|
< Earlier Kibitzing · PAGE 34 OF 91 ·
Later Kibitzing> |
Aug-26-08
 | | kutztown46: <Xenon Oxide>
What kind of communication do you mean? The actual <crying> part of your job will of course be done on the main page. But you may certainly post here anytime you wish. |
|
Aug-26-08
| | kutztown46: <whiteshark>
Thank you for the honest feedback - I appreciate it.
There are hundreds of us on the World team. Each of us has a unique set of talents, interests, and life situations. I understand that many will not wish to be <tied down> to a certain task. I also want our best analysts to be free to pursue the most important lines and not to be encumbered with administrative tasks. On the other hand, there are many of us who desire to help create an infrastructure that will enable the team to thrive. I hope to convey the message to the team that the idea of Forum Assistants is tied to the philosophy of "many hands make light work". If I can recruit 24 Forum Assistants, none of the FAs and none of the forum hosts will be seriously overburdened. |
|
| Aug-26-08 | | dotsamoht: <kutz> I have made changes to The Sticky per your direction. I have organized the order of things in what I believe is a logical fashion: greet new members, ask them if they want a job, then discuss openings, describe how engines work, etc. Let me know what you think. |
|
| Aug-26-08 | | zanshin: <kutz> Like you, I appreciate <whiteshark>'s candid response. I am pretty sure he's not the only analyst who feels this way. While the 'roaming analyst' will always serve an essential role in this game, his sentiment reinforces our main problem: you cannot make analysts commit to a line where they have no personal interest - at least not for an extended period. I got your email on Rybka 3. Fyi, I have heard it does not go down ply levels as fast as Rybka 2.3. I'll confirm this with <RV> together with other questions I need to ask him. |
|
Aug-26-08
 | | Open Defence: <Boss> wonder if we should divide the work on the Najdorf, could be a job for the analysts and summarizers
though most of the work would need to be with reference to databases and annotated games IMHO |
|
Aug-26-08
| | kutztown46: <Open Defence>
Are you talking about the library, or is this a restatement of your alter ego's question on the main page? |
|
Aug-26-08
| | kutztown46: <dotsamoht>
Looks great! Can you add "The List" (Hootey)? Thanks. |
|
Aug-26-08
| | kutztown46: <zanshin>
Speed is relative, but I am expecting to get results a lot faster with Deep Rybka 3 on a quad core than I did using (non-deep) Fritz 10 on a dual core. If that does not happen, I'll be up against it with two days per move. |
|
| Aug-26-08 | | zanshin: <kutz> You're obviously right about speed being relative. After you get Aquarium, you might want to see how Rybka 3 does in comparison to the now free Rybka 2.3. I'm just a little concerned that I may not be able to use my 25-ply rule of thumb for analyzing any single position, especially with the 2-day move rule. I'll let you know. |
|
Aug-26-08
| | Open Defence: restatement of what <Red October> said kind of like research and analysis before the Opening Library fully takes off but the library can host it |
|
| Aug-26-08 | | sentriclecub: Hey, I found out that we overlooked a chessgames.com page solely devoted to the 6. Bg5 Nadjorf. I volunteered to allow people to feed me any comments they want, and I'll sort them and turn them back over to the team if we ever reach the Najdorf (which is at best a 50% shot--probably is all wasted effort) Isn't if funny that I overlooked Sicilian, Najdorf (B94) Do I have permission to add
6. Bg5 of the Sicilian, Najdorf (B94) into the sticky? I will add it, but anyone is free to remove it, because I can't remember the sticky manager--I think it is dohtsamnot. |
|
Aug-26-08
| | kutztown46: <s>
No problem adding the link, although we should let <dotsamoht> maintain the sticky. If we end up in B94, I'm sure <Open Defence> will have this link in the library, also. Remember, pages like this are not closed to view from GMAN so discretion should be used in kibitzing there. |
|
| Aug-27-08 | | zanshin: <kutz> I've started playing around with Aquarium. You can tell it was designed by analysts. They probably all sat down somewhere and drew up a wish list of all the features they wanted in a chess program. I have confirmed that Rybka 3 does <appear> to run more slowly than Rybka 2 if you base your assessment on node count rate and ply depth. Vas states that they cannot be compared between different versions. You know that a ply-depth of x+1 in Rybka 3 is deeper than x, but cross-engine comparisons are meaningless. As for node counts per second, I would get 300-400 kN/s in R2.3.2 on my Quad, compared to <100 kN/s for R3. I think I'll have to throw out my 25-ply rule. I might have to base the decision on whether anaysis is sufficient on time spent analyzing a position rather than ply levels. Persistent hash should improve the quality of analysis. One nice feature though is that you can specify a "delta" value which corresponds to our playability calibration. So, if you have delta=1.0, and have specified 10 PVs, any lines evaluating at less than 1.0 of your top line are not even shown. Right now, the results of analysis are not what we are used to. Here's 5-minute run on my quad using the terminal position of the GMT game: [+5.12] d=15 62...Ne5 63.Kd1 (0:04.29)
