ChessPub Forum - Argument over chess statistics

(UTC)

ChessPub Forum › General › General Chess › Argument over chess statistics

(Moderators: trw, proustiskeen, RoleyPoley)

‹ Previous Topic | Next Topic ›

Page Index Toggle

Pages: 1 [2]

Topic Tools

Very Hot Topic (More than 25 Replies)

Argument over chess statistics (Read 10753 times)

Markovich God Member Offline Posts: 6099 Location: Columbus, Ohio Joined: 09/17/04	Re: Argument over chess statistics Reply #11 - 03/22/11 at 01:16:27	Post Tools	Print Post
	Uhohspaghettio wrote on 03/20/11 at 22:46:04: Markovich wrote on 03/20/11 at 21:37:33: But I have to wonder about your Spanish numbers as well. When I checked my own data base, I found 591 games where both opponents were in the 1900-2100 range and 6...Be7 was played, which is where most people would say the Closed Spanish begins. Black only drew 33% of them, not the 39% that would've been consistent with your numbers. I suggest, indeed, that you've fouled up the numbers and reversed the White winning and Black drawing percentages in the Spanish row. Hmmm... it seems C84-C89 is also the closed Spanish (I've never seriously played/studied the Spanish myself, at least not yet). Regardless, I clearly showed I was filtering for C90-C99 in my post. In Megadatabase 2010, from the years 2000-2010, 1900-2100. Are you sure you're doing it right Richard? I know computers can be tricky sometimes... My name is well known here. Yours is not. I suspect you're a notably unstable visitor here, named Richard, who visits here from time to time under various handles and has conceived some sort of animus toward me. Your replies here and before rather strongly suggest it. However if you're not him, I am perfectly willing to accept the possibility that you're some other similarly unstable, provocative person. I searched on those ECO codes and found that in games in my database (Opening Master 2, about 5,000,000 games total) with both opponents rated 1900-2100, Black draws 33% just like I said , not 39%. You have your numbers backward, it's obvious. It would be rather strange if White only won 33% while Black drew 39%. It's the other way around. Check it. But my larger point would be that nobody sub 2250 or so can play any system with any degree of reliance on its supposed drawish properties. Even if your numbers were correct, and even if they weren't subject to selection bias -- which is a very important determinant of chess opening statistics, they have essentially no practical importance. Just look how s--tty, for instance, the supposedly solid Petroff performs in your stats. If you think otherwise, fine, play the Petroff to your heart's content and give a big hug your statistics.

	The Great Oz has spoken!
	IP Logged

TN God Member Offline Posts: 3420 Joined: 11/07/08 Gender:	Re: Argument over chess statistics Reply #10 - 03/22/11 at 00:38:51	Post Tools	Print Post
	The best way to achieve a draw is to consistently play good moves. One should not be afraid of 'drawish' openings or variations, but it's helpful to be aware of them, in case you wish to avoid them. Over the years I've seen a lot of people confuse symmetrical positions with drawish positions. They are not the same.

	All our dreams come true if we have the courage to pursue them.
	IP Logged

Uhohspaghettio God Member Offline I Love ChessPublishing! Posts: 515 Joined: 02/23/11	Re: Argument over chess statistics Reply #9 - 03/22/11 at 00:18:11	Post Tools	Print Post
	Fromper wrote on 03/21/11 at 03:43:00: And yes, I'm sure certain openings lead to more draws than others, even at lower levels, but is 5% more draws really enough to call something a drawish opening? Check out the stats on the Petroff vs the Sicilian among games where the ratings are 2500+ and see how much more often the Petroff leads to draws at that level. I don't know where to check these types of stats, but I'd guess that the Petroff leads to draws at least 20% more often at that level. It may be a safe drawing line for GM's, but a 35% draw rate at 1900-2100 level doesn't exactly qualify as "drawish" in my book. Well it depends on your perception of "drawish" I suppose. I think a more cautious claim would be something like "of course below expert/master level, no opening is ever really that drawish, you're talking a few percent more draws", with one player generally able to try something speculative for a win without it being suicide. But nonetheless it's true that some openings are still MORE drawish than others. I do agree though that playing for a dead draw right from the opening while the other side is trying to win probably isn't a very good idea if you're significantly below master level. Sort of like a football (soccer) team playing for a 0-0 from the start. If you think you will settle for a draw later on in the game, great. But foregoing all your own opportunities for an initiative probably isn't a good idea at below master level because one slip and...


