{ "id": 6702, "title": "Extend competition again?", "url": "/competitions/flight2-main/discussion/6702", "author_name": "", "post_date": "2013-12-28T14:10:44.547Z", "votes": null, "comment_count": 8, "views": 4681, "content": "

Hi Admins,

\n

It has been suggested by Gabriele Seppi and DerDasMachenMuss to extend the competion again.

\n

http://www.gequest.com/c/flight2-main/forums/t/6648/third-data-release-now-available/36732#post36732

\n

and 

\n

http://www.gequest.com/c/flight2-main/forums/t/6648/third-data-release-now-available/36736#post36736

\n

Allthough I absolutely agree with Gabriele Seppi and DerDasMachenMuss that the issues with training3 dataset need to be resolved as soon as possible, I don't think it is a good idea to extend the competion again. It seems to me a bit unfair for the people currently high on the LB. For myself it could be advantageous to have more time but I simply want to go on with my life after jan 18th. That having said it would be good for everybody I think to have some clarity on this matter.

\n

If another extention of the deadline is to be considered, could we have a vote on this?

\n

Admins could you please tell us if another extension is, or will be, under consideration?

\n

Thanks.

\n

 

\n

 

", "messages": [ { "id": "36752", "postDate": "12/28/2013 14:10:44", "content": "

Hi Admins,

\n

It has been suggested by Gabriele Seppi and DerDasMachenMuss to extend the competion again.

\n

http://www.gequest.com/c/flight2-main/forums/t/6648/third-data-release-now-available/36732#post36732

\n

and 

\n

http://www.gequest.com/c/flight2-main/forums/t/6648/third-data-release-now-available/36736#post36736

\n

Allthough I absolutely agree with Gabriele Seppi and DerDasMachenMuss that the issues with training3 dataset need to be resolved as soon as possible, I don't think it is a good idea to extend the competion again. It seems to me a bit unfair for the people currently high on the LB. For myself it could be advantageous to have more time but I simply want to go on with my life after jan 18th. That having said it would be good for everybody I think to have some clarity on this matter.

\n

If another extention of the deadline is to be considered, could we have a vote on this?

\n

Admins could you please tell us if another extension is, or will be, under consideration?

\n

Thanks.

\n

 

\n

 

", "rawMarkdown": "", "votes": null }, { "id": "36758", "postDate": "12/28/2013 14:59:52", "content": "

[quote=Jules;36752][...] I simply want to go on with my life after jan 18th[/quote]

\n

 

\n

You have no idea how much I'm looking forward to that day actually :D

\n

[quote=Jules;36752]If another extention of the deadline is to be considered, could we have a vote on this?[/quote]

\n

I totally agree on this. Last time it felt a bit forced on us all.

", "rawMarkdown": "", "votes": null }, { "id": "36909", "postDate": "01/02/2014 10:53:46", "content": "

Since we started de facto to vote on another extension of the deadline:

\n

https://www.gequest.com/c/flight2-main/forums/t/6648/third-data-release-now-available/36902#post36902

\n

Score so far:

\n

3 in favour of extension

\n

1 against

\n

I call on all participants of FQ2 to vote on this.

\n

 

", "rawMarkdown": "", "votes": null }, { "id": "36925", "postDate": "01/02/2014 19:49:37", "content": "

If Kaggle changed the timeline significantly again, ~2 weeks before the deadline, I would feel this as an incorrect behavior. Even the first deadline extension was strange in my opinion, since Kaggle did not disclose any strong reason why it was needed.

\n

I think that a small extension due to the data release problems is unnecessary either, but this is only one opinion. Anyway, voting in this question is problematic, since not surprisingly, the votes correlate with the actual standings (just take me as an example ;-).

\n

Whatever will be the decision, please let us know soon.

", "rawMarkdown": "", "votes": null }, { "id": "36941", "postDate": "01/03/2014 08:58:59", "content": "

Hi Gabor,

\n

I couldn't agree more! ;-)

\n

However I probably had an advantage with the first extension so I probably have even less rights to say anything...

