3/1 UDPATE: Rules violators have officially been removed from the leaderboard, and the private leaderboard is now final. If you are no longer on the leaderboard, it is because you were removed, likely due to holding multiple accounts or identified collusion with other team(s). Evidence supporting removal is largely conclusive, but you are welcome to submit a compliance appeal if you believe your team was wrongfully removed.
\n\nThe leaderboard has been updated and the private leaderboard revealed. Note: The final standings are pending verification by our standard algorithms to remove multiple accounts violators, hence the disclaimer message on the leaderboard. Due to Kaggle's scheduled downtime, this final leaderboard verification may take a few days to post.
\n\nIn the meantime, a tentative congratulations to the winners!
\n\nA huge thank you to everyone for your participation. We recognize there were a number of unexpected issues that plagued this competition, and I'm sorry for the impact this may have had on your experience. This was atypical for Kaggle, and we tried our very best to sustain the competition so that the work you put in could be recognized.
\n\nPlease note the following:
\n\nAs described when the competition's design was revised, the private leaderboard is now based on future-looking predictions from the Stage 2 timeframe of Dec 1, 2018 to Jan 31, 2019.
The public leaderboard now contains no scores, as a result of bullet #1.
As is Kaggle policy for 2-stage competitions, the team count for points and medals calculation is based on original team volumes and has been adjusted as such.
We are still working on the leaderboard removal of rules violators. Due to downtime and migration activities over the weekend, we had to delay these finalization activities. Now that some migration dust has settled, we expect the finalization should be completed in the next day or two. Likewise, we also waited to adjust the team count for medals and points calculations. Do not fret, we have every intention of closing out these activities as quickly as possible.
", "rawMarkdown": "We are still working on the leaderboard removal of rules violators. Due to downtime and migration activities over the weekend, we had to delay these finalization activities. Now that some migration dust has settled, we expect the finalization should be completed in the next day or two. Likewise, we also waited to adjust the team count for medals and points calculations. Do not fret, we have every intention of closing out these activities as quickly as possible.", "votes": 16, "replies": [ { "id": 481765, "postDate": "2019-03-01T20:19:02.187Z", "content": "Hi Julie, thank you for the efforts the Kaggle team has made to sustain this competition. Just wondering if there are any updates on the final process?
\n\nThanks,
", "rawMarkdown": "Hi Julie, thank you for the efforts the Kaggle team has made to sustain this competition. Just wondering if there are any updates on the final process?\n\nThanks,", "votes": 2 }, { "id": 481828, "postDate": "2019-03-01T22:18:37.510Z", "content": "Apologies for the delay. We had to overcome a few post-migration issues with our algorithm to remove rules violators. Please see the 3/1 update in my original post above. The leaderboard is now final.
", "rawMarkdown": "Apologies for the delay. We had to overcome a few post-migration issues with our algorithm to remove rules violators. Please see the 3/1 update in my original post above. The leaderboard is now final." } ] }, { "id": 476373, "postDate": "2019-02-22T03:28:46.550Z", "content": "3/1 UDPATE: Rules violators have officially been removed from the leaderboard, and the private leaderboard is now final. If you are no longer on the leaderboard, it is because you were removed, likely due to holding multiple accounts or identified collusion with other team(s). Evidence supporting removal is largely conclusive, but you are welcome to submit a compliance appeal if you believe your team was wrongfully removed.
\n\nThe leaderboard has been updated and the private leaderboard revealed. Note: The final standings are pending verification by our standard algorithms to remove multiple accounts violators, hence the disclaimer message on the leaderboard. Due to Kaggle's scheduled downtime, this final leaderboard verification may take a few days to post.
\n\nIn the meantime, a tentative congratulations to the winners!
\n\nA huge thank you to everyone for your participation. We recognize there were a number of unexpected issues that plagued this competition, and I'm sorry for the impact this may have had on your experience. This was atypical for Kaggle, and we tried our very best to sustain the competition so that the work you put in could be recognized.
\n\nPlease note the following:
\n\nAs described when the competition's design was revised, the private leaderboard is now based on future-looking predictions from the Stage 2 timeframe of Dec 1, 2018 to Jan 31, 2019.
The public leaderboard now contains no scores, as a result of bullet #1.
As is Kaggle policy for 2-stage competitions, the team count for points and medals calculation is based on original team volumes and has been adjusted as such.
