No Longer Blindly Reaching

Numsgil · Post by **Numsgil** » Fri Jan 13, 2012 2:29 pm

I envy you. I just go crazy trying to do more than 30 minutes of problems a day.

OtakuViking · Post by **OtakuViking** » Fri Jan 13, 2012 2:37 pm

Numsgil, 10000 go problems from 10k to dan sounds absolutely ridiculous. If you need that many, you're not doing it right. You need to mix it up with playing games too. So far I've solved about 2000 go problems, some being opening, endgame and middlegame problems from the graded go problems set... and I'm 1kyu on KGS. So you see ... solving 10000 go problems to get to dan level seems a little ridiculous to me :lol:

karaklis · Post by **karaklis** » Fri Jan 13, 2012 2:51 pm

shapenaji wrote:Most of the time, you're better off with 2-3 10-20 minute games than 1 30 minute game + review.

This is probably not the whole truth. I've seen people playing hundreds and thousands of such games and haven't improved a single stone.

Numsgil · Post by **Numsgil** » Fri Jan 13, 2012 2:59 pm

OtakuViking wrote:Numsgil, 10000 go problems from 10k to dan sounds absolutely ridiculous. If you need that many, you're not doing it right. You need to mix it up with playing games too. So far I've solved about 2000 go problems, some being opening, endgame and middlegame problems from the graded go problems set... and I'm 1kyu on KGS. So you see ... solving 10000 go problems to get to dan level seems a little ridiculous to me

It's in the thousands certainly, and it's more than 1K and less than 100K. Beyond that, I dunno dude. It's probably different for different people. I don't know any dan players who made it to dan without doing a metric ton of problems. I stand by my number as a good estimation of the median.

illluck · Post by **illluck** » Fri Jan 13, 2012 3:03 pm

I also think 10k is absolutely ridiculous. I think I did less than 1000 before hitting kgs 1d, and I don't think they were a significant factor (they definitely sped up the process, though).

Numsgil · Post by **Numsgil** » Fri Jan 13, 2012 3:12 pm

I find it hard to believe anyone could legitimately make dan without doing thousands and thousands of problems. I just don't see how you'd drill in the tesuji, life and death, and shapes otherwise. It'd be like becoming a concert pianist only practicing piano twice a week for 20 minutes. The guys I knew doing music majors were practicing hours a day. It's just the way it is to become good at something.

Some people do have a natural affinity for the game, and if you have a stronger player constantly playing against you that probably counts for quite a bit, but I just don't see anyone getting to dan without significant effort from either problems or the equivalent of problems (game reviews maybe, if you really stop and concentrate and figure out the life and death positions yourself that you missed in game).

Numsgil · Post by **Numsgil** » Fri Jan 13, 2012 3:35 pm

OtakuViking wrote:Numsgil, 10000 go problems from 10k to dan sounds absolutely ridiculous. If you need that many, you're not doing it right. You need to mix it up with playing games too. So far I've solved about 2000 go problems, some being opening, endgame and middlegame problems from the graded go problems set... and I'm 1kyu on KGS. So you see ... solving 10000 go problems to get to dan level seems a little ridiculous to me

Just another follow up (not to derail this thread any more, but I guess it's too late. Sorry hailthorn), is your main account Kamomille? I don't mean to insinuate that anyone that disagrees with me is weaker than they claim! But you haven't played many ranked games, so the sample size is pretty narrow. KGS ranks at low sample sizes are basically pretty random (the +/- factor is several stones). You actually haven't played any ranked games as a 1 kyu yet, even. If you had 50-100 games as a 1 kyu I'd be more inclined to consider you a valid counter data point. A dan in my book means they can play an infinite number of games against other dans and win half of them. Someone who manages to get an account with a game or two to 1 dan doesn't quite count.

Again, I don't mean to be dismissive at all. I'm just saying that if you're presenting yourself as a counter case I'd argue that you need more data points. I don't mean this as a personal attack or anything (it's hard to do unbiased analysis when people are themselves data points :/)

OtakuViking · Post by **OtakuViking** » Fri Jan 13, 2012 3:49 pm

I've hidden my response to Numsgil because I don't want to hijack the thread any more than we've done already:)

Just to clarify. The point of doing Life and death problems is to increase your reading ability. Which helps mainly with your middle and endgame calculations. The point of Tesuji problems is to learn the tesuji and remember it, it's ok to get it wrong, ok to look at the answer. Other types of problems speak for themselves, opening, endgame etc. Notice no shape. Shape comes naturally when you play games, study books like Shape up! Charles Matthews and study joseki.

If you simply glance at a life and death problem, find something that looks like a vital point and immediately jump to the answer. You have totally missed the point of the problem. The solution is not NEARLY as important as the fact that you tried your best to read out all the possible variations to make sure your answer is correct. The main point of life and death problems is simply to train your reading muscle. Life and death problems are usually not limited to shapes found in games, many are artificial and you'll never encounter them in a game. The main point is, I repeat, to exercise your reading muscle.

shapenaji · Post by **shapenaji** » Fri Jan 13, 2012 3:58 pm

karaklis wrote:
shapenaji wrote:Most of the time, you're better off with 2-3 10-20 minute games than 1 30 minute game + review.

This is probably not the whole truth. I've seen people playing hundreds and thousands of such games and haven't improved a single stone.

