chess club

Chess Club

Classify and you will find!
Quirky name, real interests
[ Sign up | Log in | Guest ] (beta)
wschmidt 31 ( +1 | -1 )
Novice Nook #12 Wow, hard to believe we've been at this 12 weeks already! This week's article is entitled "Analysis and Evaluation" and can be found at:

-> www.chesscafe.com

Sorry I didn't contribute last week. Too many irons in the fire. ws
ionadowman 41 ( +1 | -1 )
Interesting... To be honest, I didn't think the article that helpful to begin with, but the practical examples makes it clearer what he is driving at. The examples are not at all difficult, but their simplicity, with Heisman's commentary, leads one through the process he is describing. To be sure, the process becomes intuitive with experience, as Heisman observes, but it is kinda nice to know what's supposed to be going on in one's mind!
Cheers,
Ion
ccmcacollister 37 ( +1 | -1 )
A mistake ... I think he makes a mistake when he says you must analyze to "quiescence"; and rather it should be that you must analyze beyond it. A "quiet" move at the end of a combination, when apparently all is still, can kill you !
Actually too, in corr Chess, that is exactly the point where you should Start analyzing all over again, if you want to be really very good.
ccmcacollister 10 ( +1 | -1 )
PS// An example of my last contention: I used to always do that - now I don't !
2200+ then, 1700ish now .... hmmm }8-))
wschmidt 72 ( +1 | -1 )
For someone who's played chess... for a long time, the idea of analyzing to quiescence (however well or poorly one does it) is pretty much understood. But for the beginner, I think this is a pretty good explanation of a key part of chess thinking. Once you understand the principle, it seems obvious on it's face, but until you've figured it out for yourself or until someone explains it to you, that "what do I do now?" sense makes chess thinking a pretty overwhelming fog.

This is actually the first time I've read the term "quiescence" used as applied to chess evaluation. Anybody familiar with other chess writers who use the term?

Craig, I understand what you mean, but if it's a "quiet" move that kills you, then I don't think you've really reached quiescence! ws
ionadowman 89 ( +1 | -1 )
Right ws! Check out some of Alekhine's games: the sting in the tail of the combination. I think a lot of the kinds of advice Heisman (and others) dispense is very useful until you get to fairly advanced levels - though how advanced I'm not sure. Then you have to reassess the advice in the light of your own experience.
But I don't think Craig's remark is really all that contradictory of Heisman's advice. You have to make some kind of judgement whether the 'quiet' position is really quiet. Recall - was it Marshall's? - remark that you should make your combinations so that "when the fire is out, it isn't out!" So failure to detect that flickering ember constitutes one of the mistakes that Heisman mentions: Number 4 (or maybe 3 or 5).
"Quiescent" is quite a good choice of word (c.f. "quiet"), suggesting dormancy, a kind of momentary stasis that could at any time re-erupt into violent action...
Cheers,
Ion
mattdw 105 ( +1 | -1 )
Interesting! Interesting read! The concept of 'quiescense' is something that I had thought about, though not with that name - just a fuzzy bunch of concepts in my mind. I hadn't really thought about specifying where to truncate analysis, but I think due to the systematic nature the basic analysis process - "If I play this, the best reply would (probably) be, to which I would reply...etc.." I would simply continue onwards until I got to a point where I cannot see a 'best move' ...or until I got distracted or hungry, whichever comes first. ;) The above type of process is just a simple reflection of the concept of quiescense, and due to the logical nature of it I believe that many people would be applying it subconciously. It's probable that the increased reliabilty in move selection that could arise from realising that the point at which we are ceasing to analyse is not arbitrary, but it is in fact a particular situation upon the board to which we are aiming to locate, would greatly improve the efficiency of (and confidence in) the thought process.
cascadejames 80 ( +1 | -1 )
quiescense I like the term and I like his definition, ". . . all the most forcing moves, such as checks captures,
and strong threats, have been resolved." It's probably a better way to think about it than what I
have been trying to do --which is to try to evaluate a position until the number of possible moves
makes it impractical to continue.