@juliaelliott\nHey Julia - many thanks for an update\nI wonder could the medals and score be calculated based on the original number of participants - I would think that if you have joined and then dropped because of some perturbations in competition - you still have already joined
", "rawMarkdown": "@juliaelliott\nHey Julia - many thanks for an update\nI wonder could the medals and score be calculated based on the original number of participants - I would think that if you have joined and then dropped because of some perturbations in competition - you still have already joined ", "votes": 17, "replies": [ { "id": 476607, "postDate": "2019-02-22T11:58:17.413Z", "content": "I agree with you @ValeriyBabushkin. On Quora Insincere competition, More than 4000 people joined and only 1400 of them made submissions. The medals were given according to the number of 4000.
", "rawMarkdown": "I agree with you @ValeriyBabushkin. On Quora Insincere competition, More than 4000 people joined and only 1400 of them made submissions. The medals were given according to the number of 4000.", "votes": 17 }, { "id": 476817, "postDate": "2019-02-23T08:35:12.867Z", "content": "Absolutely agree with Valeriy, that would be honest to count all joined participants in the final standings.
", "rawMarkdown": "Absolutely agree with Valeriy, that would be honest to count all joined participants in the final standings. ", "votes": 14 }, { "id": 476839, "postDate": "2019-02-23T09:48:33.930Z", "content": "Also, I agree with @ValeriyBabushkin!\nIt would be fair to evaluate rank (%) per all participants who joined the competition because the data and task strongly haven't changed. Probably, participants who submitted before the competition changes also should be evaluated in the total rank(%)
", "rawMarkdown": "Also, I agree with @ValeriyBabushkin!\nIt would be fair to evaluate rank (%) per all participants who joined the competition because the data and task strongly haven't changed. Probably, participants who submitted before the competition changes also should be evaluated in the total rank(%)", "votes": 14 }, { "id": 476867, "postDate": "2019-02-23T10:58:06.407Z", "content": "I also agree with @ValeriyBabushkin
", "rawMarkdown": "I also agree with @ValeriyBabushkin", "votes": 12 }, { "id": 476908, "postDate": "2019-02-23T12:43:16.247Z", "content": "Agree with @ValeriyBabushkin
", "rawMarkdown": "Agree with @ValeriyBabushkin", "votes": 13 }, { "id": 477224, "postDate": "2019-02-24T06:41:56.700Z", "content": "Also Agree with @ValeriyBabushkin
", "rawMarkdown": "Also Agree with @ValeriyBabushkin", "votes": 12 }, { "id": 477297, "postDate": "2019-02-24T10:03:36.783Z", "content": "This will be an additional motivation to participate in the competitions after rerun or important changes. Agree with @ValeriyBabushkin
", "rawMarkdown": "This will be an additional motivation to participate in the competitions after rerun or important changes. Agree with [@ValeriyBabushkin](https://www.kaggle.com/venheads) ", "votes": 13 }, { "id": 477770, "postDate": "2019-02-25T08:34:07.877Z", "content": "Can not agree more
", "rawMarkdown": "Can not agree more", "votes": 3 }, { "id": 477838, "postDate": "2019-02-25T10:49:09.953Z", "content": "I can also agree with @ValeriyBabushkin. The competition deadline was not postponed after the changes, so it would be fair not to reduce the number of participants as well and compute ratings & medals based on the total number teams including the ones that submited before the changes. I don't think it would be possible to rescore old submissions because the data is now different, but it would still be more honest for the people that continued to work on this competition during the second stage.
", "rawMarkdown": "I can also agree with @ValeriyBabushkin. The competition deadline was not postponed after the changes, so it would be fair not to reduce the number of participants as well and compute ratings & medals based on the total number teams including the ones that submited before the changes. I don't think it would be possible to rescore old submissions because the data is now different, but it would still be more honest for the people that continued to work on this competition during the second stage.", "votes": 6 } ] }, { "id": 477073, "postDate": "2019-02-23T20:04:27.957Z", "content": "I also agree with @ValeriyBabushkin. I noticed there is (now) no disclaimer message on the leaderboard, points and medals are not yet awarded it seems?
\n\nedit: disclaimer is only visible for me if I return to the LB via the back button on the browser
", "rawMarkdown": "I also agree with @ValeriyBabushkin. I noticed there is (now) no disclaimer message on the leaderboard, points and medals are not yet awarded it seems?\n\nedit: disclaimer is only visible for me if I return to the LB via the back button on the browser", "votes": 10 }, { "id": 478091, "postDate": "2019-02-25T18:01:14.103Z", "content": "Guys, have a look at the leaderboard. Not sure if it will stay like this but in this moment there are a lot more medal positions... if it's not a bug there is a chance that suggestion was accepted :-)
", "rawMarkdown": "Guys, have a look at the leaderboard. Not sure if it will stay like this but in this moment there are a lot more medal positions... if it's not a bug there is a chance that suggestion was accepted :-)", "votes": 4, "replies": [ { "id": 478097, "postDate": "2019-02-25T18:22:31.523Z", "content": "Which means all 0 sample submission have the chance to get a silver medal. Pretty sad for this kind of situation happened again and again.