\n

But there doesn't seem to be a valid reason to extend this time as there is no change to the important parts of the competition (like the simulator or allowed data to use).

\n

However it is most important to know whether or not the competition is extended really soon as it has a major impact on the next steps of our to do lists (start finalizing and tuning the last bits or start working on some new ideas that potentially take weeks to complete).

", "rawMarkdown": "", "votes": null }, { "id": "36942", "postDate": "01/03/2014 10:08:11", "content": "

[quote=Willem Mestrom;36941]

\n

But there doesn't seem to be a valid reason to extend this time as there is no change to the important parts of the competition (like the simulator or allowed data to use).

\n

[/quote]

\n

Well, I spent weeks working on the allowed datasets and trying to understand why what I was doing wasn't helping me improve (actually, my results were worse). Then, as you can see in other topics, we discovered there were problems with such datasets (which are still being analyzed and corrected by Kaggle as we write).

\n

Considering my position on the leaderboard now and the one I had before the first extension, well, my position got worse. I agree that another extension is unfair, especially so close to the final deadline, but there may be other people in my same situation that may feel like not extending the deadline in unfair to them.

\n

I think that the real problem started, as Gabor hinted after the first extension, when Kaggle struggled to meet deadline even before starting (if I remember correctly FQ2 was delayed one month). This is a complex competition and many problems arose: unfortunately these problems influenced results, deadlines and people's work. I don't think there will be a solution that can make everyone happy.

\n

Too bad, it had potential for a fun competition, but the more time passes, the more I feel sad by how things evolved. And no, my feelings (as of now, at least, since there is still time) are not driven by frustration over scores or leaderboard position...

", "rawMarkdown": "", "votes": null }, { "id": "36951", "postDate": "01/03/2014 14:07:12", "content": "

Hi Everyone,

\n

The timeline currently stated on the competition pages will not be changed. Sorry for either raising hopes or blood pressures! 

", "rawMarkdown": "", "votes": null }, { "id": "36957", "postDate": "01/03/2014 18:31:38", "content": "

When Kaggle has extended the deadline by a month, I have actually taken a month off, as there was too much stress.

\n

This competition was delayed twice at launch, and once again in November. I am happy to hear there will be no more delays.

", "rawMarkdown": "", "votes": null }, { "id": "36962", "postDate": "01/03/2014 20:24:55", "content": "

[quote=charango;36957]

\n

When Kaggle has extended the deadline by a month, I have actually taken a month off, as there was too much stress.

\n

[/quote]

\n

But it has been a great deal for the sponsor who is getting extra work done, presumably without spending any extra money. The scores improved quite a bit after the original deadline. By extension, I guess it has been a great deal for Kaggle too.

\n

On the other hand, Kaggle never fixed the problem with the ascent model in the simulator, so there is a good chance that they will end up with solutions that don't translate well to the real world (just like what happened with the milestone winning solution).

", "rawMarkdown": "", "votes": null } ], "comments": [ { "id": 36758, "author_name": "gabrieleseppi", "author_url": "", "post_date": "12/28/2013 14:59:52", "content": "

[quote=Jules;36752][...] I simply want to go on with my life after jan 18th[/quote]

\n

 

\n

You have no idea how much I'm looking forward to that day actually :D

\n

[quote=Jules;36752]If another extention of the deadline is to be considered, could we have a vote on this?[/quote]

\n

I totally agree on this. Last time it felt a bit forced on us all.

", "votes": null, "replies": [] }, { "id": 36909, "author_name": "julesvanligtenberg", "author_url": "", "post_date": "01/02/2014 10:53:46", "content": "

Since we started de facto to vote on another extension of the deadline:

\n

https://www.gequest.com/c/flight2-main/forums/t/6648/third-data-release-now-available/36902#post36902

\n

Score so far:

\n

3 in favour of extension

\n

1 against

\n

I call on all participants of FQ2 to vote on this.

