Since a narrative is a verbal exercise of system II, and system I doesn't understand a word of it.
When you solve an exercise, you make use of a narrative to keep you on track. That is already an advantage in comparison to trial and error at random. But I'm pretty sure there is something we should do after system II has build the initial narrative.
How does system I learn to handle the brakes when entering a curve? It has been said that it works by means of feedback. A gentle feeling of success when things go right and a feeling of awkwardness when things go wrong seems to impress the right motion into system I. Nobody gets tempted to learn how to handle a curve with system II. Albeit it all starts with a verbal instruction the first few times.
But how does that translate to chess?
 |
| Black to move |
[solution]
After construction, a narrative may sound something like this:
- I would like to put a rook on e2 since that is a double attack on K and Q
- e2 is covered by the bishop
- can I remove the bishop?
- yes, by capturing it with my rook
- bummer, e2 has now one attacker less
- oh wait, when white takes back on f1, the rook clears the diagonal to e2
- can white postpone the recapture of f1 by harassing my Queen?
- I can take his two rooks in return
- Let's go!
I'm not sure. But the next reasonable step seems to be to translate the narrative into the standard scenarios from the tree of scenarios.
- double attack on K and Q
- remove the defender
- add attacker by clearing the diagonal
- counter attack (harass black queen)
- tit for tat take the rooks
I have my baseline set long ago. When I improve my ATH by 50 to 1850, I consider this method possible. When I get to 1900, I call it plausible. When I get to 1950, I consider it a success. When I reach 2050: mission accomplished. Let's have a look at positional chess now!
I train in blitz mode. Since that reflects at which trivial things I fail. But there is of course a difference between training and measuring progress. Since training is way more time consuming.
This is an example of an unoccupied B.A.D. square. Your first thought is the fork of the White King and Queen. You can "see" that there are two attackers (Black Rooks on e1 and e4) and two defenders (Qd2 and Bf1). However, this is where "looking along the LoA to the edge of the board for each piece" helps me to "see" potentialities. The Black Queen is also "attacking" e2 (albeit indirectly at the moment), so there is a potential superiority ratio (2:3) already on e2. Squares (by definition) are always immobile. So, can I add an attacker to e2? Nope, because the Black Bishop is on the wrong color, and the Black Knight is too far away. That only leaves removing a defender. Coincidentally (there are NEVER coincidences in chess!), there is an attacker that can remove a defender: 1. ... Rxf1, but that also removes one of the attackers. KEEP IN MIND THAT THE ATTACKING/DEFENDING RATIO WAS ALREADY FAVORABLE! 2. Rxf1 is forced (to maintain the material balance). AHA!! Now the ratio on e2 is (1:2) because the line of the Black Queen is now open. 2. ... Re2+ 3. Qxe2 Qxe2 4. Kg1 Qxc4 and Black is winning.
ReplyDeleteI started training to establish a "new" baseline rating today.
Chess Tempo - New Baseline Quest - Session 1
Training Time: 0.6 Hours
STARTING:
Stats for blitz tactics
Rating: 1637 (RD: 92.92) (Best Active Rating: 1658 Worst Active Rating: 1443)
Active Rank: Not Active/2571 (Best Active: 1172 Worst Active: 1814)
Problems Done: 305 (Correct: 250 Failed: 55)
Percentage correct: 81.97%
Average recent per problem time spent 46 seconds
Need to do 73 more problems to get a blitz tactics FIDE estimate.
ENDING:
Stats for blitz tactics
Rating: 1687.3 (RD: 34.31) (Best Active Rating: 1696 Worst Active Rating: 1443)
Active Rank: 1119/2573 (Better than: 56.53% Best Active: 1081 Worst Active: 1814)
Problems Done: 371 (Correct: 304 Failed: 67)
Percentage correct: 81.94%
Average recent per problem time spent 37 seconds
Need to do 7 more problems to get a blitz tactics FIDE estimate.