Incidentally, in the second game, isn't the 8th move 8. dxc5 dxc5,
instead of 8. dxc6 dxc6 as shown in the text? I admit, I still have trouble with algebraic notation
because I learned the old style when I was a kid, so to me the 8th move looks like, 8. pxB5 pxB4.
I am sure that the old style notation looks bizzare to the young guys, because the idea of having
two different names for the same square probably sounds totally screwy.
mattdw 418 ( +1 | -1 )
Minimax.. I have a thought regarding the term 'Minimax', I had a question about this a few weeks ago when I was using Fritz to analyse one of my games but never got round to asking it. Basically, is it always best to play with the minimax idea in mind? I'll write an example to illustrate my question: Assuming for the sake of argument that we are able to judge the situation to a hundredth of a pawn...Imagine we have three possible moves to which our opponent has five possible replies: (in order from their best to their worst move).

Now say that if we played move #1:
The evaluation after our opponent selects their move could be:
-0.02, 0.00, +0.02, +0.04, +0.09.

Say that if we select move #2:
The evaluation after our opponent selects their move could:
-0.50, +1.00, +1.12, +1.23, +1.28.

And finally, say that if we played move #3:
That the evaluation after our opponent selects their move could be:
-1.00, +2.46, +2.76, +2.88, +3.10.

...by the applying the minimax rule we would play move #1 since our opponents best reply would result in the evaluation of -0.02 with move #2 (-0.50) and #3 (-1.00) being second and third respectively. But my thought is that since in every game, no matter what the ability of the players, there will always be some sub optimal moves. This implies that there is some kind of inherent probability that we will select a particular move based on the particular situation, our style, knowledge and experience and consequently there should exist some probability that we will select the move which produces the best (arbitrary) evaluation for us.

If anyone is familiar with the concept of expectation in probability then may see where I'm going with this. Imagine it was possible to attach a probability between 0 and 1 to each move where given the same situation an infinite number of times that the move would either be never played or always played respectively, and everything inbetween. Now if we applied this concept to the move list for our opponent it may make then we may feel that a different move is better.
Say that each move has the following probability (in the same order as above) then the expected evaluation resulting from our move would be:

Move #2
Probabilites (with corresponding evaluation in brackets):
0.95 (-0.02), 0.02 (+0.00), 0.01 (+0.02), 0.01 (+0.04), 0.01 (+0.09)
Expected evaluation:
-0.0175

Move #1
Probilities (with corresponding evaluation in brackets):
0.20 (-0.50), 0.20 (+1.00), 0.20 (+1.12), 0.20 (+1.23), 0.20 (+1.28).
Expectation:
+0.826

#3
Probabilites (with corresponding evaluation in brackets):
0.05(-1.00), 0.20(+2.46), 0.20(+2.76), 0.25(+2.88), 0.30(+3.10)
Expected evaluation:
+2.644

So now the best move by expected evaluation is #3 with +2.644, this is mathematically the evaluation you would expect to have if this exact position & conditions were played an infinite number of times. #2 is still second with +0.826 but #1 is now last with -0.0175. What this basically translates to is that move #1 is safe but you can expect to end up marginally behind, #3 is a bit more risky, but in the vast majority of cases you would expect to come out quite a bit ahead and #2 is somewhere inbetween. I doubt that this has ever been programmed into a computer as it would be impossible to determine the exact probabilities of the moves, and I believe that we as humans would do this naturally through intuition and experience rather than concrete numbers. But the point (in plain english!) I'm making is that, 'is it sensible to try to read the board AND opponent in light of our experience and knowledge to increase our overall win percentage by playing risky moves where we believe that the majority of cases we will come out on top?'.

After writing all that I realise this is very similar to something that I have discussed with wschmidt, though not in terms of probability. I think we agreed that if you want your games to be exiciting then go for that dubious sacrifice if you believe your opponent will not be able to handle it well or if you want your games to be quiet in the hope you can grind out a win - then dont.




