Done.

"Wrong answer, that was so bad!"

_________________
- Brady
Want to see videos of low-dan mistakes and what to learn from them? Brady's Blunders

Yeah, that hit me in the face, too, today not really a great motivator <shrug>

My “rank”, as estimated there, has been moving in interesting ways: from (IIRC) something around 20k–17k to 10k–8k to 15k–13 to 12k–10k. I “like”


Grtz, Tom

_________________
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS?

Well, emotions do improve the ability to retain a memory afaik.

http://en.wikipedia.org/wiki/Emotion_and_memory

But ya, need to keep working on the exact wording and other details. Still far from what I want it to be at the end.

Well, yeah … I remember punishment well from my childhood, and it was more physical than many here will imagine.

Meanwhile, though, I respond better to positive reinforcement http://en.wikipedia.org/wiki/Reinforcement

And I need MUCH more than a green number

But really, what I’d need most, both with “mistakes” and “correct” moves, would be an explanation, and if it would only show the negative consequences of the mistake or the “profit” in case of a correct move. But I guess this would be quite difficult.


And good work it already is, I really love your site, and I’m sure it will do me good.

Thanks, Tom

_________________
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS?

I hear many kyu players wishing for the explanation. In truth many times there is no trivial explanation possible. The way I approach a wrong move is I first try to read what local disadvantages my move has compared to correct move. If I find none I look for global reasons, trying to estimate how fuseki would progress with my choice compared to correct move. If I found a reason or something that I think might be a reason I try to memorize it. If I find none I just try to memorize the approximate position and come up with a generic rule such as "answer approach on hoshi with keima" and memorize that.

Not surprising, is it?

Yes, I thought (and thought I wrote) so.

(underlined by me)
Yes, I think it’s exactly these questions that are hard for a kyu player:

- What are the local disadvantages?
- how would/could the fuseki progress?

OK, that’s quite the same way I do it, it’s just that I lack the experience for answering the above questions.

I think it would really be good, somehow, to see some continuations of the “bad” and “good” moves, though I believe this could be very difficult to do, because both could respond somewhere else, and the “punishment” of bad moves and profits of good moves will often be seen much later only.

<sigh> Perhaps I should switch to Tic Tac Toe.

Greetings, Tom

_________________
“The only difference between me and a madman is that I’m not mad.” — Salvador Dali ★ Play a slooooow correspondence game with me on OGS?

You could try and "crowdsource" continuations from strong players.

A lot of times, the most interesting projects are those that let a technical or algorithmic tool enable collaboration.

_________________
Occupy Babel!

I find it already fantastic to be able to see the continuation of the "good" moves, with the links to pro games. There I can sometimes understand what was the intention behind a move. Everytime I find a meaning like that, I learn something new that I missed. Then I try to find a good name for it and to memorize it.

It would indeed be very helpful if strong players could add comments like "extend and attack at the same time" or "make your weak group stronger and opponent's group weaker" and the like.

Just look at the games that have the given position - there's your explanation (as well as the source of the reason for the "correctness" of the answer). You can how the pro games turned out in that case.

_________________
be immersed

Yes, I suggested this earlier, especially for 'highly contested positions' i.e. positions where many people make the same 'mistake'. I have access to a number of professional players and wouldn't mind helping out where I could, since I like this project a lot.

Very nice site, fun to try. But I agree with others that the rating function could use improvement. I think most users would like feedback on whether they can choose good moves, rather than moves which happen to be fashionable in recent professional play.

For consideration in your collection, the position should really occur many times in a large database. Otherwise, how do you know that there is only one correct move, or that you have a good statistical sample of correct moves?

The cleanest problems would be positions in which only one next move ever occurs in professional play. If this does not provide enough difficult problems, then expand the list of correct answers to include any move selected by a professional in any game. A good problem should still have a fairly limited number of correct answers.

As Bill has tried to point out, broadening the set of correct answers this way will improve the ability of your system to distinguish strength differences at all amateur levels. If your system then loses the ability to distinguish a 1-dan professional from a 9-dan professional, that is probably a loss you can live with.

+1 & Thank you. This has put my earlier question in much better words.

Yes, there is a bug right now because of which it doesn't show as many high quality positions as it could. Fixing it today. Important part here is tho, what is considered correct? Should a move that only loses 0.1 points of value still be considered good enough? The way its worded now is that it actually doesn't ask you for the right move, it asks you for the most common move among pro players in the given situation. Also when calculating the rank it actually takes into account the reliability of the position. If it only has a few pro examples it weights much less in the rank calculation than some position with lots of examples.


From a tsumego perspective yes, I agree. Standard requirement for good tsumego is that it only has one correct answer. However when studying fuseki, I would hesitate to use that criteria. You basically want to see what is possible from pros point of view. And I find much more important to have a distinct difference between moves that can be played and moves that cant be played. What of a situation which have occurred 150 times in pro games. Move A has been played 135 times, move B 14 times, move C 1 time. Now should move C be considered viable or not. Maybe it was a brainfart or a missclick or just some crazy experimentation that turned out to be really badly never to be tried again. B should almost definitely be still included as an answer, but why is it A played 10 times more often? So I much rather take those 150 possible answers as 30 for A, 30 for B, 30 for C, 30 for D, 30 for E. Then its clear that there are 5 possible choices, all reasonable, and nothing outside those 5 should be considered much.