So, in 0.6 hours, I did 61 problems with 81.94% correct AND chopped 9 seconds (on average) off my solving time. My previous best at Chess Tactics Server (before it went defunct) was no higher than 1650 with 79% correct. I'm somewhat (pleasantly!) surprised at the average reduction in time per problem, without a corresponding drop in percentage correct. If I can retain that speed and accuracy, that should reduce the amount of time needed to establish a baseline rating by half.
It's a starting point!
As part of my training at blitz on Chess Tempo, I'm copying all of the incorrect problems into a Word file. There are only three pieces of data: (1) The problem number (so I can revisit it on Chess Tempo if needed); (2) A diagram of the position after the first move (by the opponent); (3) Which side is to move.
ReplyDeleteIMHO, the diagram is the key to grinding the appropriate pattern recognition into System 1. I don't write down a verbal narrative. Instead, I mark the important PoPs with a red square. I then draw LoA from the critical attacking pieces to the PoPs. If there are alternative variations, I use multiple colors of squares and lines. This becomes a VISUAL narrative of what to focus on when I "see" this position (or a similar one) in the future.
After drawing all this stuff, I then work out a "narrative" in my head that binds it all together into variations. BUT, I don't write down the narrative and I don't write down the variations. My hope (is this a BAD variation of Heisman's "Hope Chess"? - LOL!) is that I can train the pattern recognition by "seeing" the PoPs and LoAs, since System 1 does not understand verbalization.
The "proof" will come later, when I get a repeat problem. If I miss a problem, I can search to see if I previously missed it. If so, that's an indication of a serious lack of pattern recognition for the specific motifs/themes involved. That will have to be repeated over and over again until it is MEMORIZED.
I haven't addressed how to use spaced repetition yet. I suspect that creating a set of missed problems on Chess Tempo will eventually be required to use that feature.
It sounds logical that system I might understand pictures better than words.
DeleteI have done my share though with diagrams, arrows, colored squares and visualizing them in my head and stuff, but to no avail. So at least matters aren't as easy as they might seem.
One of the problems I encountered was that a measly pawn on a3 in stead of a2, might alter the assessment of a position from loosing to winning (or vice versa). Meaning that all the geometrical patterns you recognize being equal, they have no necessary relation with the final judgment of the position. That happened too often, and renders the method as being unreliable.
Verbal logic can explain sharply the difference between a pawn on a2 and a3 though. That's why words are way more reliable.
System I is specialized in many things. One of them is retrieving words that are needed to build a sentence. So although system I doesn't understand words, it sure knows how to handle them.
In other words, we need more precise observations since we don't have a clue.
After thinking a bit more about it. The problem is not that system I doesn't retrieve the geometrical patterns at lightning speed. The problem is that the light of attention doesn't fall on the right area's. The light of attention is the trigger that system I orders to start its retrieval.
DeleteWe must find ways to guide our attention to the right area's. The key to that is the tree of scenarios. There is more than one way to implement that tree. Verbal and visual, for instance. Or a combination of both. Either way, it only works when the tree is internalized. One way or another.
https://chesstempo.com/chess-tactics/94924
DeleteI did not find 2.Bc7, why not?
Partly because the bishop was "in the other corner", especially bishops at a2 and h2 are easy to forget.
---But---
I was looking for other methods to unskewer ,like remove the 1. and protect the 2., ...; should have looked for interference. The traditional set of tactical pattern ( most of them are methods ) with its natural extensions ( un-pin, un-skewer...) is rich enough to feed both systems for a while ;)
You can create own tags now as premium. You may use the tags later to create special trainingssets.
ReplyDeleteMaybe for a start use -BS for beeing to slow, -BB for Blundered
the minus will put this tag on top of the tagging list.