", "rawMarkdown": "Which means all 0 sample submission have the chance to get a silver medal. Pretty sad for this kind of situation happened again and again. ", "votes": 9 }, { "id": 478187, "postDate": "2019-02-25T21:33:21.440Z", "content": "Unfortunately, the case you would not hope happens again. But lottery tickets are free. :P
", "rawMarkdown": "Unfortunately, the case you would not hope happens again. But lottery tickets are free. :P\n" }, { "id": 478193, "postDate": "2019-02-25T21:45:49.050Z", "content": "You could use a quote.
\n\nAll these flawed competitions just remind me more of the importance of clean data and meaningful evaluation metric for my projects at work.
", "rawMarkdown": "You could use a [quote](https://www.kaggle.com/c/ga-customer-revenue-prediction/discussion/73091#430209). \n\nAll these flawed competitions just remind me more of the importance of clean data and meaningful evaluation metric for my projects at work. " }, { "id": 478381, "postDate": "2019-02-26T04:55:24.680Z", "content": "What is the error for all 0 submission? In the future, there better be some cutoff of winner or not based on some pre-defined naive submission. Otherwise, it is just hilarious.
", "rawMarkdown": "What is the error for all 0 submission? In the future, there better be some cutoff of winner or not based on some pre-defined naive submission. Otherwise, it is just hilarious.", "votes": 1 }, { "id": 478428, "postDate": "2019-02-26T06:14:19.340Z", "content": "All 0 sub silver-medal is too silly. It's unfair for people who tackle other difficult and well-designed competitions , and this is disparaging the value of silver medals.\nThis competition was \"relaunched\". Medals should not be distributed based on too many teams.\n@juliaelliott, Please re-think and don't disappoint me (and maybe other kagglers) any further.
", "rawMarkdown": "All 0 sub silver-medal is too silly. It's unfair for people who tackle other difficult and well-designed competitions , and this is disparaging the value of silver medals.\nThis competition was \"relaunched\". Medals should not be distributed based on too many teams.\n@juliaelliott, Please re-think and don't disappoint me (and maybe other kagglers) any further.", "votes": 5 }, { "id": 478490, "postDate": "2019-02-26T08:14:02.987Z", "content": "It would not be first time, not even the second for sample submission get a silver in a competition. \nThe difference now -I think- is in other cases was an issue of the competition design/data and this time it would be a measure of how sorry Kaggle team is for all the problems in this one (clearly a reboot is not a two stage). \nPersonally, I decided not to overthink it much when I saw the \"happiness for all\" rain of medals (you are all comenting on silver but bronze used to be something meaningful too).
", "rawMarkdown": "It would not be first time, not even the second for sample submission get a silver in a competition. \nThe difference now -I think- is in other cases was an issue of the competition design/data and this time it would be a measure of how sorry Kaggle team is for all the problems in this one (clearly a reboot is not a two stage). \nPersonally, I decided not to overthink it much when I saw the \"happiness for all\" rain of medals (you are all comenting on silver but bronze used to be something meaningful too).\n\n\n", "votes": 4 }, { "id": 480903, "postDate": "2019-02-28T20:41:30.757Z", "content": "Agree with @NotaAaTuner! Looks like the naive 0 submission without any efforts could be awarded a silver medal, which is unfair for people taking efforts on \"real\" model. @juliaelliott, Please re-think and don't disappoint kagglers taking efforts in this competition! thanks
", "rawMarkdown": "Agree with @NotaAaTuner! Looks like the naive 0 submission without any efforts could be awarded a silver medal, which is unfair for people taking efforts on \"real\" model. @juliaelliott, Please re-think and don't disappoint kagglers taking efforts in this competition! thanks", "votes": 5 }, { "id": 481010, "postDate": "2019-03-01T00:59:01.637Z", "content": "Agree with all above comments. Competition reboot is not the same as a 2-stage competition. And all-zero submissions getting silver really devalues all medals awarded by this competition (not only silver and bronze) and effort/time that people actually spent in this competition.
