Classify Math Publications (experimental)

Fully evaluated agents (requires 100 runs)

The current rating shows the rating of your agent according to the more recent runs. The best rating is the highest value that the current rating column ever had, and it is the only rating that actually matters in the end (e.g. for grading). In other words, you do not have to worry if current rating is substantially lower than best rating as long as your code is able to produce a rating similar to best rating.

Agent Best rating Current rating
Agent384 0.7414000000000001 0.732
Agent397 0.7406 0.728
Agent950 0.7243999999999999 0.7134
Agent441 0.7136 0.7116
Agent093 0.7123999999999999 0.71
Agent708 0.6886 0.6626000000000001
Agent336 0.6808 0.6628000000000001
Agent925 0.6798000000000001 0.6668000000000001
Agent890 0.6602 0.6408
Agent519 0.66 0.6384000000000001
Agent202 0.6568 0.6459999999999999
Agent787 0.655 0.6382
Agent054 0.4166 0.396
Agent881 0.23920000000000002 0.2284
Agent389 0.0674 0.0632
Agent494 0.0656 0.0636

Not fully evaluated agents

The following agents have not had enough runs to be considered fully evaluated.

Agent Number of runs Estimated rating
TestAgent 15 0.12533333333333335

Note that agents that had no runs might not be listed.