Zettelkasten Forum


Discussion: Learning from Chess about Thinking (What is a good idea?)

Hi Zettlers,

this is a new idea of mine that I don't understand well yet:


The value of a position in chess consists of its probability of victory for the player. Moves are evaluated according to how they affect this probability of victory.

We can apply this to our thinking: knowledge-based value creation consists of enriching data, information, and pre-existing knowledge with value-giving properties. Value-giving properties include: Accuracy/Completeness, Reliability, and Usefulness.

An evaluation of thinking would look like this:1

symb. designation comment
!! Brilliant Best idea, not easy to find
! Great Thought changes the course of the debate
Best move Theoretically best thought
Excellent Almost as good as best thought
Good All right
Book Conventional
?! Inaccuracy A bad thought
? Error Directly worsens the position
Missed Opportunity
?? Blunder Very bad thought
loss of value, loss of ability, loss of character

I always remember a discussion between Prof. Beckermann and Prof. Carrier (two of the high royalties of the university of Bielefeld, yes I created a heraldry table for my faculty back then) which started by them just barking bullet points at each other. It was just like a chess opening, which is fairly researched, and most moves are known in their effect and values. In philosophy, there are typical moves as well. A debate on the freedom of will almost certainly will hit distinct positions and moves like soft vs hard determinism etc.

An example of a blunder is: Mentioning an exception as an objection to a rule, like stating that there are horses with three legs, to object to a concept of a horse that includes horses having four legs. It is a blunder because you don't challenge the rule by an exception. To me, it is a big blunder, since it even undermines the ability of the participants to establish common ground.

This idea is still in the ruminating stage. So, I mostly collect claims, isolated connections to other ideas, anecdotes, examples and stuff like that.

It is an idea that I often ruminate, since I feel that this could be some basis for a training method for correct (consistent, coherent, cohesive) thinking. I am often inspired by a particular paper on training in Science and Engineering.2 At first, it gave me a framework on how to think about problem-solving. But I feel that there is way more hidden. My guess is that there is something that could reaminate Logical Positivism. I'd exchange the strict and inflexible premise of deductive reasoning with the more flexible assumption that there are good and bad strategies and tactics in thinking which increase or decrease the probability of value-creation. Then, it is not any longer the goal of proving something right or wrong, since both are too far in the future. The goal is to develop tactics and strategy similar that improve your position.

The core of the idea is not new or revolutionary at all. But still, I have the strong intuition that using chess as a metaphor and building a model similar to chess theory is a treasure chest hidden on an island.

What do you think?


  1. The evaluation of thought is worded to be expressed in a fictional debate. ↩︎

  2. Argenta M. Price, Candice J. Kim, Eric W. Burkholder, Amy V. Fritz, and Carl E. Wieman (2021): A Detailed Characterization of the Expert Problem-Solving Process in Science and Engineering: Guidance for Teaching and Assessment, CBE---Life Sciences Education 3, 2021, Vol. 20, S. ar43. ↩︎

I am a Zettler

Comments

  • @Sascha This is an interesting approach to assessing the value of an idea. I suspect if you rated each idea or each zettel, that rating would change over time - as you fine-tuned a zettel (because of better thinking or changed data) or simply because your sense of its value changed (without judging in what direction that change would most likely be). I can see how the chess analogy might be helpful, but I wouldn't want to get stuck thinking of it in just that way.

  • @GeoEng51 I think that Sascha's referring more to intellectual debate. Though you could view Zettling as a sort of debate, I think the value of a note (as opposed to an idea) is more usefully gauged by the depth of processing, which develops over time in the scheme of a ZK.

    Of course, there are different types of value. "Accuracy/Completeness, Reliability, and Usefulness" seem to be intrinsic properties that Sascha is talking about, which remind me of those on the flower of knowledge.

    @Sascha However, what I got from the chess analogy was a description of value in a discussion, or at least "thinking", like some sort of logical chain of thoughts...? And I think that value in that context of a debate is quite different, where your mention of "blunder"s fits in. The goal would appear to be some sort of mutual understanding/development of thought. The issue is that, in my experience of chess, only one person wins.

    But I can also see why you chose chess. It is that balance of being simpler than life without being solved, and so it can be used to frame unsolvable problems in a way in which our dealing with them can be improved. The logical extension is that we improve at this best by using tactics and strategies.

    And my ZK reminds me,

    "Tactics are what you do when there's something to do, strategy is what you do when there's nothing to do".
    – Savielly Tartakower

    There might be value—whatever that means—in the chess connection if you able to mechanise it, but perhaps not as a complete analogy. If you use a criteria of intrinsic properties of value, how does this apply to ideas in the context of a discussion?

    It reminds me of my vexation at the chess game review, when it told me that I made an error when I had been trying something new. In the long run, those mistakes were more than worth it.