	IP Logged

MNb God Member Offline Rudolf Spielmann forever Posts: 10777 Location: Moengo Joined: 01/05/04 Gender:	Re: Argument over chess statistics Reply #8 - 03/22/11 at 00:10:23	Post Tools	Print Post
	I dare to say that even on our level (1700-1800) 1.e4 e5 2.Nf3 Nc6 3.Nc3 Nf6 4.Bb5 Nd4 5.Nxd4 exd4 6.e5 is quite drawish.

	The book had the effect good books usually have: it made the stupids more stupid, the intelligent more intelligent and the other thousands of readers remained unchanged. GC Lichtenberg
	IP Logged

Fromper Senior Member Offline GrandPatzer Posts: 378 Location: Chicago, IL, USA Joined: 03/12/10 Gender:	Re: Argument over chess statistics Reply #7 - 03/21/11 at 03:43:00	Post Tools	Print Post
	Ok, so my choice of the word "amateur" was a probably a poor choice. I meant "class level" players, ie those below master level. I stand by my statement that there's no such thing as a drawish opening at my level (1700), yet I've actually heard players rated lower than myself refer to the Exchange French or other lines as "drawish". And yes, I'm sure certain openings lead to more draws than others, even at lower levels, but is 5% more draws really enough to call something a drawish opening? Check out the stats on the Petroff vs the Sicilian among games where the ratings are 2500+ and see how much more often the Petroff leads to draws at that level. I don't know where to check these types of stats, but I'd guess that the Petroff leads to draws at least 20% more often at that level. It may be a safe drawing line for GM's, but a 35% draw rate at 1900-2100 level doesn't exactly qualify as "drawish" in my book.

	GrandPatzer!!! 1777 peak USCF rating - currently 1620 from coming back rusty
	WWW IP Logged

Uhohspaghettio God Member Offline I Love ChessPublishing! Posts: 515 Joined: 02/23/11	Re: Argument over chess statistics Reply #6 - 03/20/11 at 22:46:04	Post Tools	Print Post
	Markovich wrote on 03/20/11 at 21:59:40: I added more to my post, so check it out. Okay. Markovich wrote on 03/20/11 at 21:37:33: But I have to wonder about your Spanish numbers as well. When I checked my own data base, I found 591 games where both opponents were in the 1900-2100 range and 6...Be7 was played, which is where most people would say the Closed Spanish begins. Black only drew 33% of them, not the 39% that would've been consistent with your numbers. I suggest, indeed, that you've fouled up the numbers and reversed the White winning and Black drawing percentages in the Spanish row. Hmmm... it seems C84-C89 is also the closed Spanish (I've never seriously played/studied the Spanish myself, at least not yet). Regardless, I clearly showed I was filtering for C90-C99 in my post. In Megadatabase 2010, from the years 2000-2010, 1900-2100. Are you sure you're doing it right Richard? I know computers can be tricky sometimes... If it makes you feel any better, I also tried the French for getting a high draw percentage and it wasn't very high. However this may be because of all the other variations in the French Defence which mightn't be drawish at all.... and that is the only other result that I looked at. Even if it turned out that the exchange variation isn't drawish at this level, that doesn't mean there are no drawish lines at this level. The pawn structure the opening leaves is obviously particularly important. C84-C99 has 1-0: 34.9%, ½-½: 36.2% and 0-1:28.9% Both are between 1900 and 2100. Rechecking, yes, C90-C99 has 38.9% draws. Interestingly C84-C89 only has 33.0% draws. This is probably because the diversions from the line that leads to the C90s must be quite sharp diversions that change the game. Here are the alternatives to the main line (C90-C99) from Wikipedia: * 6.Bxc6 (Delayed Exchange Variation Deferred) * 6.d4 (Centre Attack) * 6.Qe2 (Worrall Attack) * 6.Re1 d6 (Averbakh Variation) * 6.Re1 b5 7.Bb3 o 7... Bb7 (Trajković Variation) o 7... 0-0 8.c3 d5 (Marshall Attack) Bingo. Now I know next to nothing about the Ruy Lopez, but even I've heard of the notoriously aggressive Marshall Attack....


