@SummerThereof said in #3:
> My intuition is telling me that facial recognition is different from what chess masters do. It seems that someone who is good at facial recognition would be able to ignore "noise", (which was part of the Cambridge test), like with a beard or without a beard. Whereas chess masters very often point out things like "position x is just like position y except that the pawn is on a rook file..." Chess recognition seems much more dynamic than facial recognition. Maybe I'm oversimplifying facial recognition.
>
> I got an 86% accuracy on the Cambridge test, which is better than my chess accuracy.
I like the idea about noise in face images and chess positions. It's a neat question as to what constitutes noise in a configuration of pieces, and I wonder if the Einstellung effect (which I keep meaning to write about) is a relevant thing to think about in this context. Even GMs can sometimes get distracted by one aspect of a position and miss variation that changes the best move, which may be a clue as to what their definition of signal vs. noise is like.
Also, I should say that I don't mean to suggest that face recognition and chess share common mechanisms. There is the paper I referenced trying to make a connection between the two, but my feeling is that there are probably more mechanistic differences than similarities. The idea here was more to consider how we define expertise in face recognition and how that can lead to selecting a population for study that may be more or less informative about typical processing.
There is a small z and a big Z. Now, I get it, A typical continuous distribution of some extrinsic "trait" (combining many factors) in one dimension the Z axis, and that density is phi(Z), while the under slice areas under the surface, are parametrized by z (lower case), is that other Greek letter, now my memory suggested psi, but it could be the ancient Greek uppercase of phi (as an integral, it would not be a bad arbitrary choice, for convergent (cooperating?) ).
The other T is also an integral under the curve of phi(Z). and it depending on the variable named lower case z.
So, Phi(z) = Integral ( phi (Z), Z = [ -infinity, z] ), and T(z) = Integral ( phi (Z), Z = [ z, +infinity ]
And I understand that there were enough visual signals, barely, to distinguish all the mathematic function symbols on display, given it was used suggestively, and not demonstratively, as cutting short of all the scripture I just produced, that would have been overkill verbal language, when the image was already doing the job (which was not to prove anything, but a nice wink to the running gag that is often obsessing a bunch of us).
I think I just finished reading the blog. (I do chunks of many things over many days).
This blog and the preceding, have the commonality of sharing a bit of the trade of your profession, which as often relied on life accidents, or outlier individuals in general in their observable behavioral traits, that we could also link to neurobiological observable (that is the most concise non rambling, but ends up a bit clinical, technical, or even pompous? of me?).
Vascular accidents, or even developmental ones, or injuries to the brain. Those have allowed us to inverse research normal function hypotheses (us=you=them=all of us in general because we share in science, we do not hoard secrets, still all of us in general, because secrets never last anyway.. just wait a generation or 2).
I see now the discussion about nature versus nurture, and it reminded me of other work on the perception of music.
We are all able to enjoy music. We might even become avid music listeners and know a bunch as listener.
But then the music instrument trained individual we would call musicians, they have a different training, that is not only passive enjoyment. They end up also completing their experience expertise as listener, the "holistic" high level part of your model (and the emerging one, still, I am a bit outdated about such things, I know that model has succeeded to the art/logic split model).
The musician trained, is able to dissect a lot more information, with the same listening experient. Isolate instruments, time armatures, keys, harmonies, and some might even combined with some visual decrypting of music score notation linking to that music production training. Music theory in between. But even without music theory, there are musicians that can do that fine grain cognitive work, from having had to produce the notes. (using other encodings of the fingering to hearing tasks, bypassing the written code). Sorry, I go where the begginning of my sentences bring me, typing being a pain with accumulation, and my thoughts while writing as well. This last theory and written notation part is bonus rambling, p0erhaps a question too, maybe relevant to chess? or of getting myself updated.
where was I going. this was about things I gathered from your blog, that I wanted to review positively here..... Well, i think i covered it.. but I just feel like I flushed another bit from my working memory, by going into the music notation thing.. Bah,
I can always go back, and ramble some more, when it comes back.. upon another iteration at reading..
image illustration: check
holistic - analytic dichotomy check
tools of the empirical part of the wheel of science (outliers) check
hmmmm.. I might have wanted to talk about the intermediate thing. But now need physical neck shoulder arm rest...
If we do study GrandMasters and understand Chad’s and the mind, what will we get out of it? And How will it benefit chess?
Hi, I just wanted to say thanks for shouting out my book. I recently discovered your blog and am loving it and learning a lot, keep it up!
Ben
@PerpetualChess said in #15:
> Hi, I just wanted to say thanks for shouting out my book. I recently discovered your blog and am loving it and learning a lot, keep it up!
> Ben
Thank you for reading and for the encouragement! I'm also enjoying your book a lot and learning a great deal from it. If you ever think it would be fun to talk about any common interests, let me know!
noise or irrelevant information. There can be features existing that might not be noise intrinsically, but that might be irrelevant to a certain degree when combined with other co-existing feature in the same position.
a feature might already be a pattern building block. Like a pin for example, it might not be noise like where a line piece is on an important diagonal but not exactly same location on the diagonal, that slight difference if it does not affect the continuation or the desirability of the position if considered in imagination or future possibility. But the pin as goal of some plan internal debate, might not be that important give the context. I think given that chess is "locally" deterministic, we can account for all part of the board as not noise, or maybe noise in the evaluation maybe.. Just chewing..
I did not look at the test noise. but now I wonder if that was not was noise meant up there. irrelevant information.
The word noise might just be not the right word in both cases.. It is not about uncertainty, or masking. it is about weighing patterns toward an explicit goal. so. just a word problem then.