An regarding the notation:
I think I'm lucky when it comes to those two types of notation, I learnt them both simultaneously - which I think has resulted in neither being more awkward to use than the other, though I do think algebraic is a bit more sensible.

ionadowman 74 ( +1 | -1 )
Pillsbury... ...was probably the author of my earlier quotation, now that I think about it. Reading through mattws's 'minimax' discussion, I'm reminded of Tartakover's remark: "the second-best move is sometimes the best move." I think we shouldn't lose sight of the psychological aspect of the game, wherein evaluation and judgement are not altogether objective, but subject to individual character, personality and experience... Any position in which the difference between the two sides is +-0.50 seems to me acceptable. The latter two examples mattdw gives indicates I would be likely to reject #2 and #3 fairly quickly, once I've noticed that a possible enemy's response would put me more than a pawn down (quantitatively or qualitatively). (I'm assuming an assessment of +1.00 is the equivalent of 1 pawn ahead)...
Cheers,
Ion
ccmcacollister 95 ( +1 | -1 )
Think about this ... especially for corr play where no limit of analytical abilities is imposed upon you. You analyze to quiesence, and say to yourself "that looks pretty even" "I'll just stop analyzing". Is your intuition really that good??? To just say lets stop here and hope that my glance is correct? ?
Or do you go on to analyze there looking for the subltle positional pressure with pull towards winning, even if there is not a quite kill shot on the scene?
Granted I am speaking mainly to perhaps A/B players who want to be expert or master. And X/M should already know it! Tho that has not been my eperience
. But the really good ones do. Always. (Bearing in mind I've played people who have been in the world corr championship. I might know what I'm saying here :)
At the same time, I dont intend to be critical of the article, it IS a very important step for those who dont know, to analyse At Least to there.
}8-)
ionadowman 135 ( +1 | -1 )
Agreed... ...Heisman's article is a bit like Terry Pratchett's 'lies to children': not quite the truth, but a very good and useful ... lie. Well, not altogether a lie, neither. More of a working assumption.
One of the difficulties one will no doubt encounter later on is something I think one of Bill Hartston's books mentioned. Occasionally we find positions in which one side looks to have a decided edge, and it is only with exploration of the possibilities available to both that we are led to a different view. Consider a situation in which one side is well placed, but pretty much has its men on their optimal squares, whilst the enemy's position could stand considerable improvement. Now suppose the former has no good way of improving its good position, but the latter can improve quite a bit without any real interference. I've seen situations like that - I suspect I might have one in one of my current games (unfortunately). It transpires that the latter has the better of things if its potential proves the greater. Clearly in evaluating a quiescent position, one has to be somewhat aware of its dynamics, namely its potential and capacity for improvement. I suspect this kind of judgement is a very high order, and one not easy to practise, at that!
There seems to be 'levels' of development, each 'level' somewhat modifying or qualifying what has been learned at earlier 'levels'. (Sounds horribly like RPGs, don't it?).
Cheers,
Ion
ccmcacollister 345 ( +1 | -1 )
clarification Someone pointed out to me that my prior posts might be seen as critical or defensive, so I just wanted to assure readers that such is Not the case. And to wschmidt , my apologies to you if it seemed so to you as well. I agree they could have been read that way, and so just want to clarify that the certain vehemence apparent in the post is soley because of the weight of importance I place upon the suggestion of analyzing a bit further.
***
And wschmidt makes a good point about When IS Quiesence Reached, that we should not consider it so if there are any shots left on the board. And that sounds very reasonable indeed; but it does seem that just to determine if it Has been reached, we must look somewhat beyond the point it first Seems reached. Just to be sure. And again, any work you do on the position from that point on is only to your benefit.
The article itself points out that one of the primary errors to watch for is thinking mistakenly that Quiesence was reached. And so to the writers credit to make that caution. I simply cannot tell you how many corr games I've won or staved off a loss in because the opponent looked one move too short. And remember all to well the loss taken when I thought I had seen one move further, but he saw a move beyond that. A very important concept. It is something that WINS. Like zwishenzugs, double-checks and eliminating wishful thinking do. Look further than your opponent, to win.
***
[As I've mentioned before; to win you must KNOW more aor EXECUTE BETTER what you do know. This is an easy way to KNOW MORE than your opponent about that particular game, without cracking a book! ~And marks superior execution as well.]
***
So from that there is a principle that even applies to the novice player ... to Look Beyond The Recapture. At least a move (or two), and you will avoid a lot of grief. And look for chances to win from opponents who fail to do that. (And it would gratify me a lot to hear from anyone who wins or saves a game from that circumstance.)
***
Just to be clear, I have liked this series of articles a lot. It can be very difficult to simply write clearly and comprehensively about so vast a subject as Chess,without leaving out a whole lot that is important. And even moreso to write towards an audience of mixed levels and experience. I think the author of the series does a good job of it overall, and wschmidt has done us all a service in presenting them.
And since it IS "Novice Nook" I feel regretful in calling it a "mistake" for the author not to have suggested going further. As wschmidt pointed out, it is a very important basic concept to have introduced to novices. Thus I think instead I should have offered the additional as an "addendum" for stronger players, rather than the wording used. But an addendum I must feel compelled to add, for the simple reason; the subject had not reached Quiesence ! :))
Regards to all, CAC
PS// I'd like to offer one final observation, that being ... Working hard at your game ALWAYS pays off. If naught else, by the feeling of confidence from knowing you've paid your dues, and so should feel you have every right and reason to defeat anyone you play ... !
GOOD CHESS, to all. Keep pushin on ! }8-)