[+5.12] d=14 62...Ne3 63.Rb5 (0:01.42)
[+5.24] d=14 62...c2 63.Kd2 (0:01.52)
[+5.43] d=14 62...Nh2 63.Rf5 (0:01.59)
[+5.73] d=14 62...Ke4 63.Rxa5 (0:02.12)
[+6.11] d=14 62...Nf6 63.Rxa5 (0:02.47)
[+6.14] d=14 62...Nf2 63.Rxa5 (0:03.20)
[+6.24] d=14 62...Kc4 63.Rxa5 (0:04.27)
Ply level per move is shown, which is nice, but no game/engine details or fen diagram. I have asked an Aquarium developer if there's any way around this. I'll cross-post to <RV> and see if he has any suggestions. |
|
| Aug-27-08 | | truefriends: If it isn't a problem being away for a few days once in a while (mostly at my girlfriends home, not having any chesssoftware) i would like to volunteer for a spot of <Temporary Forum Assistant> in our team. <truefriends>
|
|
Aug-27-08
| | kutztown46: <truefriends>
Thank you! I'll let you know later which forum host you will be assisting. |
|
| Aug-27-08 | | truefriends: Q: am i allowed to do some analysis on my own forum on other moves than on the forum i am assisting? |
|
Aug-27-08
| | kutztown46: <truefriends>
Absolutely. And thanks for asking the question. Maybe others are wondering the same thing and not volunteering because of it. When I get time, I need to make another post and better explain the role of Forum Assistant. |
|
| Aug-27-08 | | dotsamoht: <kutz> Just checking in... I update the Sticky after GMAN's move, right? Do I take any action after our move? |
|
Aug-27-08
| | kutztown46: <dotsamoht>
Once we kick off the temporary forums, they can and often do change both after our move and after his move. Until we start the temporary forums, you just need to add things as they come up. Things will come up at irregular intervals. |
|
| Aug-28-08 | | Xenon Oxide: <kutztown46> I don't think that the system of Temporary Forum Assistants is gathering much interest. I as the town crier will post a request on the main page to garner interest in the position. Is that alright? |
|
Aug-28-08
| | kutztown46: <XO>
Let me handle that, hopefully today sometime. |
|
| Aug-28-08 | | zanshin: <kutz> As promised, a couple more observations on using Aquarium. Here's sample evaluation I ran last night on a game from <ganstaman>. I ran it for about 30 mins on a dual core. The first observation is that the ply levels didn't get very far. We should expect that now. More interesting is the disparity in depths. Previous versions analyzed all lines to the same depths, such that you knew all ply levels were the same, or the familiar grayed out ones were 1 ply less. Here, they range from 14 in the promising lines to 5 in the unpromising one. One problem is that the engine and game specs are not included, nor the fen diagram. If you select the top line to clip, it will show you then fen. However, if you select all lines, it does not. The Aquarium developers told me there is currently no way around this except to paste twice. However, the feature of including all info + fen is scheduled for the next Aquarium update. It looks like there have already been two since the release. >>>>>>>>>>>>>>>>>>>>>>>> [+9.78] d=14 16.Rh8 (0:11.11)
[+3.70] d=14 16.g6 Rf5 17.Qg4 Nxe5 18.Qxf5 Nxg6 19.Qxg6 Qe8 20.Qh7 Kf8 21.O–O–O Qf7 22.Rde1 Bf6 23.g4 (0:15.24) [+2.11] d=11 16.Qh5 Nxe5 (0:00.42)
[-2.13] d=5 16.Qd3 Ba6 (0:00.00)
|
|
Aug-28-08
| | kutztown46: <z>
I have not received my copy yet, but it has been shipped. I would be interested in knowing how deep how fast it gets on a quad in an early middle game position when there are lots of pieces on the board and many playable moves. |
|
| Aug-28-08 | | Karpova: <kutztown46: I would like to clarify something. A few weeks back, you said on my forum that you were willing to host a temporary (analysis) forum during GMAN2. Then last week, you made the post on the main page quoted above. Are you willing to be an active forum host as you were last game (updating your header as required, participating in the discussion, etc.) or are you instead offering your forum for analysis if needed but have no plans to be active yourself?> I strictly separate between maintaining a poll and being a forum host so the fact that I won't go on doing the former doesn't have any impact on the latter. My forum can be used for analysis and I will update the header. I will be an active forum host if necessary. |
|
Aug-28-08
| | kutztown46: <Karpova>
Thank you. That is what I thought but I wanted to make sure. |
|
 |
|
|
< Earlier Kibitzing · PAGE 34 OF 91 ·
Later Kibitzing> |
Kibitzing
A majority vote for 1-0 means we resign. 