours before deadline, available analysis summarized and posted to main forum both as commented (easy to read) lines and pgn ready format.

<Forum B - Conservative Alternates>

Alternate 1 - Allow g4 and Counterattack 10...b5 11. g4 Nxd4 12. Bxd4 b4 13. Ne2 e5 14. Be3 Qa5 <Movsesian v. K. Miton, 2004>

Alternate 2 - Allow g4 and Counterattack 10...0-0 11. g4 Nxd4 12. Bxd4 Bc6/b5 or 10...Nxd4 11. Bxd4 0-0 12. g4 (transposes) <G. Ardelian v. Valentin Raceanu, 2004>

Alternative X -

1) Can be any number of lines that evaluate within a certain number of centipawns from the main line. We can use a similar method of selection as the main line to designate the lines as conservative and aggressive (not a big deal).

<Forum C - Aggressive Alternates >

Alternate 1 - Immediate Attack 10...Qa5 11. Kb1 b5 12. Nb3 Qc7 13. Bxb5 axB <Zsofia Polgar vs D J Zecevic, 1994>

<Forum D - Creative Alternates >

I like the attitude of being open-minded and flexible about how to organize these games.

But I'm a bit worried that with the above forum-assignment scheme, there will be difficulty in reaching a consensus about which forum a specific line should go into. Might not a creative alternative also be aggressive? Who decides which idea goes where? We might just be replacing one headache with another.

I think that our current concept of having an analysis forum hosting a specific line has proven to be effective. However, perhaps forums should be created on a more dynamic, demand-driven way:

- We have a pool of available forum hosts, and one forum (say <kutz>'s) which is the forum center (FC).

- As the need arises, a forum host can offer to host a given line by posting his/her offer in the FC forum. The idea for hosting that line might be the forum host's own idea, or it might have been recommended by one of the analysts. The FC makes sure that two forums don't overlap ideas.

- When the FC receives an offer, he/she might accept it, reject it, or suggest a revision to it. If accepted, then the FC updates the sticky accordingly and announces the analysis forum assignment in the main forum.

This way has a few good points: (1) no analysis forum will be created unless *someone* as expressed an interest in it, (2) you always have a forum host who has taken an active interest in the line (they initiated it!), and (3) the FC doesn't have the burden of figuring out which lines are relevant.

For some moves, there might not be any analysis forums created. Sometimes an analysis forum may be very specific (based on long series of moves), while others may be very general (e.g. "pawn moves").