", "rawMarkdown": "Agree with all above comments. Competition reboot is not the same as a 2-stage competition. And all-zero submissions getting silver really devalues all medals awarded by this competition (not only silver and bronze) and effort/time that people actually spent in this competition." }, { "id": 481749, "postDate": "2019-03-01T19:41:43.193Z", "rawMarkdown": "", "votes": -1, "isDeleted": true } ] }, { "id": 481052, "postDate": "2019-03-01T02:15:17.137Z", "content": "@juliaelliott\nso I think the best strategy is that try to add a function in your standard algorithm to filter out the submission with all 0.
", "rawMarkdown": "@juliaelliott\nso I think the best strategy is that try to add a function in your standard algorithm to filter out the submission with all 0. ", "votes": 1, "replies": [ { "id": 481061, "postDate": "2019-03-01T02:33:17.627Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 478062, "postDate": "2019-02-25T17:20:11.213Z", "content": "Agree with @ValeriyBabushkin that it would be fairer to rank based on the number of original participants.
", "rawMarkdown": "Agree with @ValeriyBabushkin that it would be fairer to rank based on the number of original participants. ", "votes": 1 }, { "id": 477335, "postDate": "2019-02-24T10:58:24.073Z", "content": "can we get the right predicted values' file?
", "rawMarkdown": "can we get the right predicted values' file?", "votes": 1 }, { "id": 476511, "postDate": "2019-02-22T09:09:42.907Z", "content": "Thanks all your effort on this competition!!
", "rawMarkdown": "Thanks all your effort on this competition!!", "votes": 1 }, { "id": 478026, "postDate": "2019-02-25T16:17:58.473Z", "content": "Hey @juliaelliott can you please update when the points and the medals will be awarded?
", "rawMarkdown": "Hey @juliaelliott can you please update when the points and the medals will be awarded? ", "votes": 2, "replies": [ { "id": 481829, "postDate": "2019-03-01T22:19:37.167Z", "content": "Apologies for the delay. We had to overcome a few post-migration issues with our algorithm to remove rules violators. Please see the 3/1 update in my original post above. The leaderboard is now final. Medals & Points will be updated within the next 24 hours.
", "rawMarkdown": "Apologies for the delay. We had to overcome a few post-migration issues with our algorithm to remove rules violators. Please see the 3/1 update in my original post above. The leaderboard is now final. Medals & Points will be updated within the next 24 hours.", "votes": 1 } ] }, { "id": 476476, "postDate": "2019-02-22T07:57:01.997Z", "content": "Thank you for your efforts to enable this competition to continue.
", "rawMarkdown": "Thank you for your efforts to enable this competition to continue. " }, { "id": 478037, "postDate": "2019-02-25T16:31:07.223Z", "content": "I think Kaggle should award more people who continued to compete in this competition by caring the number of ALL people joining this competition. This competition started and a lot of abnormalities happened in this competition:
\n\n1) The metric of competition was chosen as log of sum but actually it was sum of log or vice versa(I don't remember exactly).
\n\n2) The leakage was found and revealed.
\n\n3) After the founding of leakage, people had to wait 3 weeks for the competition to be re-arranged by Kaggle Team including @juliaelliott .
\n\n4) Competitors spent just 3 weeks on this competition(compared to normal competitions, it was a short time).
", "rawMarkdown": "I think Kaggle should award more people who continued to compete in this competition by caring the number of ALL people joining this competition. This competition started and a lot of abnormalities happened in this competition:\n\n1) The metric of competition was chosen as log of sum but actually it was sum of log or vice versa(I don't remember exactly).\n\n2) The leakage was found and revealed.\n\n3) After the founding of leakage, people had to wait 3 weeks for the competition to be re-arranged by Kaggle Team including @juliaelliott .\n\n4) Competitors spent just 3 weeks on this competition(compared to normal competitions, it was a short time).", "votes": 1 }, { "id": 481801, "postDate": "2019-03-01T21:50:01.250Z", "content": "Hi, @juliaelliott why my submission is disappear???????
", "rawMarkdown": "Hi, @juliaelliott why my submission is disappear???????", "votes": -1, "replies": [ { "id": 481824, "postDate": "2019-03-01T22:16:34.907Z", "content": "See 3/1 update made to original post.
", "rawMarkdown": "See 3/1 update made to original post." } ] }, { "id": 478354, "postDate": "2019-02-26T04:02:35.967Z", "content": "\n\nJust to confirm...
\n\nCalculation of metrics is log of sum or sum of log？
\n\nI understood it as log of sum. However, the score was better for sum of log.\nAlthough there are differences in random seed between the two models, I was curious.
", "rawMarkdown": "Hi @juliaelliott\n\n\nJust to confirm...\n\nCalculation of metrics is log of sum or sum of log？\n\nI understood it as log of sum. However, the score was better for sum of log.\nAlthough there are differences in random seed between the two models, I was curious.", "votes": -2 }, { "id": 483247, "postDate": "2019-03-04T11:18:17.813Z", "content": "Thanks all your effort on this competition!
