5 ms·
Surprised at the small size of the dataset here. Surely it’s possible to record thousands if not millions of results… has someone been able to do this? Seems li
by thom 2y ago
Surprised at the small size of the dataset here. Surely it’s possible to record thousands if not millions of results… has someone been able to do this? Seems like the simplest thing to do here would just be to develop some sort of plus/minus system and not try to overthink things. I work in sports analytics, and mostly in low scoring sports so we don’t get to do that very often.
- kqr 2y ago> Surprised at the small size of the dataset here. Author here. For answering the research question that spawned the article (engage or stay away from a do-or-die encounter), the data set actually appears to be "too large" – see the significance values of the parameters. I got carried away when copy-pasting match result tables, but the effect is so clear that a much smaller data set would have sufficed for showing it. I think the thing you're asking for is what the Leetify rating tries to encode, but people in this discussion have suggested alternative rating systems which I have yet to look into.