deferred draw acceptance

The act of developing or disclosing that which is unknown.
Richard Bates
Posts: 2720
Joined: Fri Nov 14, 2008 8:27 pm

Re: deferred draw acceptance

Post by Richard Bates » Mon Dec 27, 2010 3:05 pm

PeterTurland wrote:I agree with Jack, three points for a win and one point for a draw makes a draw less likely and makes players take more of a risk, which is better for chess because it is better for the spectators.

The fact that the tourney had only one winner, is also evidence that it is probably the way forward, even McShane tho' it cost him money said in his newspaper column, that it was the way forward.
You may agree with Jack, but 3pts for a win is not Sofia rules...

Alex Holowczak
Posts: 8468
Joined: Sat May 30, 2009 5:18 pm
Location: Oldbury, Worcestershire
Contact:

Re: deferred draw acceptance

Post by Alex Holowczak » Mon Dec 27, 2010 5:22 pm

PeterTurland wrote:I agree with Jack, three points for a win and one point for a draw makes a draw less likely and makes players take more of a risk, which is better for chess because it is better for the spectators.
I'll continue to repeat this point: Statistically, that's not true at all! If you look at chess events where you have three points for a win, and chess events where you have the normal scoring system, it doesn't make any difference to the percentage of draws.

They found using similar statistics with the top division of the Football League that actually, the number of draws went up.

Out of 462 games (22 teams playing 42 matches each):
1980-81: 2 points for a win: 118 draws
1981-82: 3 points for a win: 121 draws
1994-95: 3 points for a win (last year before they reduced the number of teams): 134 draws

Note from now on, the numbers were calculated using my brain, so should be checked against the corresponding wikipedia article!

Out of 380 games (20 teams playing 38 matches each):
2009-10: 96 draws

So far this season, after 179 games:
2010-11: 57 draws, on pace for about 121 draws

Since 1980-81, the gap between the top teams and the bottom teams has dramatically increased, so although there is a slight drop in the number of draws as a %, it's clear that the number of points for a win makes no difference whatsoever. Indeed, we're on course for more draws this season than were had in 1980-81, despite the fact there are more points for a win now, and there's a bigger gap between top and bottom.

We don't know one vital piece of information: How many draws would we have had if it were two points for a win? For all we know, three points for a win could be improving the draw % (i.e. making fewer draws), but we don't have any information to compare it to (i.e. there is no information where the same teams play with a system of two points for a win). So we can't make any comment at all about whether three points for a win makes any difference.

In chess, we can, because 2/1/0 is equivalent to 1/0.5/0, and top players play that by default as opposed to any other system. I think you'll find a similar lack of correlation to the number of draws there too.

Peter Rhodes
Posts: 246
Joined: Thu Aug 20, 2009 10:53 pm

Re: deferred draw acceptance

Post by Peter Rhodes » Mon Dec 27, 2010 9:19 pm

Alex Holowczak wrote:I'll continue to repeat this point: Statistically, that's not true at all! If you look at chess events where you have three points for a win, and chess events where you have the normal scoring system, it doesn't make any difference to the percentage of draws.
I would like to see these statistics. On the other hand - in an old post I showed statistics from the CCF chess club that showed that games played with the 3-1-0 system had less draws than games played with the traditional system. http://www.ecforum.org.uk/viewtopic.php?f=2&t=1430

You made the claim
When I get some free time, I'm going to get some statistics that debunk the 3-1-0 myth.
which I still notice you haven't done, unless you're counting these football league stats which are from a completely different sport and I can think of numerous reasons why the conclusions might not apply to chess.
I'll do it over the weekend, writing it up properly in LaTeX and stuff.
never appeared either.


I think it's hard for any behaviourist to argue that the 3-1-0 system will not influence draws at all. I can see the validity of arguments along the line of "it won't make alot of difference" - but to argue that it does not make any difference just seems to go against common-sense.

And therein lies the dilemma, if you admit that the 3-1-0 scoring system might affect only 1-in-1000 games you then let the genie out of the bottle.
Chess Amateur.