PART I:
ReplyDeleteI have been re-reading portions of Dr. Daniel Kahneman’s Thinking, Fast and Slow, searching for clues as how to push knowledge into LTM so that it is (or potentially CAN BE) used by System 1. I gained an “insight” (or hallucination) into how to clarify the “vulture’s eye view” process. I had assumed that this viewpoint could be improved by relying on memory of motifs as the “trigger” or “cues” for specific tactical devices/themes (using Dr. Lasker’s definitions of these terms). I’m going to quote or paraphrase Dr. Kahneman, rearranging some of the text so that it seems more directly applicable to chess. Statements between [] are my own for clarification; emphasis is added using bold, italics and capital letters.
Dr. Kahneman: ~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~
The following anecdote offers an apt description of how System 1 functions. The great comedian Danny Kaye had a line that has stayed with me since my adolescence. Speaking of a woman he dislikes, he says, “Her favorite position is beside herself, and her favorite sport is jumping to conclusions.” [There is a play on words involved in that statement; I don’t know if it translates easily outside of USA English.]
Jumping to conclusions is efficient IF the conclusions are likely to be correct and the costs of an occasional mistake acceptable, and IF the jump saves much time and effort. Jumping to conclusions is risky when the situation is unfamiliar, the stakes are high, and there is no time to collect more information. These are the circumstances in which intuitive errors are probable, which may be prevented by a deliberate intervention of System 2.
[This suggests that there is a possibility of recognizing when we should “turn on” System 2 to either verify or override System 1’s intuitions.]
System 1 neglects ambiguity and suppresses doubts, all without any “signals” to System 2 that it is doing that. [He gives three examples of ambiguity. Two are based on the written cursive “shape” of the capital letter “B” or the number “13”. Imagine (in the case of the letter “B”) that if one writes it sloppily with a little space between the vertical line and the curved line, it might appear to be the number “13“. Or imagine (in the case of the number “13”) that the two numbers are written closer together than usual, making it appear to be a “B”.] If given “A 13 C” and “12 13 14”, you can read the first group as “A B C” and the second group as “12 13 14”. The middle items in both groups are identical. You could just as well have read them as “A 13 C” or “12 B 14”, but you did not. Why not? The same shape is read as a letter in a context of LETTERS and as a number in a context of NUMBERS. THE ENTIRE CONTEXT HELPS DETERMINE THE INTERPRETATION OF EACH ELEMENT. The shape is AMBIGUOUS, but you jump to a conclusion about its identity and do not become aware of the ambiguity that was resolved.
PART II:
ReplyDeleteIN THE ABSENCE OF AN EXPLICIT CONTEXT, SYSTEM 1 GENERATED A LIKELY CONTEXT ON ITS OWN. We know that it is System 1 because you were not aware of the choice or of the possibility of another interpretation. When uncertain, System 1 bets on an answer, and the bets are guided by experience. The rules of the betting are intelligent: RECENT EVENTS and THE CURRENT CONTEXT have the most weight in determining an interpretation. When no recent event comes to mind, more distant memories govern.
The most important aspect of the examples is that a definite choice was made, but YOU DID NOT KNOW IT. Only one interpretation came to mind, and you were never [consciously] AWARE of the ambiguity. SYSTEM 1 DOES NOT KEEP TRACK OF ALTERNATIVES THAT IT REJECTS, OR EVEN THE FACT THAT THERE W-E-R-E ALTERNATIVES. Conscious doubt is not in the repertoire of System 1; It requires maintaining incompatible interpretations in mind at the same time, which demands mental effort. Uncertainty and doubt are the domain of System 2. When System 2 is otherwise engaged, we will believe almost anything. [Re: the “invisible” gorilla] System 1 is gullible and biased to believe, System 2 is in charge of doubting and unbelieving, but System 2 is sometimes busy, and often lazy.