  • edited April 2023

    Thank you @Sascha, for sharing chess as an interesting metaphor for thinking. The part that captured my attention is:

    My guess is that there is something that could reaminate Logical Positivism. I'd exchange the strict and inflexible premise of deductive reasoning with the more flexible assumption that there are good and bad strategies and tactics in thinking which increase or decrease the probability of value-creation. Then, it is not any longer the goal of proving something right or wrong, since both are too far in the future. The goal is to develop tactics and strategy similar that improve your position.

    I want to emphasize your idea that "there are good and bad strategies and tactics in thinking which increase or decrease the probability of value-creation."

    The consequences of actions have continuously increased or decreased the probability of favorable future outcomes, not just in chess but throughout life. Some strategies and tactics lead to a high level of learning and flourishing, some only part way, and some lead us astray. It's a spectrum. Some tactics significantly increase value creation probability, and some steer us away from value creation.

    In this view, you'd want to use strategies that position you for the highest probability of winning the chess match, creating value, or flourishing in life.

    I subscribe to Richard Feynman's "The first principle is that you must not fool yourself, and you are the easiest person to fool." While my current tactics reflect my future position's probability, I am open to discovering better tactics like you.

    Will Simpson
    My zettelkasten is for my ideas, not the ideas of others. I will try to remember this. I must keep doing my best even though I'm a failure. My peak cognition is behind me. One day soon, I will read my last book, write my last note, eat my last meal, and kiss my sweetie for the last time.
    kestrelcreek.com

  • @GeoEng51 said:
    @Sascha This is an interesting approach to assessing the value of an idea. I suspect if you rated each idea or each zettel, that rating would change over time - as you fine-tuned a zettel (because of better thinking or changed data) or simply because your sense of its value changed (without judging in what direction that change would most likely be).

    Sure thing. The word "idea" might mislead. An idea would be more akin to a position.

    My core interest is more in the next step, finding good moves. (or seeing bad moves)


    @Warhead said:
    Of course, there are different types of value. "Accuracy/Completeness, Reliability, and Usefulness" seem to be intrinsic properties that Sascha is talking about, which remind me of those on the flower of knowledge.

    That is one issue of the metaphor: In chess, there is an ultimate goal, the win. That is not the case in knowledge work in general. One might, decide on the ultimate goal (usefulness as a businessman, truth as a scientist) and subdue the other value aspects. But I am not sure.

    @Sascha However, what I got from the chess analogy was a description of value in a discussion, or at least "thinking", like some sort of logical chain of thoughts...? And I think that value in that context of a debate is quite different, where your mention of "blunder"s fits in. The goal would appear to be some sort of mutual understanding/development of thought. The issue is that, in my experience of chess, only one person wins.

    I don't think there is much difference between a debate and chess if both are played with the meta goal of self-improvement in mind. Being wrong and losing are both learning experiences. I think debating is a much-needed training method for thinking technique because it provides much feedback.

    So, winning a debate is neither better nor worse than losing it if the actual goal is self-improvement. Same in chess. Same in any game/competition. (I personally liked it when I was wrong or debated in a losing position back in university, since I learned way more.)

    "Tactics are what you do when there's something to do, strategy is what you do when there's nothing to do".
    – Savielly Tartakower

    There might be value—whatever that means—in the chess connection if you able to mechanise it, but perhaps not as a complete analogy. If you use a criteria of intrinsic properties of value, how does this apply to ideas in the context of a discussion?

    I am not so much about mechanizing, but about using it as a creative technique. In chess, you can force moves which would be akin to formal arguments that "force" the truth of claims. But this is just a small part of chess, and a tiny part of thinking and understanding.

    I think I failed to communicate the priority of thinking before position. To me, ideas and positions are almost just necessary conditions for elaboration and thinking. The relations between things are primary to me. In a sense, the relationship between two ideas comes before the ideas themselves.

    It reminds me of my vexation at the chess game review, when it told me that I made an error when I had been trying something new. In the long run, those mistakes were more than worth it.

    Interesting. I for example do both. I try new things and look forward to the AI of commenting it. A better scenario is that I could tell about my intentions and the AI could comment on that. (I chess teacher would be better of course)

    I am a Zettler

  • @Sascha, have you seen this?

    How to become a successful physicist.
    This is helpful and non-trivial.

    GitHub. Erdős #2. CC BY-SA 4.0. Problems worthy of attack / prove their worth by hitting back. -- Piet Hein.

  • @ZettelDistraction said:
    @Sascha, have you seen this?

    How to become a successful physicist.
    This is helpful and non-trivial.

    Yes. I referred to the primary source (Price et al., 2021) above. This is still on my reading list. :)

    I am a Zettler

  • @Sascha said:

    My guess is that there is something that could reanimate Logical Positivism. I'd exchange the strict and inflexible premise of deductive reasoning with the more flexible assumption that there are good and bad strategies and tactics in thinking which increase or decrease the probability of value-creation. Then, it is not any longer the goal of proving something right or wrong, since both are too far in the future. The goal is to develop tactics and strategy similar that improve your position.