Roger de Coverly
Posts: 17253
Joined: Tue Apr 15, 2008 2:51 pm

Re: deferred draw acceptance

Post by Roger de Coverly » Mon Dec 27, 2010 10:19 pm

Peter Rhodes wrote:I think it's hard for any behaviourist to argue that the 3-1-0 system will not influence draws at all.
I would be extremely reluctant to play in an event with a 3-1-0 scoring system. This would be particularly the case in a Swiss with a wide ability range. From my perspective at 2100, a couple of draws with FMs at 2300 should not be given a lower score than a loss in 30 moves to a 2500 GM and a win in 15 moves against a relative novice under 1700.

If you don't want short draws, the Sofia style rules at the London Classic of directing all draw acceptances through the arbiter makes considerable sense. The main problem is that you can only really enforce it when the event is paying the players rather than the case at most amateur events, the other way round.

Alex Holowczak
Posts: 8468
Joined: Sat May 30, 2009 5:18 pm
Location: Oldbury, Worcestershire
Contact:

Re: deferred draw acceptance

Post by Alex Holowczak » Mon Dec 27, 2010 10:24 pm

Peter Rhodes wrote:
Alex Holowczak wrote:I'll continue to repeat this point: Statistically, that's not true at all! If you look at chess events where you have three points for a win, and chess events where you have the normal scoring system, it doesn't make any difference to the percentage of draws.
I would like to see these statistics. On the other hand - in an old post I showed statistics from the CCF chess club that showed that games played with the 3-1-0 system had less draws than games played with the traditional system. http://www.ecforum.org.uk/viewtopic.php?f=2&t=1430

You made the claim
When I get some free time, I'm going to get some statistics that debunk the 3-1-0 myth.
which I still notice you haven't done, unless you're counting these football league stats which are from a completely different sport and I can think of numerous reasons why the conclusions might not apply to chess.
I'll do it over the weekend, writing it up properly in LaTeX and stuff.
never appeared either.
Yeah, I had more important things to do. :oops:

I want to see statistics from the top level of chess. I didn't have these to hand; football league information though; I did!
Peter Rhodes wrote: I think it's hard for any behaviourist to argue that the 3-1-0 system will not influence draws at all. I can see the validity of arguments along the line of "it won't make alot of difference" - but to argue that it does not make any difference just seems to go against common-sense.
I don't see why it'd go against common sense at all.

Rugby had a similar dilemma. They wanted people to play more attacking rugby (i.e. score tries), because it was deemed to be more interesting. So, they increased the number of points for a try from four to five. I haven't checked any statistics on whether more tries were scored, but given that change came when rugby became professional, and the game fundamentally changed with some of its rules (legalisation of lifting at lineouts amongst others), I don't think any statistical comparison of before and after could come to much of a conclusion; you're not comparing like for like.

The mentality though did change. Before, one converted try = two penalties, and now one converted try > two penalties. So now, defences are happier to give away penalties, because they know the try is the greater evil. If instead, it was two points for a try, then defences would make sure they'd never give away any penalties! It'd be far worse than giving up a try, which is perceived to be more exciting.

The only advantage that change had was that games in which a team were losing could come from further behind to win than before, due to being able to score a maximum of 7 points with one possession, rather than 6.
Peter Rhodes wrote:And therein lies the dilemma, if you admit that the 3-1-0 scoring system might affect only 1-in-1000 games you then let the genie out of the bottle.
It's very difficult to come to any conclusion with any degree of certainty, because you're never comparing like for like. Chess is a rare exception though, so if evidence were available, some analysis would be interesting.

