Thursday, September 21, 2006

Paying For Accuracy

Three weeks ago I used to play a level of about 1460-1470 at Chess Tactics Server. I was quite happy with a 90% accuracy then, compared to my former 80%.

But still I felt uneasy when I saw my rating drop. This caused me to play faster, and I paid this with a drop in accuracy, and after such a session I felt quite unhappy. I even got angry sometimes.

Then I realized that I still had the wrong view, despite my new focus on accuracy. I mean, I still regarded accuracy as a sort of money to buy as much rating as possible, trying to be more efficient, that is, paying less accuracy loss for more rating gain.

Yesterday I looked at it the other way round: Rating was the money now, and I was ready to spend it for more accuracy. When I gained two or three points in a good series, I said to myself, hello, here are some more points to SPEND. Earlier I always was anxious to LOSE them again.

And yeah, it worked: I had my best session ever, 96.5% in a session of 143 with a new personal record of 77 correct in a row, starting with 1442 and ending with 1444.

My next goal will be a blank session of more than a hundred.

6 Comments:

At 10:30 AM, Blogger likesforests said...

Spot - Sort - Focus - Decide ... I'll have to try that. When I loan rating points for accuracy they came back, eventually. It's only a loan.

 
At 9:42 PM, Blogger transformation said...

dear mouse three things:

i. thank you for your note. i did not know that you were near retirment. so i am not alone as not one of the kids, or wiz kids if you will?

i propose that you and i form a breakaway group called The Wobbly Knights, or the Geront Knights. as in Gerontology or study of old people. we can recruit aggressively, and pull members from among the original Knights Errant. perhaps even tempo or dutch might join, since they are also not absolute kids. blueDevel does not quality, but by special exemption, we could grandfather him in without illegal recourse, since he is not a totally young man, thus strengthening our roster of members.

we need rules. no one can join unless they accept our motto: false teeth for all Knight Geronts. strict adherance to loyalty to other Knight Geronts, and posting links to their site. reporting results of hair loss, pharacological mishap, and new marriages or births of grandchildren.

ii. i love this post. percentage, to me, is the key to better over the board chess. sometimes i am very upset to go below 1500 or rather well below 1500 but remind myself sometimes it is nice NOT to be 1500 so as to be able to focus more on accurate thinking. of what matter is raing here? this is just peer to peer comparison, but understand this upset you describe.

if i get under 93% now, i feel upset. it is all relative. at one time, i was glad for 80%, then 85%, now i am 91% on a second practice ID at cts, and feel like my percentage is TOO low.

this seventy in a row+ is fantastic. i try for this every day. many, many, MANNNY at cts say they care about rating but not percentage, but, let me assure you, not many of them could do forty in a row, or thirty. it is a lot harder than it looks, since even sure things can be ruined by a mouseslip (no pun intended), or timeout or disconnect, or lag, or misrecollection.

lets say this. we try for the highest percentage, WHILE giving strong consideration to the clock. not to watch rating and clock alone, but consider them WHILE seeking the highest percentage per session while rapid or not totally slow.

iii. if you have not already seen it, and probably well already have, view the top fifty at CTS for tries, and how many of them are over 80% or how many in the top 200 percentage at CTS are over 1450? not many. this says something. it is like seven or eight who are 1500 and over 84% both.

you post is wonderfull, and so much keen inteligence, makes coming to blogger a true delight.

take care, david

 
At 12:17 PM, Blogger wormstar said...

have you guys noticed FARKINDAMIYIZ yet? 815 tries at 1586 at an incredible 97.2%! that's 33 problems correct in a row on average!

 
At 3:42 PM, Blogger Mousetrapper said...

To likes: welcome to the percentage club. I see you at position 138, I am at 254, but on my way up, looking forward to meet you in the future.

To dk: hehehe, sort of grey panther fraction of the knights, I guess Tempo might also join us, looks like retired regarding the huge amouts of time he spends for CTS and blogging.

To worm: let this guy do 10k and we'll see where he is. Impressive, though.

 
At 10:29 PM, Blogger transformation said...

This comment has been removed by a blog administrator.

 
At 12:16 AM, Blogger transformation said...

percentage club indeed. there are still 359 who have done '80%' or greater [BTW, of course, really 340 since some of those 80's are actually 79.501>], but of those, only seven or eight who are 85% or greater and =>1500 [correction] over 10,000 or 15,000 tries.

i have tried to go for 95-97% which is of course realizable, but how to do it without a gross rating adjustment...

as i have said time and time again, this spring i was 78%, then 80%, then 82%, and got to where i finally identified 1f/7s=8 tries was acceptable to me. so for every 40, id get five wrong.

then i decided to go for 87% at 30,000 tries, so now i must do 92% for the next 10,000. the issue for me is not how to stay so accurate but how to stay that accurate without getting totally slow, or having a lower incidence of 40-50 sec problems recovered by brute force--i.e. 'i will solve this no matter what'.

but as you say, there are many, many other near immediate recognitions, and just as percentage is paid for with rating as you so well articulate, but these slower problems for me are payed for with the many other very quick ones that are 'slam dunks'.

now i am upset if i get one wrong in 15 tries! id love to alway be 1f/19s=20 and like you have long runs of tries, but then there are those times you get two in a row wrong with overconfidence.

then in a very different way, there are times i get in the zone and see the board, all threats, checks, captures, counterattacks, usually after a lot of practice at the end of a long session when also NOT tired, it just all comes together.

i do want 95% but not totally willing to take any speed to do it, so trying to balance it all. hence my current obscession to stay near 1500 but probably better off without it, and just sink my feet into the ground and hit them all.

i have a practice ID that is seperate, but lower tries, and even there at a lower rating, i find 95% hard to do SUSTAINED without even concern much for rating...

at a certain point, you feel that you know, so you move the piece or touch the square, but it is still wrong.

then the problems we ALWAYS get wrong. last night i finally had one of those and i did it right, after only like seven or eight iterations since march.

do we charge money or simply require folks be over age 40 and be committed to percentage tries? i say if we form a breakaway Knight group, we make an exemption for LikeForests despite his apparent youthfullness. is that ok with you? lets lay strategy before they get us first.

dk

 

Post a Comment

<< Home