", "rawMarkdown": "Thanks all your effort on this competition!" }, { "id": 479875, "postDate": "2019-02-27T14:25:09.090Z", "content": "How many customers bought during that period?
", "rawMarkdown": "How many customers bought during that period?" }, { "id": 476379, "postDate": "2019-02-22T03:39:12.843Z", "content": "Thankyou for providing the entrance of late submission. My late submission score is 0.88398. I should trust my model and local cv(0.715) rather than my intution.
", "rawMarkdown": "Thankyou for providing the entrance of late submission. My late submission score is 0.88398. I should trust my model and local cv(0.715) rather than my intution." }, { "id": 478495, "postDate": "2019-02-26T08:18:58.063Z", "rawMarkdown": "", "votes": -1, "isDeleted": true }, { "id": 481985, "postDate": "2019-03-02T06:25:31.383Z", "content": "Thanks for your efforts!
", "rawMarkdown": "Thanks for your efforts!" }, { "id": 476650, "postDate": "2019-02-22T13:32:23.117Z", "content": "Thanks for your efforts!
", "rawMarkdown": "Thanks for your efforts!" } ], "comments": [ { "id": 478114, "author_name": "Julia Elliott", "author_url": "", "post_date": "2019-02-25T18:53:30.363000", "content": "We are still working on the leaderboard removal of rules violators. Due to downtime and migration activities over the weekend, we had to delay these finalization activities. Now that some migration dust has settled, we expect the finalization should be completed in the next day or two. Likewise, we also waited to adjust the team count for medals and points calculations. Do not fret, we have every intention of closing out these activities as quickly as possible.
", "votes": 16, "replies": [ { "id": 481765, "author_name": "ChrisCC", "author_url": "", "post_date": "2019-03-01T20:19:02.187000", "content": "Hi Julie, thank you for the efforts the Kaggle team has made to sustain this competition. Just wondering if there are any updates on the final process?
\n\nThanks,
", "votes": 2, "replies": [] }, { "id": 481828, "author_name": "Julia Elliott", "author_url": "", "post_date": "2019-03-01T22:18:37.510000", "content": "Apologies for the delay. We had to overcome a few post-migration issues with our algorithm to remove rules violators. Please see the 3/1 update in my original post above. The leaderboard is now final.
", "votes": 0, "replies": [] } ] }, { "id": 476568, "author_name": "ValeriyBabushkin", "author_url": "", "post_date": "2019-02-22T10:52:34.370000", "content": "@juliaelliott\nHey Julia - many thanks for an update\nI wonder could the medals and score be calculated based on the original number of participants - I would think that if you have joined and then dropped because of some perturbations in competition - you still have already joined
", "votes": 17, "replies": [ { "id": 476607, "author_name": "MuhammedBuyukkinaci", "author_url": "", "post_date": "2019-02-22T11:58:17.413000", "content": "I agree with you @ValeriyBabushkin. On Quora Insincere competition, More than 4000 people joined and only 1400 of them made submissions. The medals were given according to the number of 4000.
", "votes": 17, "replies": [] }, { "id": 476817, "author_name": "Mikhail Chesnokov", "author_url": "", "post_date": "2019-02-23T08:35:12.867000", "content": "Absolutely agree with Valeriy, that would be honest to count all joined participants in the final standings.
", "votes": 14, "replies": [] }, { "id": 476839, "author_name": "Stanislav Sopov", "author_url": "", "post_date": "2019-02-23T09:48:33.930000", "content": "Also, I agree with @ValeriyBabushkin!\nIt would be fair to evaluate rank (%) per all participants who joined the competition because the data and task strongly haven't changed. Probably, participants who submitted before the competition changes also should be evaluated in the total rank(%)
", "votes": 14, "replies": [] }, { "id": 476867, "author_name": "AmirH", "author_url": "", "post_date": "2019-02-23T10:58:06.407000", "content": "I also agree with @ValeriyBabushkin
", "votes": 12, "replies": [] }, { "id": 476908, "author_name": "Lyalikov Artyom", "author_url": "", "post_date": "2019-02-23T12:43:16.247000", "content": "Agree with @ValeriyBabushkin
", "votes": 13, "replies": [] }, { "id": 477224, "author_name": "0xFunky", "author_url": "", "post_date": "2019-02-24T06:41:56.700000", "content": "Also Agree with @ValeriyBabushkin
", "votes": 12, "replies": [] }, { "id": 477297, "author_name": "Andrei Dukhounik", "author_url": "", "post_date": "2019-02-24T10:03:36.783000", "content": "This will be an additional motivation to participate in the competitions after rerun or important changes. Agree with @ValeriyBabushkin
", "votes": 13, "replies": [] }, { "id": 477770, "author_name": "Haoran Zhou", "author_url": "", "post_date": "2019-02-25T08:34:07.877000", "content": "Can not agree more
", "votes": 3, "replies": [] }, { "id": 477838, "author_name": "Nikita Kozodoi", "author_url": "", "post_date": "2019-02-25T10:49:09.953000", "content": "I can also agree with @ValeriyBabushkin. The competition deadline was not postponed after the changes, so it would be fair not to reduce the number of participants as well and compute ratings & medals based on the total number teams including the ones that submited before the changes. I don't think it would be possible to rescore old submissions because the data is now different, but it would still be more honest for the people that continued to work on this competition during the second stage.
