Sunday, October 25, 2015

Back in the saddle

At least somewhat.  My work schedule has gotten less crazy, although it's not ideal.  I was able to pick up my training back in September.  I started doing Chess Tempo tactics full time mid-November 2014.  Here's what my training looked like:

 Month - Problems (Non-duplicate problems)

November '14 - 1,717 (656) *
December '14 - 2,743 (630)
January - 1,934 (399)
February - 1,243 (138) **
March - 647 (0)
April - 339 (1)
May - 341 (1)
June - 48 (0)
July - 329 (4)
August - 162 (29)
September - 504 (118) ***
October - 2,215! (500)

* Half a month worth of training
** Start a new job 2nd week of February
*** One week worth of training


I go from doing problems almost every day, to problems every 3-4 days to just doing a big set of problems over the course of 3-4 days and nothing else the rest of the month.  You can see the big decline in non-duplicate problems right when I take a new job.  From that point on, any problems I'm doing is just keeping up with problems I've seen before.  Treading water if you will.

Since coming back in late September, it looks like I'm back to where I was when I was training regularly.  500-600 non-duplicates is equivalent to a good book's worth of problems.    That's a lot of tactics problems when you think about it.

Now for some interesting numbers.  I made several custom sets and tracked my stats weekly to look for progress/improvement.  The next numbers are where I left off in February and where I'm at now

All ratings are based on blitz mode.

Mate in 2 (1200-1600) - Feburary
95.5% accuracy
45.2% in under 15 seconds
3 seconds slower vs average solver

Mate in 2 (1200-1600) - October
96.8% accuracy
54.2% in under 15 seconds
2 seconds faster vs average solver

Decoys* (1200-1600) February
91.9% accuracy
34.6% in under 15 seconds
4 seconds slower vs average solver

Decoys (1200-1600) October
91.4% accuracy
34.8% in under 15 seconds
2 seconds slower vs average solver

Core** (1200-1600) February
93.2% accuracy
22.2% in under 15 seconds
16 seconds slower vs average solver

Core (1200-1600) October
93.6% accuracy
24% in under 15 seconds
10 seconds slower vs average solver

* Decoys = Attraction, Blocking, Clearance, Coercion, Interference
* Core = Capturing Defender, Distraction, Fork, Pin

I also have some mating motifs based on Yusupov's first book, but there aren't enough problems in each set to get any meaningful info from.  I just use them because I assume they're more important mating motifs than some of the others.

I was surprised to see that I'm generally faster and more accurate after a month of training compared to where I was back in February.

I haven't played any OTB games since August.  I'm expecting that to change soon.  Honestly, I'm still not sure if I feel motivated enough to start playing like I used to.