sum of absolute errors, I think, is what you mean. Otherwise your fitness function should always return 0 ;-)
You may be better off using one of the statistical error calculations (Portal:Statistics). IIRC (and stats was not my strong suit 15+ years ago, so apply salt as necessary), some of the more complex (well, more complex than summation anyway) error functions would give "better" results.
--MidLifeXis
In reply to Re^3: Average Price Algorithm
by MidLifeXis
in thread Average Price Algorithm
by camelcom
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |