I am dimly aware that cars, like trees, come in different types, but I see no need to learn their names or to differentiate between them.

Can we please have a forum for the <2011 FRENCH CHAMPIONSHIP>? 2700s Vachier-Lagrave and Bacrot are playing.

Thanks

However, conceivably, for some players we could have the simple name for search purposes, but--for *display* purposes only--we could use the fancy-version. Maybe that way searches could be easy but display would be perfect.

There are a few exceptions even now. Our page on François André Philidor uses non-English characters; consider it a kind of an experiment. The conclusion? It's caused no ends of bugs and complaint emails all stemming from the use of non-standard characters. Clearly a better method is needed.

<A.) The Crafy engine that you use (here) to occasionally post analysis?>

Well, like all engines it is stronger when it looks at more nodes. We intentionally make it run sort of slow (about 500 million nodes per second) because it runs on the server and we don't want to delay web requests due to Crafty.

Note that every time Crafty makes a statement it shows the number of nodes it looked at as well as a depth. If it says 25M nodes and a depth of 12 ply, then I would say it's probably making a statement at a 2000-2200 level or so. If it says 500M nodes I'd guess it's around 2400-2500 level. If Crafty can play at a GM level I suppose it would come around a billion nodes or so. This is just my guess, based on what I've seen other engines do.

<B.)"Lil Chess Partner?" >

Now we're talking apples and oranges--Little ChessPartner runs on YOUR machine, so how strong it is depends on your platform. However, it runs in Java, which is an interpreted language. Therefore even if your machine is a exceptionally fast LCP will be woefully slow, and probably only play at the level of an strong expert at best. On most people's machines it might not even play that strong, but of course even the weakest computer can be a tactical genius when it comes to two-move zingers.

Could diacritiqued (is that a word?) letters be deemed to be one letter for purposes of the search function, or is that a matter of <been there tried that> or even <don't be silly, it doesn't work like that>?

TWIC PGN - http://www.2seeitlive.co.uk/twic/as... >

Merci.

You have the included the <CLASSICAL> games from the China vs Russia (2010). Could you please also add the <RAPID> games?

Thanks

Chessbase Men's all games PGN ->
http://www.chessbase.com/news/2010/...

<"Well, like all engines it is stronger" ... <<snip>>>>>>>>

Thanks for the answer, (!!!) I am thinking that you probably UNDER-estimated Crafty while slightly over-estimating LCP ... but its more of a hunch than solid proof ... just one man's opinion ... for what its worth.

7th CHN-RUS Rapid (2010)

♔ QUICK SUMMARY ♔

Some pages (a) in the OE now will have interesting facts that go beyond the normal # of games and year range.

These data include:

<Notable Games> - A list of up to 5 games that contain the diagrammed position, and are "notable" in the usual CG sense of the word. (b)

<Game Collections> - A list of up to 3 game collections which either exclusively or nearly exclusively feature the diagrammed position. (c)

<Debut> - The game in which the diagrammed position was first seen(d); in other words, the move you just clicked on was first played as a novelty in this game.

<Recent> - A recent example of this position in the database.(e)

This is the first significant update to the OE we've made in a long time, and we think the "notable games" and "game collections" are both very cool applications of what has lately been called "crowd-sourcing."

Although this feature is mostly designed for premium members to enjoy, non-premium members can use it as well by checking out the Opening of the Day on the homepage.

Your feedback is welcome, either right here or emailed to chess@chessgames.com.

Thanks,
The Chessgames Staff

♘ NITTY-GRITTY FOOTNOTES ♘

(a) Unlike other data that appears on OE pages, the tabulation of this data is very computationally expensive. For that reason, we can't just generate it "on the fly" as you cruise through the OE. At the time of this writing there are only a few thousand OE pages with the new information presented, but they happen to be mostly very popular pages so you'll notice them a great deal. At the heart of this system is a "bot" that slowly spiders our own site and adds or updates data. The data will NOT be instantly updated when new games are added, game collections change, etc., but we are hoping that it will update frequently enough to never be "stale". The logic that the bot uses to prioritize its spidering is an interesting issue in itself; one suggested improvement is to make it look at our own logfiles and spider whatever people happen to be examining.

(b) Actually, it's not the usual CG sense of the word: we raised the bar considerably for purposes of this feature. We require more than 10 collections to exist on a game before it gets triggered here, where most features require 2 or 3 collections. The reason for this decision is that we only want to alert the user to games which are truly famous and notable, not just happens to be in a few people's favorite game collections. We're not sure if 10 is really the magic number to use, but so far it seems to work fairly well.

(c) The algorithm which figures out which collections to show is a little weird, and that's on purpose. We don't want it to be so obvious that people can intentionally "rig" their collections to show up on a page of their choosing. At the same time, if there really is a collection all about that opening it ought to appear. You may notice some surprises, for example a page on the Sicilian may trigger a game collection "Bobby Fischer's Greatest Games", even though that collection contains a few Ruy Lopez games as well. That's OK, in fact it's a good thing.

(d) The "Debut" algorithm is not quite right, and probably needs some tweaking. The problem is that if the move was played many times in the same year (which is almost always the case with good novelties) it will often find itself unable to know with 100% certainty which one came first. For instance, if Topalov plays a novelty on "2009.09.15" it could be trumped by another game that reads "2009.??.??", even if the Topalov game was in fact the novelty. One solution is to simply label the field differently ("Early Example"?) to emphasize that it's not guaranteed to be literally the debut. Other ideas involve applying special sorting rules which penalize "??.??" style dates over specific information.

(e) After some consideration we decided that the "Recent" feature should be restricted to only very strong players, so we made an Elo cutoff of 2400. At first we even considered a 2500 cutoff, but decided that was being a bit too extreme, since it would even exclude some grandmasters. This way, if you look up (say) the King's Gambit you will find the most recent occurrence of that opening being played at the GM level, or darn close to it. This cutoff not only makes the information more interesting, but it also discourages low-rated members from uploading vanity games just to see their name on important OE pages.

Where a rating-related cutoff is concerned, I can see the logic -- 'vanity PGNs'. One problem might be the number of games in the database with incorrect or inadequate rating data. There are a lot of these, and fixing them is going to be a big task.

By and large, recent GM tournaments are not affected. But pre-2000 games seem to have been uploaded on a fairly random basis. In some cases, a player's rating at the time of submission (rather than the date of the game) was given - leading to absurdities such as players with 2400 ratings before the Elo system was in use.

In most cases, the emphasis on recent games will tend to cancel out the errors. But this is going to need repair work in the long run.

This is gratifying, but I don't think the game should be held up as any kind of example. It might be an idea to include a subroutine excluding simul games, even though one of the players is a high-rated GM.