", "votes": 6, "replies": [] } ] }, { "id": 477073, "author_name": "Jules", "author_url": "", "post_date": "2019-02-23T20:04:27.957000", "content": "I also agree with @ValeriyBabushkin. I noticed there is (now) no disclaimer message on the leaderboard, points and medals are not yet awarded it seems?
\n\nedit: disclaimer is only visible for me if I return to the LB via the back button on the browser
", "votes": 10, "replies": [] }, { "id": 478091, "author_name": "miguel perez", "author_url": "", "post_date": "2019-02-25T18:01:14.103000", "content": "Guys, have a look at the leaderboard. Not sure if it will stay like this but in this moment there are a lot more medal positions... if it's not a bug there is a chance that suggestion was accepted :-)
", "votes": 4, "replies": [ { "id": 478097, "author_name": "Ren", "author_url": "", "post_date": "2019-02-25T18:22:31.523000", "content": "Which means all 0 sample submission have the chance to get a silver medal. Pretty sad for this kind of situation happened again and again.
", "votes": 9, "replies": [] }, { "id": 478187, "author_name": "KALE", "author_url": "", "post_date": "2019-02-25T21:33:21.440000", "content": "Unfortunately, the case you would not hope happens again. But lottery tickets are free. :P
", "votes": 0, "replies": [] }, { "id": 478193, "author_name": "Ren", "author_url": "", "post_date": "2019-02-25T21:45:49.050000", "content": "You could use a quote.
\n\nAll these flawed competitions just remind me more of the importance of clean data and meaningful evaluation metric for my projects at work.
", "votes": 0, "replies": [] }, { "id": 478381, "author_name": "NotaAaTuner", "author_url": "", "post_date": "2019-02-26T04:55:24.680000", "content": "What is the error for all 0 submission? In the future, there better be some cutoff of winner or not based on some pre-defined naive submission. Otherwise, it is just hilarious.
", "votes": 1, "replies": [] }, { "id": 478428, "author_name": "ykaneko", "author_url": "", "post_date": "2019-02-26T06:14:19.340000", "content": "All 0 sub silver-medal is too silly. It's unfair for people who tackle other difficult and well-designed competitions , and this is disparaging the value of silver medals.\nThis competition was \"relaunched\". Medals should not be distributed based on too many teams.\n@juliaelliott, Please re-think and don't disappoint me (and maybe other kagglers) any further.
", "votes": 5, "replies": [] }, { "id": 478490, "author_name": "miguel perez", "author_url": "", "post_date": "2019-02-26T08:14:02.987000", "content": "It would not be first time, not even the second for sample submission get a silver in a competition. \nThe difference now -I think- is in other cases was an issue of the competition design/data and this time it would be a measure of how sorry Kaggle team is for all the problems in this one (clearly a reboot is not a two stage). \nPersonally, I decided not to overthink it much when I saw the \"happiness for all\" rain of medals (you are all comenting on silver but bronze used to be something meaningful too).
", "votes": 4, "replies": [] }, { "id": 480903, "author_name": "Edward Cao", "author_url": "", "post_date": "2019-02-28T20:41:30.757000", "content": "Agree with @NotaAaTuner! Looks like the naive 0 submission without any efforts could be awarded a silver medal, which is unfair for people taking efforts on \"real\" model. @juliaelliott, Please re-think and don't disappoint kagglers taking efforts in this competition! thanks
", "votes": 5, "replies": [] }, { "id": 481010, "author_name": "luyaxin", "author_url": "", "post_date": "2019-03-01T00:59:01.637000", "content": "Agree with all above comments. Competition reboot is not the same as a 2-stage competition. And all-zero submissions getting silver really devalues all medals awarded by this competition (not only silver and bronze) and effort/time that people actually spent in this competition.