	IP Logged

Markovich God Member Offline Posts: 6099 Location: Columbus, Ohio Joined: 09/17/04	Re: Argument over chess statistics Reply #5 - 03/20/11 at 21:59:40	Post Tools	Print Post
	I added more to my post, so check it out.

	The Great Oz has spoken!
	IP Logged

Uhohspaghettio God Member Offline I Love ChessPublishing! Posts: 515 Joined: 02/23/11	Re: Argument over chess statistics Reply #4 - 03/20/11 at 21:50:43	Post Tools	Print Post
	Markovich wrote on 03/20/11 at 21:37:33: Uhohspaghettio wrote on 03/20/11 at 21:30:28: Markovich wrote on 03/20/11 at 21:14:01: @uhoh: Sheesh, who is being smug? You could have made your point with a whole lot less superiority and even open hostility. How old are you, about 16? Or are you our unstable chessfriend who keeps coming back under various aliases, R.DeC. from New Paltz? In any case before you make arguments of this kind, I suggest you read up on selection bias. Selection bias, are you serious? I took the people rated from 1900-2100 FIDE, which are in theory (are....) the same everywhere. All of their games were random, what you would expect them to score in any other game. In fact this is one of the unusual instances where we can say without doubt that there are no selection bias or confounding factors whatsoever. Do you think they took the most "interesting" games and decided to put them in? That's not how comprehensive databases work... Yes, and when players sit down to play chess, they discover the opening they're going to play by reading a lot that they pull out of a jar, right? Therefore, their purposes in any given game have nothing to do with the opening they select, right? Okay, that's actually a logical argument. Still though, I doubt it really accounts for all of those statistics. How often are amateurs looking for draws in these? The numbers wouldn't stack up. I bet even if you took statistics from games played at fics on blitz, where people definitely aren't trying for a draw, you would find significantly more drawing openings (although it might not be by too much, since the standard of play is a lot poorer at blitz).


	IP Logged

Markovich God Member Offline Posts: 6099 Location: Columbus, Ohio Joined: 09/17/04	Re: Argument over chess statistics Reply #3 - 03/20/11 at 21:37:33	Post Tools	Print Post
	Uhohspaghettio wrote on 03/20/11 at 21:30:28: Markovich wrote on 03/20/11 at 21:14:01: @uhoh: Sheesh, who is being smug? You could have made your point with a whole lot less superiority and even open hostility. How old are you, about 16? Or are you our unstable chessfriend who keeps coming back under various aliases, R.DeC. from New Paltz? In any case before you make arguments of this kind, I suggest you read up on selection bias. Selection bias, are you serious? I took the people rated from 1900-2100 FIDE, which are in theory (are....) the same everywhere. All of their games were random, what you would expect them to score in any other game. In fact this is one of the unusual instances where we can say without doubt that there are no selection bias or confounding factors whatsoever. Do you think they took the most "interesting" games and decided to put them in? That's not how comprehensive databases work... Yes, and when players sit down to play chess, they discover the opening they're going to play by reading a lot that they pull out of a jar, right? Therefore, their purposes in any given game have nothing to do with the opening they select, right? I would also point out that in your statistics, these 1900-2100 rated Whites should thank their lucky stars whenever their opponent plays the Petroff, since relative to the Sicilian, almost every additional Black draw is purchased at the cost of a Black win. Apparently the Petroff is an excellent system for sub-middling players who would much rather draw than win. But I have to wonder about your Spanish numbers as well. When I checked my own data base, I found 591 games where both opponents were in the 1900-2100 range and 6...Be7 was played, which is where most people would say the Closed Spanish begins. Black only drew 33% of them, not the 39% that would've been consistent with your numbers. I suggest, indeed, that you've fouled up the numbers and reversed the White winning and Black drawing percentages in the Spanish row.

