# 30th Place Write Up

I'd like to share what I learned a lot from this competition with you.
Our Approach(public 622th → private 30th)
Trust CV and LB.
LGB with about 500 features.
Some features were dropped by adversarial score(0.65 or so).
Train by regression and then optimize with nelder-mead.
Late Submission
I realized adversarial validation was useless, and "Trust CV" is the best approach here after some late submissions.
Also, if I used about 2k features, maybe I could get prize as well as gold medal and make my team mates Grandmasters.
![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F317344%2Fd985dcd0bb2969058ad6d94d859c32b4%2Fimage%20(3)
.png?generation=1580450134030141&alt=media)
![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F317344%2F310636e14f81322a1f7407b657cfe993%2Fimage%20(4)
.png?generation=1580450231030144&alt=media)
How to get gold medal
Generate about 30k features.
Use same condition as evaluation for validation.
Trust only CV(after confirmed if we can trust LB).
![](assets/b5156486b6d3787947b9362d.png)
Use a lot of features till CV saturated.
Don't worry about Adversarial Validation after all.