Feedback from previous weeks is lost when re-using the same submissioin?

Insurance pricing game
#1

Hi!
I mostly reuse the same submission since week 1, except when I forget to lock it in and my “latest” submission gets reused by default.

I have noticed that older feedbacks appears to disappear. For example, when I view feedback from week 3 it actually links me to week 6 so I can’t get feedback from week 3 and #2… I think all three weeks used the 'latest submission".

Also, is there any way to lock the submission “until further notice” instead of “just for next Saturday” ?

thanks!

#2

Hi @simon_coulombe

When you lock a submission it is actually “until further notice” as you say. In fact if you leave the same submission locked in then over a few weeks this is what you see:

image
image728×306 35 KB

Having said that, feedback only appears as a link under your submission for weeks 4 onward. Weeks 1 - 3 were sent to your emails I believe :incoming_envelope:

In your case the feedback for each of the weeks 4, 5 and 6 are under your relevant submissions:

The reason that clicking on your week 3 submission links you back to your week 6 submission is because they are exactly the same submission and week 3 feedback was emailed to you.

Please let me know if that doesn’t clarify things and we can discuss :left_speech_bubble:

1 Like
#3

that makes sense, I guess I had “selected” but not “locked” my submission on week5.

That being said, I would love to get to explain my pricing strategy for week 6.

thanks @alfarzan

#4

Haha you can!

If you are not worried about giving away your strategy you can start a notebook and discuss it.

Alternatively if you would rather wait until the end of the challenge we will be hosting a “Town Hall” event where we will invite contributions from participants and even give people the chance to present and speak if they wish :slight_smile:

For that one stay tuned :headphones:

#5

I’ve been planning to release my submission as soon as it became irrelevant, but for some reason it still manages to grab a good spot on the profit leaderboard so I’ll wait a bit more :slight_smile:

Here’s a fun one: I created a “R starterpack” using {recipes} and {xgboost} late last year to help get people started. I did the absolute minimum to estimate an xgboost model and set the hyperparameters to something I thought would be crazy. Guess what, that is my best submission on the RMSE leaderboard by far, so I can’t even share that. I don’t trust it enough for the profit leaderboard, but that was very humbling :slight_smile:

1 Like