	The Great Oz has spoken!
	IP Logged

Uhohspaghettio God Member Offline I Love ChessPublishing! Posts: 515 Joined: 02/23/11	Re: Argument over chess statistics Reply #2 - 03/20/11 at 21:30:28	Post Tools	Print Post
	Markovich wrote on 03/20/11 at 21:14:01: @uhoh: Sheesh, who is being smug? You could have made your point with a whole lot less superiority and even open hostility. How old are you, about 16? Or are you our unstable chessfriend who keeps coming back under various aliases, R.DeC. from New Paltz? In any case before you make arguments of this kind, I suggest you read up on selection bias. Selection bias, are you serious? I took the people rated from 1900-2100 FIDE, which are in theory (are....) the same everywhere. All of their games were random, what you would expect them to score in any other game. In fact this is one of the unusual instances where we can say without doubt that there are no selection bias or confounding factors whatsoever. Do you think they took the most "interesting" games and decided to put them in? That's not how comprehensive databases work...


	IP Logged

Markovich God Member Offline Posts: 6099 Location: Columbus, Ohio Joined: 09/17/04	Re: Argument over chess statistics Reply #1 - 03/20/11 at 21:14:01	Post Tools	Print Post
	@uhoh: Sheesh, who is being smug? You could have made your point with a whole lot less superiority and even open hostility. How old are you, about 16? Or are you our unstable chessfriend who keeps coming back under various aliases, R.DeC. from New Paltz? In any case before you make arguments of this kind, I suggest you read up on selection bias.

	The Great Oz has spoken!
	IP Logged

Uhohspaghettio God Member Offline I Love ChessPublishing! Posts: 515 Joined: 02/23/11	Argument over chess statistics 03/20/11 at 20:14:19	Post Tools	Print Post
	Fromper wrote on 03/20/11 at 16:49:16: Markovich wrote on 03/20/11 at 02:24:07: Nor do I think the Petroff is particularly conducive to Black's taking the half point when played at these lower levels. Yeah, it always amuses me to see amateur players refer to openings like the Petroff as "drawish". The Exchange French is another one that people always say that about. I always tell them that below master level, there's no such thing as a drawish opening. As a 1700ish player who has played the French quite a bit as black, I can't recall ever scoring a draw in the Exchange variation. And as a 1. e4 player, I only remember getting a draw once against the Petroff. And that had nothing to do with the opening, and everything to do with a hard fought game against an evenly matched opponent, which came down to a typically drawish rook and pawn endgame. As for playing the Ruy as black, I don't think players at my level should play the main lines. I played the Chigorin for a bit, and it was just too closed and positional. I've since switched to the Open Ruy, just to get more open positions in my games. I filtered all the games in Mega Database between players 1900-2100 for the years 2000-2010, and they show some openings do draw significantly more than others at even this level (which is far lower than master). The Sicilian B20-B99 got 1-0: 37%, ½-½: 30.7% 0-1: 32.3% (out of 21704 games). The Petrof C42-C43 got 1-0: 38.0, ½-½: 35.4% 0-1: 26.4% (out of 920 games). The Closed Ruy Lopez C90-C99 got: 1-0 33.3%, ½-½: 38.9% 0-1: 27.8% (out of 565 games). And you have what to back up your claim, a ridiculous anecdote? IMO you need to be extra careful you're right yourself in a situation like that, where you find the supposed ignorance of other amusing and smugly go about feeling you're "in the know" about something. What struck me the most about your post is that you claimed you would have to be a professional to be able to call something a drawish opening. There aren't that many professional chess players, they are a very small group of people in the world. Not every master is a professional. I don't know how anyone can say some lines in chess aren't drawish, unless they are at the level of just not seeing basic tactics. According to these statistics, you can have drawish lines below 2000, nevermind below 2300, nevermind professional playing level which would usually need to be close to 2500 in this day and age (unless you're a woman maybe). "It always amuses me", well looks like the "amusement" is on you now.


	IP Logged

Page Index Toggle

Pages: 1 [2]

Topic Tools

‹ Previous Topic | Next Topic ›

Bookmarks:

del.icio.us

Digg

Google

Google+

reddit

StumbleUpon

Yahoo

« Board Index ‹ Board Top

Top

ChessPub Forum » Powered by YaBB 2.6.12!
YaBB Forum Software © 2000-2026. All Rights Reserved.

Valid XHTML

Valid CSS

Powered by Perl

Source Forge