wschmidt 229 ( +1 | -1 )
No apology necessary, Craig.... ...the issue of when quiescence is reached and how do you analyze your position at that point seems a fair one to me. I think that Heisman is looking at quiescence from a novice's point of view - the end of a series of strictly forced moves. He would then say look at the resulting position. You, as a more advanced player, would probably say "look at that 'quiet' position and recognize that there are some additional moves that, if not "forced", are so clearly positionally strong for one side that the quiescence really hasn't been reached. Thus, Ion's observation that at a higher level of play one's judgement will change.

Matt, your thoughts about move selection are an interesting mathematical approach to the long-standing chess question of whether one should play the opponent or the board. Different players at all levels have their own philosophies about this.

For the rest of you on the thread, this discussion between Matt and myself came up in the context of a game I played a few months back that I suggested Matt take a look at. My opponent was playing somewhat passively and as a result was in a bit of cramped position. At some point, rather than proceed with a slow positional game, I sacrificed a bishop for two pawns and an attack against a denuded king position. My thought was that given the nature of his previous play in the game (we were rated about equal) that he probably wasn't going to be a strong defender and, at the worst, I would be able to force a return of material and I'd still be positionally better.

That thought process proved correct and I won the game. I was surprised, however, when I had Fritz analyze the game afterward, at how much the machine disapproved of my sac. I was "behind" much further and much longer than I thought during the course of the game. And that, I guess, is the twist that isn't present in your question, Matt. I didn't know that at the time I made the sac. I thought for sure that the numbers would show me close to equal -down material but with a good attack. If I had known what Fritz thought before hand, I'm not sure I would have made the decision I did. (And probably wouldn't have enjoyed the game nearly as much! ) ws
mattdw 90 ( +1 | -1 )
I've experienced many times in my education (and I'm sure you all have too) the situation where I am tought something that is not scrictly correct due to many assumptions being made to simplify the matter. This makes it much easier to concentrate on that matter at hand without clouding it with information that is unnecessary at that time. Upon finding out that what I had learnt previously was in fact a simplification or in some cases slightly misleading (wave/particle duality!) was not that imortant as the new knowledge would put the previous teachings in a new light, thus building upon the foundations that the already been laid. I think this is the case with this arcticle, especially as it is named 'Novice Nook'! He probably feels that complicating the matter for the sake of accuracy would sacrifice too much of the main point that he is trying to get accross.