###
Moving Along, more annoying problems

The current problem, for reference, is Advanced100-1.pgn #20. The tags are as follows:
[Annotator "T3H"]
[Black "Adamski"]
[Date "1989.??.??"]
[Event "Kopenhagen"]
[EventDate "1989.??.??"]
[FEN "1r6/4kp2/3n2p1/2KPp3/4P2p/2NR2P1/7P/8 b - - 0 1"]
[PlyCount "7"]
[Result "0-1"]
[Round "?"]
[SetUp "1"]
[Site "?"]
[White "Christensen"]
I scored 10% on this problem twice in a row, which is no good. For a quick understanding of my scoring system, there are four positions that you need to solve in this problem, according to the PGN. With these four positions, Imagine on the first move, you do not get it correct (3 guesses, all wrong). The next move, you do the same (3 guesses all wrong). On the next move, you guess it correct on the first try (1 guess, correct). Finally, on the last move, you guess wrong on all three guesses (3 guesses, all wrong).
Thus, my program would do this math:
First, number of correct answers:
0 + 0 + 1 + 0 = 1
Divided by number of guesses
3 + 3 + 1 + 3 = 10
which is 10% correct. Now, here's where it's really uncool, and it really hurts your score. If you have a problem next with three positions, and you get it right on every try on the first shot, getting 100%, your new overall score is not 55%. It is actually 30%, because each problem is weighted by the number of guesses. Therefore, the previous problem, with 10 guesses is weighted more heavily than the current problem, with 3 guesses.
Thus, it allows one to take their results table with a grain of salt, after understanding the weighting mechanism.

## 0 Comments:

Post a Comment

<< Home