# Give more than 1 chance for submission

**URL:** <https://community.drivendata.org/t/give-more-than-1-chance-for-submission/4565>\
**Category:** Hateful Memes\
**Created:** [May 31, 2020, 12:16pm UTC](https://community.drivendata.org/t/give-more-than-1-chance-for-submission/4565 "2020-05-31T12:16:15Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![pritish\_mishra](https://avatars.discourse-cdn.com/v4/letter/p/b77776/32.png) [@pritish\_mishra](https://community.drivendata.org/u/pritish_mishra)\
**Post date:** [May 31, 2020, 12:16pm UTC](https://community.drivendata.org/t/give-more-than-1-chance-for-submission/4565/1 "2020-05-31T12:16:15Z")

</div>

Please give more then 1 chance per day for submission…

---

<div class="post-metadata">

**Author:** ![adalseno](https://avatars.discourse-cdn.com/v4/letter/a/e47c2d/32.png) [@adalseno](https://community.drivendata.org/u/adalseno)\
**Post date:** [June 1, 2020, 10:41am UTC](https://community.drivendata.org/t/give-more-than-1-chance-for-submission/4565/2 "2020-06-01T10:41:55Z")

</div>

You actually are allowed up to 3 submission per day

---

<div class="post-metadata">

**Author:** ![VictorCallejas](https://avatars.discourse-cdn.com/v4/letter/v/8491ac/32.png) [@VictorCallejas](https://community.drivendata.org/u/VictorCallejas)\
**Post date:** [June 6, 2020, 6:43pm UTC](https://community.drivendata.org/t/give-more-than-1-chance-for-submission/4565/3 "2020-06-06T18:43:34Z")

</div>

It is one by day.

And if you are in a team, it is one per day per team

---

<div class="post-metadata">

**Author:** ![VictorCallejas](https://avatars.discourse-cdn.com/v4/letter/v/8491ac/32.png) [@VictorCallejas](https://community.drivendata.org/u/VictorCallejas)\
**Post date:** [June 6, 2020, 6:48pm UTC](https://community.drivendata.org/t/give-more-than-1-chance-for-submission/4565/4 "2020-06-06T18:48:11Z")

</div>

By the way, in a 5 months competition I think one by day is good.

If you think validation set does not represent well the test set, one option is to join train and validation and use k-fold validation or your own split.

With more submissions on test set you will not create a better model, you will just overfit to test set. And the second phase has a different test set, keep it in mind.

---

<div class="post-metadata">

**Author:** ![pritish\_mishra](https://avatars.discourse-cdn.com/v4/letter/p/b77776/32.png) [@pritish\_mishra](https://community.drivendata.org/u/pritish_mishra)\
**Post date:** [August 9, 2020, 7:11am UTC](https://community.drivendata.org/t/give-more-than-1-chance-for-submission/4565/5 "2020-08-09T07:11:52Z")

</div>

Okay… I understood