    As you may know, there were many mid-20th-century thinkers who developed argumentation theory explicitly in response to the limitations of logical positivism/logical empiricism. Karl Popper (mentioned in the Critics section of the Wikipedia article on logical positivism that @Sascha linked above) and his theory of "conjectures and refutations" was one early example. Popper and others influenced Else Barth and Erik Krabbe's work in dialogical logic, which formalizes argumentation in terms of "moves" in dialogue games. (The chapter on "Formal dialectical approaches" in the Handbook of Argumentation Theory provides an overview of much of this work.)

    More informally, Frans van Eemeren and Rob Grootendorst's theory of pragma-dialectics talks about strategic maneuvering and provides rules for critical discussion (which are the "rules of the game" of argumentation, and any violation of them is a fallacy). And there is other similar work by scholars of argumentation theory (including all the work on argumentation schemes).

    The education researchers Carl Bereiter and Marlene Scardamalia, known for their work on knowledge building, wrote an about their conception of good strategic "moves" in knowledge-building discourse (Carl Bereiter & Marlene Scardamalia (2016), "'Good moves' in knowledge-creating dialogue", Qwerty: Open and Interdisciplinary Journal of Technology, Culture and Education, 11(2), 12–26):

    The notion of "good" moves is partly inspired by the intriguing finding of Chase and Simon (1973) that chess grand masters do not actually think farther ahead and consider more moves than lesser players, they only consider good moves! How in the world can one identify good moves in advance of analyzing them? The explanation apparently lies in superior knowledge. Chase and Simon emphasized the vast tacit knowledge chess masters accumulate, knowledge that provides them with an efficient "vocabulary" or set of categories for considering moves. It has further been established that master chess players read a lot; in one large study, amount of time devoted to study was found to be the strongest predictor of competitive standing (Charness et al., 2005). This accumulated knowledge is what we have called knowledge of "promisingness" (Bereiter, & Scardamalia, 1993). It is a cognitive resource that supports not only wise judgment but also creative thinking.

    In fact, any teacher who teaches argumentation needs to have a rubric for evaluating their students' work, and it's easy to find such rubrics online (here, here, and here are the first three of a vast number of English-language Google search results), so giving "points" for "moves" in knowledge-building argumentation probably already happens in very many schools during every academic term! It seems like a good way to teach the skill.

  • Also, remember Stephen Toulmin's point in The Uses of Argument (1958) that while the most general procedural forms of argumentation are "field-independent" (independent of the subject being discussed), many of the criteria for evaluating discourse are "field-dependent" (related to the subject being discussed). Strategic "moves" in knowledge building could be either field-independent or field-dependent. Knowledge-building skill requires both general and subject-specific training. This is a way in which knowledge building, considered in general, is not like chess, because chess skill is limited to one field (chess), but knowledge building can happen in any field, and different fields have different field-dependent evaluative criteria. (But, yes, chess skill is a good example of knowledge building, even if it is not a universal model of knowledge building.)

    For example, the excellent recent book Qualitative Literacy: A Guide to Evaluating Ethnographic and Interview Research by Mario Luis Small and Jessica McCrory Calarco (University of California Press, 2022) provides evaluative criteria specific to qualitative social research. The authors argue that people who are only familiar with the evaluative criteria for quantitative research can lack the knowledge necessary for evaluating qualitative research:

    Large-sample surveys are optimal means to accurately describe a large population; experiments, to precisely determine the effects of a cause in a controlled setting; participant observation, to directly observe phenomena in their natural contexts; and in-depth interviews, to elicit explicitly how people understand themselves and their circumstances. The techniques required to effectively perform each task differ dramatically from those needed to perform others.
    By extension, each task calls for its own evaluation criteria. For example, we believe that while a sufficiently large sample is essential for an effective survey, that characteristic is perfectly fine for a high-quality interview study but not essential to it. Similarly, while an adequately selected comparison group—that is, control group—is indispensable for an effective experiment, it is not essential for an effective ethnography. Just as we would not fundamentally assess a laboratory experiment by whether participants were selected with known probability from a large population, and just as we would not fundamentally assess a survey by whether respondents were randomly assigned to a treatment, we also should not fundamentally assess open-ended interviewing or participant observation by criteria, or with language, immaterial to their core strengths. For these reasons, language such as "selection bias," "sampling bias," "variance," and "control group" will not form part of our discussion.
    Given the core strengths of in-depth interviewing, the method should be assessed primarily on whether the researcher effectively elicited how people understand themselves and their circumstances. Given the strengths of participant observation, it should be evaluated primarily on whether the ethnographer effectively observed social phenomena in their context. Our task in this book is to specify precisely what "effectively" means in each case.

    Small and Calarco explain five criteria for evaluating qualitative social research: cognitive empathy, heterogeneity, palpability, follow-up, and self-awareness. Not all of these criteria would be relevant for all other fields.

Sign In or Register to comment.