", "votes": 0, "replies": [] }, { "id": 481749, "author_name": "", "author_url": "", "post_date": "2019-03-01T19:41:43.193000", "content": "", "votes": -1, "replies": [] } ] }, { "id": 481052, "author_name": "Jianhua Ma", "author_url": "", "post_date": "2019-03-01T02:15:17.137000", "content": "@juliaelliott\nso I think the best strategy is that try to add a function in your standard algorithm to filter out the submission with all 0.
", "votes": 1, "replies": [ { "id": 481061, "author_name": "", "author_url": "", "post_date": "2019-03-01T02:33:17.627000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 478062, "author_name": "ChrisCC", "author_url": "", "post_date": "2019-02-25T17:20:11.213000", "content": "Agree with @ValeriyBabushkin that it would be fairer to rank based on the number of original participants.
", "votes": 1, "replies": [] }, { "id": 477335, "author_name": "merazga7", "author_url": "", "post_date": "2019-02-24T10:58:24.073000", "content": "can we get the right predicted values' file?
", "votes": 1, "replies": [] }, { "id": 476511, "author_name": "0xFunky", "author_url": "", "post_date": "2019-02-22T09:09:42.907000", "content": "Thanks all your effort on this competition!!
", "votes": 1, "replies": [] }, { "id": 478026, "author_name": "David Kagan", "author_url": "", "post_date": "2019-02-25T16:17:58.473000", "content": "Hey @juliaelliott can you please update when the points and the medals will be awarded?
", "votes": 2, "replies": [ { "id": 481829, "author_name": "Julia Elliott", "author_url": "", "post_date": "2019-03-01T22:19:37.167000", "content": "Apologies for the delay. We had to overcome a few post-migration issues with our algorithm to remove rules violators. Please see the 3/1 update in my original post above. The leaderboard is now final. Medals & Points will be updated within the next 24 hours.
", "votes": 1, "replies": [] } ] }, { "id": 476476, "author_name": "MichaelP", "author_url": "", "post_date": "2019-02-22T07:57:01.997000", "content": "Thank you for your efforts to enable this competition to continue.
", "votes": 0, "replies": [] }, { "id": 478037, "author_name": "MuhammedBuyukkinaci", "author_url": "", "post_date": "2019-02-25T16:31:07.223000", "content": "I think Kaggle should award more people who continued to compete in this competition by caring the number of ALL people joining this competition. This competition started and a lot of abnormalities happened in this competition:
\n\n1) The metric of competition was chosen as log of sum but actually it was sum of log or vice versa(I don't remember exactly).
\n\n2) The leakage was found and revealed.
\n\n3) After the founding of leakage, people had to wait 3 weeks for the competition to be re-arranged by Kaggle Team including @juliaelliott .
\n\n4) Competitors spent just 3 weeks on this competition(compared to normal competitions, it was a short time).
", "votes": 1, "replies": [] }, { "id": 481801, "author_name": "Jianhua Ma", "author_url": "", "post_date": "2019-03-01T21:50:01.250000", "content": "Hi, @juliaelliott why my submission is disappear???????
", "votes": -1, "replies": [ { "id": 481824, "author_name": "Julia Elliott", "author_url": "", "post_date": "2019-03-01T22:16:34.907000", "content": "See 3/1 update made to original post.
", "votes": 0, "replies": [] } ] }, { "id": 478354, "author_name": "atfujita", "author_url": "", "post_date": "2019-02-26T04:02:35.967000", "content": "\n\nJust to confirm...
\n\nCalculation of metrics is log of sum or sum of log？
\n\nI understood it as log of sum. However, the score was better for sum of log.\nAlthough there are differences in random seed between the two models, I was curious.
", "votes": -2, "replies": [] }, { "id": 483247, "author_name": "Elvira Astashkina", "author_url": "", "post_date": "2019-03-04T11:18:17.813000", "content": "Thanks all your effort on this competition!
", "votes": 0, "replies": [] }, { "id": 479875, "author_name": "merazga7", "author_url": "", "post_date": "2019-02-27T14:25:09.090000", "content": "How many customers bought during that period?
", "votes": 0, "replies": [] }, { "id": 476379, "author_name": "Simon Wang", "author_url": "", "post_date": "2019-02-22T03:39:12.843000", "content": "Thankyou for providing the entrance of late submission. My late submission score is 0.88398. I should trust my model and local cv(0.715) rather than my intution.