It's nowhere near as simple as saying if you increase the number of points for a win, the number of drawn games will decrease. There are far too many variables going on for perceived wisdom such as "increasing the number of points for a win reduces the number of draws" to apply. That's the problem with perceived wisdom - it's only perceived. :wink:

User avatar
Mats Winther
Posts: 194
Joined: Thu Jun 04, 2009 5:27 pm
Location: Stockholm, Sweden
Contact:

Re: deferred draw acceptance

Post by Mats Winther » Tue Dec 28, 2010 7:59 am

Well, if we don't come up with something now, then we will have to bear with the said 3-1-0 scoring rule, as in the recent London tournament. A draw is such a natural outcome of chess so this scoring rule will never be generally accepted. As black, many defensive systems, unlike "Najdorf - poisoned pawn", etc., are designed to produce a draw. One cannot cut away such an immense portion of chess, namely defensive play. It can be beautiful and clever and it belongs in chess. To reward attacking play more than defensive play is very unfair, as attacking play is already somewhat advantageous, not the least for psychological reasons. The London scoring rule does not harmonize with the very nature of chess. A game when both players play perfect chess will end in a draw. A game when one or both players make mistakes ends in a win. So incorrect chess is rewarded with the London rule. This is unwholesome. But to remove early draws, some way or the other, is a better strategy.

Nor is the Sofia rule any good because defensive play will create games that go on for much longer time, and thus they take a greater toll.
/Mats
Last edited by Mats Winther on Tue Dec 28, 2010 11:41 am, edited 1 time in total.

Roger de Coverly
Posts: 17253
Joined: Tue Apr 15, 2008 2:51 pm

Re: deferred draw acceptance

Post by Roger de Coverly » Tue Dec 28, 2010 9:20 am

Mats Winther wrote:Nor is the Sofia rule any good because defensive play will create games that go on for much longer time, and thus they take a greater toll.
Why is this a problem? Games can and do continue for ages with thirty second increments, particular those involving K***h A****l.

LozCooper

Re: deferred draw acceptance

Post by LozCooper » Tue Dec 28, 2010 9:42 am

[quote="Mats Winther"
Nor is the Sofia rule any good because defensive play will create games that go on for much longer time, and thus they take a greater toll.
/Mats[/quote]

I thought the idea was to remove quick draws in which case Sofia rules serve that purpose. If you want to dictate the style of play then you have to alter which players you invite.

Alex Holowczak
Posts: 8468
Joined: Sat May 30, 2009 5:18 pm
Location: Oldbury, Worcestershire
Contact:

Re: deferred draw acceptance

Post by Alex Holowczak » Tue Dec 28, 2010 2:25 pm

By coincidence, I was watching this on YouTube: http://www.youtube.com/watch?v=WFyoQX59USY

David Howell says, on the issue of 3 points for a win, "I don't think three points makes any difference, and personally I would never use it in a chess tournament."

I assume that means he approaches the game in the same way as a game with the conventional scoring system. It would be interesting to hear the views of other players.

raycollett
Posts: 207
Joined: Wed Jun 25, 2008 1:54 pm

Re: deferred draw acceptance

Post by raycollett » Wed Dec 29, 2010 12:39 pm

Mats Winther wrote:Well, if we don't come up with something now, then we will have to bear with the said 3-1-0 scoring rule/Mats
I belive Basman's UK Chess Challenge uses 3-1-0 for the first level (school club) competition but 1-0.5-0 for later stages.

Alex Holowczak
Posts: 8468
Joined: Sat May 30, 2009 5:18 pm
Location: Oldbury, Worcestershire
Contact:

Re: deferred draw acceptance

Post by Alex Holowczak » Wed Dec 29, 2010 1:40 pm

raycollett wrote:
Mats Winther wrote:Well, if we don't come up with something now, then we will have to bear with the said 3-1-0 scoring rule/Mats
I belive Basman's UK Chess Challenge uses 3-1-0 for the first level (school club) competition but 1-0.5-0 for later stages.
The points system at the preliminary round is 3-2-1, which is of course identical to 1-0.5-0, but gives everyone some points. We just used 1-0.5-0 at my school, with the qualification at 4/7, which is just the same as 15/21.

The stages thereafter are 1-0.5-0 until the Terafinal, which as you say is 3-1-0. Amongst the five people from our school who have played at the Terafinal over the years, a total of five of them don't like it. I have no idea what the others think; five is not really a good sample out of the two-hundred or more who play in it.

benedgell
Posts: 1244
Joined: Thu Apr 05, 2007 12:43 pm
Location: Somerset
Contact:

Re: deferred draw acceptance

Post by benedgell » Wed Dec 29, 2010 5:06 pm

LozCooper wrote:[quote="Mats Winther"
Nor is the Sofia rule any good because defensive play will create games that go on for much longer time, and thus they take a greater toll.
/Mats
I thought the idea was to remove quick draws in which case Sofia rules serve that purpose. If you want to dictate the style of play then you have to alter which players you invite.[/quote]

My understanding of the Sofia Rule is that a draw can be agreed when the position is stalemate, there's a three-fold repetition, the fifty-move rule, insuficient material to mate, or if the arbiter declares it a drawn position.