The operations of associate memory contribute to a general confirmation bias. When asked, “Is Sam friendly?” different instances of Sam’s behavior will come to mind than would if you had been asked, “Is Sam unfriendly?” A DELIBERATE SEARCH FOR CONFIRMING EVIDENCE, KNOWN AS POSITIVE TEST STRATEGY, IS ALSO HOW SYSTEM 2 TESTS A HYPOTHESIS. Contrary to the rules of philosophers of science, who advise testing hypotheses by trying to refute them [Popperian falsification], people (and scientists, quite often [AND CHESS PLAYERS]) seek data that are likely to be compatible with the beliefs they hold. The confirmatory bias of System 1 favors uncritical acceptance of suggestions and exaggeration of the likelihood of extreme and improbable events.
Dr. Kahneman: ~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~+~ END
PART III:
ReplyDeleteI think I tried to put too much into the “vulture’s eye view.” The insight (if it is such a thing) is that the purpose of the “vulture’s eye view” should be to establish the context for further investigation of a problem or position in more depth. We don’t “see” as much as we should when calculating variations because the context that would disambiguate the problem or position is missing. In other words, we don’t “see” the forest FIRST before we start looking at the trees.
In that regard, I think Aox’s “thinking process” (approach) to solving makes perfect sense. First, determine which side is to move. This establishes the context for the initiative. Second, assess the relative material on both sides. This establishes one of the most important cues: how “urgent” is the necessity to “make something work”? Third, make an INITIAL assessment of what the problem (or position) is all about: (1) checkmate, (2) gaining material, or (3) Pawn promotion. These three steps establish a context for solving the problem. Perhaps there are more required steps, but I think these may be the minimum steps needed for the “vulture’s eye view.”
Or, I could just be hallucinating. . .
"I think I tried to put too much into the “vulture’s eye view.” The insight (if it is such a thing) is that the purpose of the “vulture’s eye view” should be to establish the context for further investigation of a problem or position in more depth. We don’t “see” as much as we should when calculating variations because the context that would disambiguate the problem or position is missing. In other words, we don’t “see” the forest FIRST before we start looking at the trees."
DeleteExactly.
We are inclined to tag matters with the labels "good" or "bad". Remembering is good, forgetting is bad. Objectivity is good, bias is bad. But what we call bad, always has a function. Forgetting is good, because it prevents memory interference. Bias is good since it speeds up decisions.
The mind is very smart organized. And if we are smart too, we should try to act in accordance with its nature. I don't belief in the system hack of spaced repetition. If the mind wants to forget something, it has a clear reason to do so.
Take, again, the example of the appointment to join a chess tournament in may. You don't need spaced repetition to remember that. If you find it important enough, you will remember it no matter what. If you deem it unimportant, you forget it straight away.
If you deem the tournament as important, you forget the appointment right after you visited it. You no longer deem it worthy to remember the appointment when the day of the appointment has passed.
So if you want to store something in stm, you need to deem it as being important enough to remember. You should add the context that makes it important.
"Or, I could just be hallucinating. . ."
DeleteI consider that to be a different matter. . .
Tempo
ReplyDeleteCould you share your thoughts related to chess puzzles at chesscom? I mean the new feature called PUZZLE RUSH.
I would like to ask you if you could create an article and your opinion towards this "system".
https://www.chess.com/puzzles/rush
I would be interested to test if this method works! If there would be more experimental rabbit, we could see if such approach leads to chess tactics improvement (progress).
Let me know what you (and others) think about it and if you have already checked it out!
I see no reason why this should be any better
DeleteI agree with Aox on this one. The feedback is not addressed. Meaning that system I learns nothing.
DeletePART I:
Delete@ Tomasz:
Aox (in a comment on 15 JUL 2018 to the blog post “Deep Understanding”) gave a link to Blitz Tactics [https://blitztactics.com/] which is (somewhat) a replacement for or an alternative to the old Chess Tactics Server, which no longer exists. It’s free to join Blitz Tactics. The puzzles on the site are from the Lichess tactics trainer.
I think (but I’m not sure) that the Chess.com Puzzle Rush and the Blitz Tactics Countdown are similar in intent, but different in execution.
