Ticket #103 (closed enhancement: fixed)
owl tuning
| Reported by: | gunnar | Owned by: | gnugo |
|---|---|---|---|
| Priority: | normal | Milestone: | 3.7.9 |
| Component: | patterns | Version: | |
| Severity: | normal | Keywords: | |
| Cc: | patch: | yes |
Description (last modified by gunnar) (diff)
Owl tuning. Regression results:
neurogo:28 FAIL F14 [O2] slightly bad nicklas1:1901 PASS B7 [B7] slightly good niki:5 PASS S9 [S9|S10] good global:4 FAIL F5 [Q6] neutral global:5 FAIL F5 [O4] neutral strategy4:201 PASS D5 [D5] mostly good safety:3 PASS F13 [H15|H16|G14|F13|C17] good gunnar:30 PASS F18 [F18] slightly good 9x9:550 PASS C6 [C6] very good, primary target kgs:240 PASS B7 [B7] somewhat bad kgs:530 PASS P14 [P14] neutral 8 PASS 3 FAIL Total nodes: 1682783132 3317474 12608312 (+0.11% +0.13% +0.031%)
global:4,5 are both affected by a changed owl reading for E4. Both with and without patch the reading is almost completely clueless. It would be good if these testcases could be changed so that this faulty reading does not disturb the more interesting aspects.
Attachments
Regression Results
| Attachment | Rev. | PASS | FAIL | Nodes | Status | |
| gunnar_7_9.12.diff | never tested |
Change History
Note: See
TracTickets for help on using
tickets.