\n

 

", "votes": null, "replies": [] }, { "id": 36925, "author_name": "gtakacs", "author_url": "", "post_date": "01/02/2014 19:49:37", "content": "

If Kaggle changed the timeline significantly again, ~2 weeks before the deadline, I would feel this as an incorrect behavior. Even the first deadline extension was strange in my opinion, since Kaggle did not disclose any strong reason why it was needed.

\n

I think that a small extension due to the data release problems is unnecessary either, but this is only one opinion. Anyway, voting in this question is problematic, since not surprisingly, the votes correlate with the actual standings (just take me as an example ;-).

\n

Whatever will be the decision, please let us know soon.

", "votes": null, "replies": [] }, { "id": 36941, "author_name": "wmestrom", "author_url": "", "post_date": "01/03/2014 08:58:59", "content": "

Hi Gabor,

\n

I couldn't agree more! ;-)

\n

However I probably had an advantage with the first extension so I probably have even less rights to say anything...

\n

But there doesn't seem to be a valid reason to extend this time as there is no change to the important parts of the competition (like the simulator or allowed data to use).

\n

However it is most important to know whether or not the competition is extended really soon as it has a major impact on the next steps of our to do lists (start finalizing and tuning the last bits or start working on some new ideas that potentially take weeks to complete).

", "votes": null, "replies": [] }, { "id": 36942, "author_name": "gabrieleseppi", "author_url": "", "post_date": "01/03/2014 10:08:11", "content": "

[quote=Willem Mestrom;36941]

\n

But there doesn't seem to be a valid reason to extend this time as there is no change to the important parts of the competition (like the simulator or allowed data to use).

\n

[/quote]

\n

Well, I spent weeks working on the allowed datasets and trying to understand why what I was doing wasn't helping me improve (actually, my results were worse). Then, as you can see in other topics, we discovered there were problems with such datasets (which are still being analyzed and corrected by Kaggle as we write).

\n

Considering my position on the leaderboard now and the one I had before the first extension, well, my position got worse. I agree that another extension is unfair, especially so close to the final deadline, but there may be other people in my same situation that may feel like not extending the deadline in unfair to them.

\n

I think that the real problem started, as Gabor hinted after the first extension, when Kaggle struggled to meet deadline even before starting (if I remember correctly FQ2 was delayed one month). This is a complex competition and many problems arose: unfortunately these problems influenced results, deadlines and people's work. I don't think there will be a solution that can make everyone happy.

\n

Too bad, it had potential for a fun competition, but the more time passes, the more I feel sad by how things evolved. And no, my feelings (as of now, at least, since there is still time) are not driven by frustration over scores or leaderboard position...

", "votes": null, "replies": [] }, { "id": 36951, "author_name": "joycenv", "author_url": "", "post_date": "01/03/2014 14:07:12", "content": "

Hi Everyone,

\n

The timeline currently stated on the competition pages will not be changed. Sorry for either raising hopes or blood pressures! 

", "votes": null, "replies": [] }, { "id": 36957, "author_name": "sergeykozub", "author_url": "", "post_date": "01/03/2014 18:31:38", "content": "

When Kaggle has extended the deadline by a month, I have actually taken a month off, as there was too much stress.

\n

This competition was delayed twice at launch, and once again in November. I am happy to hear there will be no more delays.

", "votes": null, "replies": [] }, { "id": 36962, "author_name": "anlthms", "author_url": "", "post_date": "01/03/2014 20:24:55", "content": "

[quote=charango;36957]

\n

When Kaggle has extended the deadline by a month, I have actually taken a month off, as there was too much stress.

\n

[/quote]

\n

But it has been a great deal for the sponsor who is getting extra work done, presumably without spending any extra money. The scores improved quite a bit after the original deadline. By extension, I guess it has been a great deal for Kaggle too.

\n

On the other hand, Kaggle never fixed the problem with the ascent model in the simulator, so there is a good chance that they will end up with solutions that don't translate well to the real world (just like what happened with the milestone winning solution).

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "36752": "", "36758": "", "36909": "", "36925": "", "36941": "", "36942": "", "36951": "", "36957": "", "36962": "" }, "source": "meta" }