Strength

Blikskottel 0.7 represents another noticeable improvement in playing strength. The code significantly outperforms the previous release, and several kinks have been eliminated from the engine's knowledge. Although Blikskottel 0.7 is better than ever before, it is still quite naive knowledge wise and could be significantly improved.

Blikskottel 0.5, playing as Skottel on FICS, has performed consistently in the mid-1800s (ELO) at Blitz and Standard time controls, but over 2100 for lightning. These figures were achieved on a Mobile PIII 800 with 128 MB RAM. Version 0.7 should move Blikskottel into the 1900s.

Further significant strength improvements could be had by increasing the chess knowledge in Blikskottel's evaluation function, as well as improving on Blikskottel's speed. Probably further improvements could also result from more correct weighting of the eval terms. In addition, Blikskottel is noticeably lacking an openings book, which would make a material difference to Blikskottel's results. Also, a whole pile of optimisation and algorithm improvement (working hash tables!!) should really be applied. In short, Blikskottel could be significantly improved. Unfortunately I seem to have lost interest a bit; but maybe it'll come back one day.