Now, I suppose this idea is fraught with problems that will be quickly spotted by experienced forum hosts and organizers, but I think the idea is to minimize unnecessary work and maximize usefulness.

~~~~~~~~~~~~~~~~~

Just for laughs, here is one other little idea I had:

We assign a "name" to each proposed line. For instance, suppose we are at move 34 with white to move, and we see 4 alternatives. We give each a name: GMANW34a, GMANW34b, GMANW34c, and GMANW34d. Later, when it's black to move and there are 3 main alternatives, we name them GMANB34a, GMANB34b and GMANB34c, etc.

Everyone who posts analysis about a given line should include the given name somewhere in their post. That way, if someone wants to gather all the analysis posted for a given line, they can do a search on the name, and all the posts relevant posts from any forum will appear.

ccGM Umansky likes to play Catalan

maybe he will try his own line which some of us already know from the loss vs ccGM Thaler:

1. d4 Nf6 2. c4 e6 3. g3 d5 4. Bg2 dxc4 5. Nf3 a6 6. O-O Nc6 7. Nc3 Rb8 8. e4 Be7 9. Qe2 Nxd4 10. Nxd4 Qxd4 11. Rd1 Qc5 12. e5 Nd7 13. Ne4 Qb4 14. Qg4 Bf8 15. a3 Qb3 16. Bd2 Qxb2 17. Bc3 Qb6 18. Rab1 Qa7

I might not have clearly explained my suggestion, or at least that's what I'm thinking based on your comments:

<One of the intended benefits of organizing the forum in the way <kutz> has suggested is to ease the ability to summarize and lessen administrative hassle. If we allow anyone to have a forum then summarization would be impossible.>

I wasn't thinking that <anyone> could host a forum. The pool of forum hosts would be a selected group of people who know how to work it, not unlike what we have now. But if someone gets tired of hosting, then they simply discontinue offering to host analysis. Of course new people could volunteer to join the pool of forum hosts, but that would be under the direction of <kutz>.

And I'm not sure why the ability to summarize would be lost. When a new forum is created, it is added to the sticky and an announcement is made that shows what all the current forum assignements are. So, any summarizer would have access to a nice directory of current analysis locations.

<Too many people are fretting over whether or not we can identify a line as aggressive or conservative...this is nonsense. <kutztown46> can make a reasonable call and be done with it. It matters little which forum it is in only that we have identified it and placed it somewhere where it can be easily found, analyzed and ultimately included in a summary. Conservative, creative, aggressive are all metaphors. We can just call them A, B and C for that matter.>

Sure, <kutz> could place it anywhere he likes. Personally, I would hate to be the one to decide whether a given line should be in the "creative" forum or the "aggressive" forum. I think that decision will often be arbitrary rather than related to the forum title.

But if, as you say, the titles could just as well be A, B, C, etc., then I'm failing to see the organization benefit from this approach. Basically, we're just having some forums share two or more lines, rather than having one line per forum. What administrative hassle is being avoided? And how does this help make analysis easier to find?

sbevan: <kutztown46: <Team> I would appreciate feedback on an idea for the analysis forums for the Umansky game. Read this post:

kutztown46 chessforum>

I think its worth a try. The one time I tried a forum the term (someone else used this but I can't remember whom) "herding cats" comes to mind.

lonepsycho: <kutztown46>

I have not been here a lot lately, but I have seen the development of the forum system since GMAN 1. I would like to comment on your proposed system.

In my opinion, the classification of the forums in your proposed system is too vague. Some might consider a move "creative," while others might consider it "careful."

Also, how are we to limit the number of "careful" or "creative" (or whatever) moves that the posters are allowed to propose? It seems as if your proposed system will flood each of the individual forums rather than flooding the main page.

Also, in the middle game especially, there will be points when multiple main lines (all within 0.01 points of each other with evaluations constantly changing) will exist. Which one is the main line in this instance? Does it change as RV's computer digs deeper?

I cannot think of a solution to the problems I have proposed.

lost in space: Hi <kuztown46>,

I think the idea is worth a try.

Please see the following as input and not as critique:

The disadvantage of our "old" forum system was that analyses disapeared after a few moves (or you have to search for these anylses in various forums to find it again) eventhough a few of the analyses were still valid.

So it was not posible to keep the overview with the forum system alone. This was the advantage of the analyse tree: Everything what is published there can be found easily.

But I was often too lazy to post my findings on the main page, then in the forum and then once again in the analyse tree.

Sometime over the weekend I will post some follow-up comments. Meanwhile, keep the feedback coming.

I like reading this discussion. Very creative and adaptive. Go team!

(plus a HUGE power battle is going down at wikipedia see below...)

<Jimbo Wales: I try to treat objections with seriousness, kindness, and concern. In return, most people are happy to work with me. Others simply choose to spit in my face. I accept that>.--Jimbo Wales (talk) 02:40, 23 January 2009 (UTC) " from http://en.wikipedia.org/w/index.php... (sorry for the linkspam, VERY good arguments from bright people, worth sharing)

I applaud Jimbo Wales (the inventor of wikipedia) for sidestepping bureaucracy and making a great, necessary change). I overestimated our inertia at this site. I hope the team will judge the 4 proposed forums based on a trial, not on speculation.

Very good discussion and I wish that this was on the original agenda. One minor comment, can someone clarify if "creative alternatives" is aka "non silicon alternatives"? I have seen minor communication differences on this semantic.

However, there's a potential problem with your proposal that may lead to the opposite effect. The Main Line forum would emphasize the influence of <RV> and become even more default status. It would be difficult for findings from other forums to have a chance in the voting.>

I am sympathetic to the idea of assigning people to work in each of the forums. However, I don't think that will happen. In the GMAN2 game, I attempted to recruit 24 forum assistants. I worked very hard at it and only found maybe 5, most of whom did not stick with it.

I agree that the biggest weakness of the proposed system is the potential for overlap. I can take some steps to minimize that. In your example, if the creative aggressive alternative is rated reasonably well by the engines, I would suggest that the aggressive forum take it. If the move or line is more speculative, I would try to nudge it into the creative forum.

<However, perhaps forums should be created on a more dynamic, demand-driven way:...>

I'm afraid that this idea reminds me of the beginning of the GMYS game, where <Artar1> assigned forums based on volunteers. One of the main lessons of that game was that the forum system works MUCH better when forums are assigned by the FC. The new proposal in a weird sort of way takes that a step further by ensuring that the analysis forums are more permanent.

