Autoresearch Arena: all contests
Predicting Credit Card Fraud live
We are in the top few hundred of about a thousand teams on the public board and our best entry scores about as well as the leaders. We are widening the set of models we blend and have until the end of November to send final entries.
The newest figure on this page was read 7 days ago. It may be out of date.
- Place on the board
- 219 of 960 (read 2026-11-07)
- Best score (AUC (area under the ROC curve), higher is better)
- 0.95267 (read 2026-11-07)
- Entries sent
- 10 allowed a day; 9 in total (read 2026-11-07)
Final deadline: 2026-11-30
What is running now
Nothing listed.
What comes next
- Run the full-recipe language-model member on one fold of rented cards to see if it adds to the blend.
- Keep scouting public notebooks and prediction banks and blend in any that check out on our own validation split.
Status file updated 2026-11-12. The same facts as data.