Well, bear in mind I wrote "experience AND experimentation"

It's very easy to do the same thing over and over and not try out new ideas. The important thing is that you're trying out your ideas, not spending a lot of time on a game.

karaklis · Post by **karaklis** » Fri Jan 13, 2012 4:17 pm

illluck wrote:I also think 10k is absolutely ridiculous. I think I did less than 1000 before hitting kgs 1d, and I don't think they were a significant factor (they definitely sped up the process, though).

I did about 2000 tsumego and got stuck at 7k before I burnt out from tsumego. Your improvement is obviously not only dependent on tsumego but many other factors as well.

Numsgil · Post by **Numsgil** » Fri Jan 13, 2012 4:18 pm

OtakuViking wrote:I've hidden my response to Numsgil because I don't want to hijack the thread any more than we've done already:)

You're correct. It is my new account. And I have played some ranked games on it. I originally started it as 3k but it quickly moved up to 2k (Old account was 4k, nibbler). Then I stayed at 2k for awhile and played a few ranked games to get to 1k. The reason I mainly play free games is to ensure it doesn't get stuck. What does it matter? I also play on tygem (2d.. 1k and 1d on tygem seem stronger or atleast more sandbagger infested). Regardless, I'd be willing to play a few games with you if you want to ascertain whether I am a worthy 'counter' Just send me a pm when I'm on KGS and we can get set up:)

It's a numbers game, you see. To me, X kyu means you've successfully held your own against a wide range of players and styles on numerous occasions. With a wealth of data points we can find a rank that fits the data. That is, it's not a mark of how strong you are currently, but a statistical statement about past performance. See sample size calculator for instance. Let's say we want to be confident that you're 1 kyu instead of 2 kyu. We'd expect you to win 60-66+% vs 2 kyus (a stone is roughly 1 standard deviation). So to determine the difference between 50% (from 2 kyu) and 60%, we want a sample size of almost 100 games. Less than that, and the margin of error swamps the difference between the two ranks. Conversely, go to the confidence interval calculator. Plug in, say, 15 games with a win rate of 70%. The +/- % it gives back is 23%. That means, statistically speaking, a 70% win rate is indistinguishable from a 50% win rate with only 15 games. You could get a 70% win rate at 15 games just playing rock-paper-scissors or flipping a coin and it wouldn't be that unusual. We just can't draw any conclusions without a proper sample size of games.

...

I'd love to play you, but win or lose wouldn't necessarily convince me one way or another. There's a prolific 4 kyu I play against from automatch on KGS, for instance, that just destroys me over and over again. He has like a 80% win rate against me over 11 games. If we remember the sample size thing I talked about above, even those 11 games isn't really enough to know if he's stronger or weaker. It's within the margin of error of 50%. We could do a series, if you want, though. The goal, to be statistically significant, is to win at least games * .5 + sqrt(games) games. The winner is statistically significantly better than the other player. That can be a lot of games, though (it easily reaches dozens of games for almost evenly matched players), and most of my friends get mad at me when I suggest that instead of "best 2 out of 3".

Anyway, yeah, hit me up on KGS. It's a long weekend for me so I'll be on a lot in the next few days. I'll add you as a friend and see if you're on when I am.

illluck · Post by **illluck** » Fri Jan 13, 2012 4:35 pm

I find it a bit amusing that less than 1k problems is considered "hard to believe". As Karaklis says, progress is dependent on a lot of factors. Problems is not a necessary condition for reaching 1d. If anything, I think the number of games played is more meaningful, but even there different players have very different rates of progress. I probably played more than 2k games before hitting 1d, whereas someone with about 700 is almost dan (and surely there are players who reach dan in significantly fewer games).

hailthorn011 · Post by **hailthorn011** » Sat Jan 14, 2012 11:38 pm

Self-Review Game #2:

Let me begin by saying that I enjoyed this game very much. But it was marred by mistakes.

1. Still playing too fast.
2. Failure to recognize Big Moves

Black could have won this if not for a few critical errors which I point out to the best of my ability.

OtakuViking · Post by **OtakuViking** » Sun Jan 15, 2012 5:28 am

A few quick comments.

Move 7. I understand why you played K16, but O17 already reduces whites potential with that opening. Maybe approach the 4-4 or make an extension of your own in the middle?

Move 9. Playable, but game variation was weird and not very good. White should C16, then black jumps away. But I would play elsewhere... There are bigger things to consider than invasions.

Move 31. Don't cut if you can't follow it up. You just end up with a weak, useless group that you look after. Better abandon those stones and aim at something else.

move 47. Yes:=)

move 51. WEAK. Jump and continue attack.

...no more time, hope it helped:=)

hailthorn011 · Post by **hailthorn011** » Sun Jan 15, 2012 10:27 am

OtakuViking wrote:A few quick comments.

Move 7. I understand why you played K16, but O17 already reduces whites potential with that opening. Maybe approach the 4-4 or make an extension of your own in the middle?

Move 9. Playable, but game variation was weird and not very good. White should C16, then black jumps away. But I would play elsewhere... There are bigger things to consider than invasions.

Move 31. Don't cut if you can't follow it up. You just end up with a weak, useless group that you look after. Better abandon those stones and aim at something else.

move 47. Yes:=)

move 51. WEAK. Jump and continue attack.

...no more time, hope it helped:=)

Thanks for the comments! As for your comment on Move 31, I completely agree. Cutting there was a bad decision that hurt me quite a bit in end game.

Life In 19x19

No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching

Re: No Longer Blindly Reaching