Puzzle Run gives you a different set of problems for each run you make. You get a different set of problems each run. I have not checked to see if you can set the problem difficulty level for a given run. If you get three problems wrong, you’re done with that run. Otherwise, the run ends after 5 minutes. The goal is to get as many problems right as possible within the 5 minute period. When you get a problem wrong, there is no feedback other than your move is wrong; it automatically goes to the next problem on an error.
Blitz Tactics’ Countdown also gives you a “run” based on 5 minutes. The problem set repeats for each run within a 34-hour period. You get an error message “TRY AGAIN” until you get the answer correct; you can “guess” as many times as needed to correctly get the problem answer without any penalty (other than losing time). When you get all of the moves correct for a given puzzle, it automatically moves to the next problem.
For some types of training (particularly hammering basic tactics into LTM), I think the Bltz Tactics Countdown MAY be more useful, primarily because of two things. You can make as many runs on a given problem set as you want, trying to increase your score (number of problems solved) for the 5 minute time period. AFAIK, there is no upper limit on the number of problems you can solve in a given run. The problem set is identical on each run (presuming you attempt the runs within a specific 24-hour period), giving repetition on the same problems. It also gives you immediate and effective feedback when you make a wrong move, and then allows you to continue with that same problem. It doesn’t tell you what the RIGHT move is; you have to figure it out yourself.
Others may have a different opinion; YMMV.
I’ve excerpted some things from Dr. Daniel Kahneman’s book THINKING, FAST AND SLOW that seem to be applicable to this training idea.
PART II:
DeleteDr. Kahneman gives TWO BASIC CONDITIONS for ACQUIRING A SKILL:
(1) An environment that is sufficiently regular to be predictable.
(2) An opportunity to learn these regularities through PROLONGED PRACTICE.
When both these conditions are satisfied, INTUITIONS are likely to be skilled. Chess is an extreme example of a regular environment. The accurate intuitions . . . are due to HIGHLY VALID CUES that the expert’s System 1 has learned to use, even if System 2 has not learned to name them. . . . If the environment is sufficiently regular and if the [player] has had a chance to learn its REGULARITIES, the associative machinery will recognize situations and generate quick and accurate predictions and decisions.
Human learning is normally efficient. If a strong PREDICTIVE CUE exists, human observers will find it, given a decent opportunity to do so. . . . In the absence of valid cues, intuitive “hits” are due either to luck or to lies.
Whether professionals have a chance to develop intuitive expertise depends essentially on [1] THE QUALITY AND SPEED OF FEEDBACK, as well as on [2] SUFFICIENT OPPORTUNITY TO PRACTICE.
Emotional learning may be quick, but what we consider as “expertise” usually takes a long time to develop. The acquisition of expertise in complex tasks such as high-level chess . . . is intricate and slow because EXPERTISE IN A DOMAIN IS NOT A SINGLE SKILL BUT RATHER A LARGE COLLECTION OF MINISKILLS. Chess is a good example. An expert player can understand a complex position at a glance, but it takes years to develop that level of ability. Studies of chess masters have shown that at least 10,000 hours of dedicated practice (about 6 years of playing chess 5 hours a day) are required to attain the highest levels of performance. During those hours of INTENSE CONCENTRATION, a serious chess player becomes familiar with thousands of configurations, each consisting of AN ARRANGEMENT OF RELATED PIECES THAT CAN THREATEN OR DEFEND EACH OTHER.
Expertise is NOT a single skill; it is a collection of skills, and the same professional may be highly expert in some of the task in her domain while remaining a novice in others. By the time chess players become experts, they have “seen everything” (or almost everything), but chess is an exception in this regard.
Learning high-level chess can be compared to learning to read. A first grader works hard at RECOGNIZING individual letters and assembling them into syllables and words, but a good adult reader perceives entire clauses. An expert reader has also acquired the ability to assemble familiar elements in a new pattern and can quickly “recognize” and correctly pronounce a word that she has never seen before. In chess, RECURRING PATTERNS OF INTERACTING PIECES play the role of letters, and A CHESS POSITION is a long word or sentence.
