Trip rating

[JJMartin]

Is there such thing as a trip comment evaluation amongst the info providers that sums up the comments about the trip as either being positive or negative?

[shouldacoulda]

Bris has PPs with comments that I've used. I think they do more harm than good. You're better off drawing your own conclusions and watching replays.

[giovanni]

Last year I looked at all trip notes using an automated system to get the gist of the trip note. After characterizing the trip notes into similar meaning clusters, I then looked at the next out performance of each group. Many groups showed no difference, but a few showed consistent improvement or worsening of up to 2 positions at the finish line compared to the race with the trip note. Mind you, the improvement may have been from 9th place up to 7th place.

The odds of the trip notes groups showing improvement or worsening were discounted appropriately to the delta of finish position. The study showed the note to have no added wager value, as the crowd set the odds correctly and accounted for the good or bad trip note completely.

A recent post on R-Bloggers discusses analysis of E-Mails, which are of a much longer size than a tripnote, but might be interesting for further research.


http://www.r-bloggers.com/a-rather-nosy-topic-model-analysis-of-the-enron-email-corpus/?utm_source=feedburner&utm_medium=email&utm_campaign=Feed%3A+RBloggers+%28R+bloggers%29