", "votes": 0, "replies": [] }, { "id": 478495, "author_name": "", "author_url": "", "post_date": "2019-02-26T08:18:58.063000", "content": "", "votes": -1, "replies": [] }, { "id": 481985, "author_name": "Yanni", "author_url": "", "post_date": "2019-03-02T06:25:31.383000", "content": "Thanks for your efforts!
", "votes": 0, "replies": [] }, { "id": 476650, "author_name": "airwindow", "author_url": "", "post_date": "2019-02-22T13:32:23.117000", "content": "Thanks for your efforts!
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "478114": "We are still working on the leaderboard removal of rules violators. Due to downtime and migration activities over the weekend, we had to delay these finalization activities. Now that some migration dust has settled, we expect the finalization should be completed in the next day or two. Likewise, we also waited to adjust the team count for medals and points calculations. Do not fret, we have every intention of closing out these activities as quickly as possible.", "476373": "**3/1 UDPATE**: Rules violators have officially been removed from the leaderboard, and the private leaderboard is now final. If you are no longer on the leaderboard, it is because you were removed, likely due to holding multiple accounts or identified collusion with other team(s). Evidence supporting removal is largely conclusive, but you are welcome to [submit a compliance appeal](https://www.kaggle.com/compliance) if you believe your team was wrongfully removed. \n\n-----\n\nThe leaderboard has been updated and the private leaderboard revealed. Note: The final standings are **pending verification** by our standard algorithms to remove multiple accounts violators, hence the disclaimer message on the leaderboard. Due to Kaggle's scheduled downtime, this final leaderboard verification may take a few days to post.\n\nIn the meantime, a tentative congratulations to the winners!\n\nA huge thank you to everyone for your participation. We recognize there were a number of unexpected issues that plagued this competition, and I'm sorry for the impact this may have had on your experience. This was atypical for Kaggle, and we tried our very best to sustain the competition so that the work you put in could be recognized.\n\nPlease note the following:\n\n- As described when the competition's design was revised, the private leaderboard is now based on future-looking predictions from the Stage 2 timeframe of Dec 1, 2018 to Jan 31, 2019.\n\n- The public leaderboard now contains no scores, as a result of bullet #1.\n\n- As is Kaggle policy for [2-stage competitions](https://www.kaggle.com/two-stage-frequently-asked-questions), the team count for points and medals calculation is based on original team volumes and has been adjusted as such.\n", "476568": "@juliaelliott\nHey Julia - many thanks for an update\nI wonder could the medals and score be calculated based on the original number of participants - I would think that if you have joined and then dropped because of some perturbations in competition - you still have already joined ", "477073": "I also agree with @ValeriyBabushkin. I noticed there is (now) no disclaimer message on the leaderboard, points and medals are not yet awarded it seems?\n\nedit: disclaimer is only visible for me if I return to the LB via the back button on the browser", "478091": "Guys, have a look at the leaderboard. Not sure if it will stay like this but in this moment there are a lot more medal positions... if it's not a bug there is a chance that suggestion was accepted :-)", "481052": "@juliaelliott\nso I think the best strategy is that try to add a function in your standard algorithm to filter out the submission with all 0. ", "478062": "Agree with @ValeriyBabushkin that it would be fairer to rank based on the number of original participants. ", "477335": "can we get the right predicted values' file?", "476511": "Thanks all your effort on this competition!!", "478026": "Hey @juliaelliott can you please update when the points and the medals will be awarded? ", "476476": "Thank you for your efforts to enable this competition to continue. ", "478037": "I think Kaggle should award more people who continued to compete in this competition by caring the number of ALL people joining this competition. This competition started and a lot of abnormalities happened in this competition:\n\n1) The metric of competition was chosen as log of sum but actually it was sum of log or vice versa(I don't remember exactly).\n\n2) The leakage was found and revealed.\n\n3) After the founding of leakage, people had to wait 3 weeks for the competition to be re-arranged by Kaggle Team including @juliaelliott .\n\n4) Competitors spent just 3 weeks on this competition(compared to normal competitions, it was a short time).", "481801": "Hi, @juliaelliott why my submission is disappear???????", "478354": "Hi @juliaelliott\n\n\nJust to confirm...\n\nCalculation of metrics is log of sum or sum of log？\n\nI understood it as log of sum. However, the score was better for sum of log.\nAlthough there are differences in random seed between the two models, I was curious.", "483247": "Thanks all your effort on this competition!", "479875": "How many customers bought during that period?", "476379": "Thankyou for providing the entrance of late submission. My late submission score is 0.88398. I should trust my model and local cv(0.715) rather than my intution.", "478495": "", "481985": "Thanks for your efforts!", "476650": "Thanks for your efforts!" } }