# Problems A-F

Problems A-F

## Performance of POB and Other Algorithms on Problems A-F

We measure the performance of POB and partial order pruning by comparing four kinds of search:Plain alpha-beta (abbreviated ab in the table) uses partial ordering for move ordering only, and evaluates wins and losses only by capture. Alpha-beta search with partial order pruning ( abPOP) uses knowledge about partial ordering to prune moves. Non-pruning POB (POB-) uses partial order leaf evaluation but no pruning. Full POB uses both partial order leaf evaluation and partial order pruning. However, all four types of search apply the pruning rule of using only one of several equivalent moves, such as outside liberties. Without this rule, the non-pruning variants would be much worse.
 BW POB POB- abPOP ab B 1 1 20 50 A W 1 1 20 50 B B 1 1 210 441 B W 1 1 250 358 C B 20 26 23 33 C W 6 21 12 21 D B 18 27 433 1443 D W 18 27 134 284 E B 304 333 1339 1523 E W 208 333 993 1367 F B 18 27 1066 4199 F W 18 27 272 662
Comparison of POB and alpha-beta