<What administrative hassle is being avoided?>

I am open to a better set of descriptions. As I said earlier, the potential for overlap is the biggest problem with my proposal. However, I hope to be able to minimize it by carefully defining the four forums. I think I can define them in a way that they are mutually exclusive. But there will be gray areas because of different opinions. If one person thinks a move is aggressive and someone else thinks it is careful, then I don't know what to do about that.

I am hopeful that the strengths of this system would outweigh the weaknesses.

<Also, how are we to limit the number of "careful" or "creative" (or whatever) moves that the posters are allowed to propose? It seems as if your proposed system will flood each of the individual forums rather than flooding the main page.>

I don't think we would try to limit the number. If we end up flooding the forums, I think that's an improvement. We WANT team members to use the forums more extensively. One of the motivations for this new idea is to take away some of the disincentives of using the forums.

<Also, in the middle game especially, there will be points when multiple main lines (all within 0.01 points of each other with evaluations constantly changing) will exist. Which one is the main line in this instance? Does it change as RV's computer digs deeper?>

That's a good question. We will have to agree on some convention to handle that. I would think if there is a body of analysis on a certain move believed to be main line, and then a different move emerges at a higher ply, then the main line would stay the same. That's open to debate.

<However, might I suggest assigning individual summarizers for each forum. This way, one summarizer does not have to sift through many forums' worth of knowledge in order to present an in-depth summary. If we have a "creative" move summarizer and a "main line" summarizer for each move, the burden is slightly alleviated from the summarizers. (This role could rotate every 6 days so that there wasn't only one "main line" summarizer.)>

That would be wonderful. I am not highly optimistic about finding team members who would be willing to do that.

I know nothing about Jimbo Wales or wikipedia, but I endose that statement.

<Very good discussion and I wish that this was on the original agenda.>

Thank you. It was not on the "original agenda" because I only thought of it three days ago, as I was mulling over <capafan>'s ideas.

<One minor comment, can someone clarify if "creative alternatives" is aka "non silicon alternatives"? I have seen minor communication differences on this semantic.>

It was such a waste for you three to do so much work, but the forum systems getting an average of 0.6 posts per forum per day.

The wikipedia argument is something like: <...but we only have 61% supporting a new radical proposal that will help wikipedia; clearly concensus is 70-75%>

The counterargument is <if a controversial solution is needed to fix a serious problem, then a supermajority will never agree to it ... how many things would get done if a 75% threshold is needed to break a bureaucratic inertial deadlock>

Jimbo Wales is being called a dictator because he only has 60% support before doing one of the biggest policy changes to effect wikipedia. From now on, you can't go to an internet cafe and anonymously vandalize your senator's wikipedia biography. With Ted Kennedy's collapse and Senator Byrd's collapse, some people from starbucks anonymously reported that they both died. My study of undermining bureaucracies is fueled by my annoyance by them. Rigidity is for a non-existent declining social intelligence medical issue. Imagine if every generation, humans shaved off 10% of their intelligence. Then a bureaucracy would be justified and perhaps brilliant.

However, they spring up everywhere and they grow and leech precious societal resources. Congrats to Zanshin, Hms123, and Kutztown. Welcome to the UMAN game without those needless shackles!

By rehashing our discussions we overcome our mental biases, therefore, I welcome your comments.

I reread your comments and hopefully better understand your suggestion. I do not believe any of us are too far apart in our suggestions and concerns.

Many of the concerns center around the continuing reliability of those working within the forums and the feasibility of attracting and retaining qualified individuals. If they had the degree of motivation, permanence, diligence and continuity of <RV> and <kutztown46> we would not have as many concerns. Regardless of the form, without the commitment of a forum host and assistants, the system will fail.

In realization of this fact, I would like to propose the following.

Those of us participating in the discussion achieve a working consensus of what we believe to be an optimum organizational model without regard to the implementation issues. A full blown attempt for the FC at assigning forum hosts as well as the organizational structure underneath is unworkable. Rather we push this task onto the forum hosts, who must agree to actively recruit their our own assistants, analysts and summarizers rather than asking for volunteers and agreeing to fill in the gaps as needed. This is a significant commitment, but critical if the system is to succeed; however, we found the initial two in <RV> and <kutz>, surely we can find five others amongst 300 recurringly active team members. I will volunteer as one of the five.

As for as recruiting assistants, analysts and summarizers I will most likely rely on a pool and not a permanent cadre of talent. As these helpers come and go, I will no doubt be required to fill in the gaps myself when necessary. <kutz>. I believe this is consistent with your concept of sub-groups.

If the system fails, I will have the satisfaction of knowing exactly why. In addition, at least I can say I am not afraid of backing up my recommendations with a personal commitment.

As to specific comments, if conservative, aggressive, and creative are improper metaphors, we can agree to label them in some other descriptive manner, although I do not believe the substance of the categorization changes despite the label. Everyone agrees we should have a main line based on a human assessment of our silicon friends top line, Rybka, Fritz, et. and a category for lower evaluated lines that may involve more human creativity. What we call those lines in between that are secondary, tertiary, etc. that show up as 2., 3., 4. recommendations by Rybka, Fritz or whatever, does not matter. We should make a call, go with it and weather whatever linguistic criticism we get on the semantics.

Suppose <ajile> hosts the main line forum and <Dionyseus> hosts the aggressive forum. Maybe we start the system up at move 5. Maybe up thru move 9 we play the main line move. Then, on move 10, the aggressive crowd convinces the team to vote for an aggressive move, rather than the main line move.

Now what? Does <ajile> continue to host the main line forum and <Dionyseus> the aggressive forum? If we do that, we will lose some continuity, because all the analysis dealing with the new main line will be at the <Dionyseus> forum, and the aggressive forum would begin looking for something more aggressive than the new main line.