If that's correct, then someone could just play the Ruy Lopez line (Nf3-g5-f3, Rf8-e8-f8) for a quick draw if they really wanted.

I think the Sofia rule has been fairly effective, but you can never completely eradicate short draws.

User avatar
Rob Thompson
Posts: 757
Joined: Wed Jul 15, 2009 12:03 pm
Location: Behind you

Re: deferred draw acceptance

Post by Rob Thompson » Thu Dec 30, 2010 10:45 pm

Alex Holowczak wrote:Amongst the five people from our school who have played at the Terafinal over the years, a total of five of them don't like it. I have no idea what the others think; five is not really a good sample out of the two-hundred or more who play in it.
You can make that at least 6 out of 6.
True glory lies in doing what deserves to be written; in writing what deserves to be read.

John Foley
Posts: 286
Joined: Thu Oct 15, 2009 9:58 am
Location: Kingston-upon-Thames
Contact:

Re: deferred draw acceptance

Post by John Foley » Sat Jun 11, 2011 9:47 pm

Spectators and organisers prefer to see games as battles rather than as exercises in risk-minimisation.
Hence the disappointment when players eschew the sporting encounter, particularly when there is plenty of play left in the game.
To address this, the introduction of deferred draw acceptance has merit in my opinion.
I am not convinced about the detail of Mats' original proposal, namely the requirement that the queen is unexchanged.
However, I think it is possible to reshape his proposal in a way that would be acceptable to most regular players.

The idea is to make a draw offer a risky intervention and hence reduce their incidence.

Suppose we call moves during the deferral period 'grace moves'. Once a draw offer is made, the offeree is allowed a number of grace moves whereby he can see the response before deciding whether to accept the offer. The simplest implementation would be for one grace move on the first draw offer.

A trivial example:
1. e4 =

Black (the offeree) now has a grace move and plays 1...e5
White continues 2. Nf3

Black has no more grace moves and must decide whether to accept the draw or not. Suppose he declines.

2...Nc6

3. d4=

With the second offer of a draw, Black is given two more grace moves.

3...ed:
4. c3 dc:
5. Nc3:

Now Black must decide whether or not to accept the draw offer.

The point about grace moves is that the offeror must reveal the next move(s) they would make before the offeree is obliged to respond.

A more pertinent example would be in a complex position where a player is faced with a difficult choice but makes a draw offer instead.
This modified approach has the associated benefit of penalising repeated draw offers without the arbiter having to be involved.
The deferral period could be extended by one move for each subsequent draw offer i.e. two grace moves are granted with the second offer of a draw.
Whether the appropriate number of grace moves upon a draw should start at a higher number is a matter for debate.

Deferred draw acceptance does not eliminate pre-arranged draws from collaborating players but it could deal with premature draws at the less exalted level.

Giulio Simeone
Posts: 154
Joined: Tue Feb 07, 2012 1:06 pm
Location: Rome, Italy

Re: deferred draw acceptance

Post by Giulio Simeone » Wed Feb 08, 2012 3:59 pm

This rule seems to be nonsense ... no player that has thought a bit about its consequences would offer a draw and, mostly, no one would accept it immediately!! Why should I accept a draw offer immediately if I know that I will be able to take the half point in every moment? That rule gives to the player who has been offered the draw a great power: he can play very aggressively, he can sacrifice pawns and pieces ... and he can also lose the queen, provided he doesn't lose it in one move!! Suppose that a player has been offered a draw, and fifteen moves later he undertakes a knight fork to king and queen: before moving the king, he can still accept the draw! I think that no player would dare to give the opponent such a great power with this "permanent" draw offer.

Locked