Strongest users (as per their systems rank, not the self provided rank) do actually affect the possible answers to some degree, and have from the very beginning. Its the reason the system tells what strength players the answers are based on. Since if we have only 1 pro example from a position, but many answers from users its only reasonable to actually diversify the possible answers even if it comes at a slight expense in quality, so it ends up combining the few strongest users answers with the pro player. But it does protect strong users by actually letting them know that the answers are based on users who are ranked close to their own strength and not pro only pro players. So in time the possible answers list will become broader for each position.

Having a more diverse valuation for different answers, than just correct, maybe, wrong, and having more possible answers, would indeed allow a user to be rated faster, with fewer questions. However even if only the one most common move was considered correct and everything else as wrong, it should still be able to figure out users rating just as accurately, it would just take longer time. Of course this would feel annoying to the user as the requirement would be too harsh. Now I'm trying to strike a balance between the service still being useful for those that need harsh criteria (strong players who need the 100% answer instead of the 99% answer). While at the same time showing as many reasonable moves as possible.

What I find most problematic is that some users have very outdated fuseki, or alternatively are have a handicap fuseki style (combined with a very fighting oriented playstyle). Both of those end up being ranked somewhat weaker than their play actually is. To fix this Im going to introduce some tsumego problems for a more balanced evaluation of users rank.

This is partially what worries me. You're looking at a relatively small sample of games over a short time period. Very, very easily a move that is making up a low % of games might actually be the currently considered best move because once it was discovered and the results seen in a few games the previous move has been avoided until a counter to the new move is found. Going simply by frequency you could be punishing the only correct answer for this position currently considered and rewarding answers that are now considered inferior but were widely played in your sample beforehand.

Honestly, I think you really need to take simple positions or get a very strong player to vet questions for you to avoid this or similar problems (e.g. common but now refuted responses etc).

What if C was played in on of the 5 most recent games in the database?

What if B was played 10 times in the last year and 4 times before that, and A 20 times in the last year and 115 times before that?

Both such situations could indicate that a new move has been found, perhaps better, or just as good. Or they could simply indicate a change in style or fashion. Or, at those kinds of numbers, even the fact that a new pro has entered the scene and that he is very much a fan of a certain fuseki pattern.

I don't think there is an easy answer. I do think that finding any move played by a pro, even if turns out to be a pro-level brainfart, strongly indicates that at least your instincts are in the right place.

Anyway, thanks for the effort, it keeps improving and I think a lot of people will learn a lot from it.

There are 15000 games over 5 years. If a position has only a few examples, then it can hardly be considered common nowdays. And for a rare position the basis for the best move is actually just one or two pro players oppinion, those who were in the position within last 5 years. The move they played are clearly their oppinion of the best move for that position. Having another pro player vet the position probably wouldnt help since then we would just be exchanging one pro players oppinion for another. Yes there are a few positions which used to be considered standard but have now been "refuted" and he might help there. However for now my choice of using 5 years worth of games is an attempt to strike a balance, between having modern enough situations that there shouldn't be too many outdated fusekis, and the few that there are should be end up in the "maybe" category instead of "correct". While at the same time providing a diverse enough choice of possible moves for a position.

In actuality in most situations, there is only one correct move. Everything else leads to a loss. Well might be that regardless what move you play you will lose anyways, assuming both players play perfectly. Since we obviously can't go by those standards, as no one knows the right answer for most positions, we just have to settle with best available knowledge. If I were to ask a pro to vet positions we would just be exchanging the collective wisdom of pro players for one particular pro players oppinion. In truth no one can create a perfect set of situations with only the correct answers, it simply cannot be done. We are left with trying to minimize the errors and have positions and answers that are educational despite their imperfections. Just like you learn from studying pro games, even tho you know they make several mistakes during every game, the sum of the parts adds up to a positive value.

This is actually on the todo list. I intend to weight the pros answers according to how recent they are. So recent examples are valued much higher than older examples, despite having both in the database. While again not perfect, it should reduce the number of absolete fuseki variations.

I don't mean this harshly, I hoping this is constructive criticism. What brings me to look at this this way was a lecture by Guo Juan that I wanted about the Small Chinese, she took time to point out several moves that would crop up as very common in the databases that were now not being played because of a flaw being found through research and pros moving to a different response in this position. The caution was that just because you can find many pro examples doesn't mean that a move is actually a good move in the fuseki. Rating moves through frequency alone will often work but almost certainly throw up some instances where you're marking as correct moves that (at least) high level players should be avoiding because they are known to be suspect now. Simple frequency analysis is a great start but you'll have to move past that to get closer to where you seem to want to go. Again, not intended as harsh just hoping to be helpful. Opening a discussion on how get an algorithm to spot these "common but refuted" moves as opposed to simple fashion changes might be an idea? (looking at oversampling from a single player's games or a single "school" of players might be an idea?)

Detecting fashion changes from refutations is probably very, very hard. As a matter of fact, detecting sudden (but minor) changes in continuous or discrete signals (in an automated manner, so you get an alert) is relatively hard (again, for minor changes you'd be able to eyeball but still wonder if it's that big of a change: you need fairly above-average techniques in signal processing just to detect small spikes or trends,) in a go board it would be almost impossible. Imagine having move A a keima, move B a 1-point jump. A has been played a lot, whereas B has been played recently. A players win 50%, B won his game(s) (given a sample size of, say, 3 games vs 50). How could you tell with this data (which is mostly what you'd get with automated analysis until go engines improve 5-6 stones) that B was a good innovation and not just a coincidence? Or even how to distinguish if it was just a change of fashion (Koreans attack, let's all play the keima! Japanese defend, let's play the 1-jump!) or a complete refutation?

_________________
Geek of all trades, master of none: the motto for my blog mostlymaths.net