Acquiring expertise in chess is harder and slower than learning to read because there are many more letters in the “alphabet” of chess and because the “words” consist of many letters. After thousands of hours of practice, however, CHESS MASTERS ARE ABLE TO READ A CHESS SITUATION AT A GLANCE. The few moves that come to their mind are almost always strong and sometimes creative. They can deal with a “word” they have never encountered, and they can find a new way to interpret a familiar one.
[EMPHASIS added]
Correction: The Countdown run problem set changes after 24 hours, not 34. Too much typing too late at night, I guess. Apology, mea culpa.
DeleteThanks for poiting out Blitz Tactics server! I have already signed up and I have done some puzzles for 2 days already :).
DeleteI hope I will practice some blitz type puzzles and after that I will reset my stats at Chess Tempo (it has a new beta version/layout)... and we will see if I can see more (faster) and more correct! :)
I've been trying to use the Countdown, Speedrun, Repetition and Infinity sets of problems. I'm too tempted to "guess" to be able to use Haste regularly. I like the Countdown and Speedrun because I get to repeat the same problems over and over, trying to speed up my "recognition" (or memory recall) on each run and thereby hammering them into LTM.
DeleteOne thing I've noticed (and I try to hammer this idea into LTM) is that whenever I "see" (recognize) a common "pattern" in various problems, I take the time to note it, put a name on it, and then repeat it back to myself a few times. Some of my runs are rather slow as a result. Examples include typical tactical things like smothered mates, back rank mates, etc. I am already FAMILIAR with these tactical concepts; after all, I solved more than 30,000 problems on Chess Tactics Server before it went away. I just try to impress on my brain that these things can occur in various configurations and are IMPORTANT to "see" quickly. Maybe it's just concentrating on them and making sure they get my attention.
In any event, it will take some time to determine if this is useful training or not.
Good luck!
mea culpa: https://www.youtube.com/watch?v=9Grr9nxK1Ts
Delete@ Aox: Je regarde à travers des lunettes roses. Je ne parle pas français - Google Translate fonctionne très bien (parfois). Passez une bonne journée!
DeleteIt took me five minutes to solve this problem, even though I had identified the f3 rook as the overworked piece.
ReplyDeleteI looked at 1...b5, 2.axb cxb, 3.cxb Nxd5, 4.QxNd5 Re2+, 5.Rf2 before noticing that 4.QxNd5+ had come with check, so that couldn't be the answer (plus, it doesn't work out anyway). Then, I spotted the simple solution. I think there is often a period of mucking-about before the simple solution is stumbled upon. Chess is a patience game.
The funny thing is that 1...RxBf1 is the first thing that I looked at, but I didn't really visualize it. There is a trade-off, the rook can't recapture on f1, and stay on f3 at the same time. ;-) Give material to get squares, I guess is the moral here.
ReplyDeleteIt's hilarious how they rate these problems - 1339.
DeleteThat is a 1300 problem for someone in a tournament game how has half an hour on their clock, and has spent half an hour trying to get to this position so that they can play the tactic. Said 1300 player has also been studying a lot of tactics lately, and that's probably all he/she was good at because if he/she was good at anything else, he/she would probably be Class C or B for managing to even get this far into a game, with a position like this.
A 1558 solves it in 55 sec with a Success Rate of 72.7% if they know there is a tactic and a 1339.8 solve it in 03 min 14 sec with a Success Rate of 63.14%
DeleteA 1558 on Chesstempo, which hardly corresponds to a real over-the-board Classical-Chess rating.
ReplyDeleteI am a fide 1911 and my ct-blitz rating is ~1950, as far as i know many ct-blitz ratings are close to the fide rating
ReplyDeleteThe standard rating at ct is a matter of thinkingtime.. as longer you are willing to spend time at these puzzles as hinger the standard rating, my standard rating is ~2300, but i did spend very much time for that at each puzzle.