![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2332181%2Fa1789488b5c823ac4b03fcd8ea1af016%2F396pss.jpg?generation=1567091991354419&alt=media\")
Congratulations to all of the winners and thanks to kaggle for hosting this competition.\nThis is by far the most stable LB I've ever seen on Kaggle! I hope it means the results will be useful for the organizers as well.
\n\nLooking forward to the top teams sharing their solutions.
", "messages": [ { "id": "610670", "postDate": "08/29/2019 00:03:47", "content": "Congratulations to all of the winners and thanks to kaggle for hosting this competition.\nThis is by far the most stable LB I've ever seen on Kaggle! I hope it means the results will be useful for the organizers as well.
\n\nLooking forward to the top teams sharing their solutions.
", "rawMarkdown": "Congratulations to all of the winners and thanks to kaggle for hosting this competition.\nThis is by far the most stable LB I've ever seen on Kaggle! I hope it means the results will be useful for the organizers as well.\n\nLooking forward to the top teams sharing their solutions.", "votes": null }, { "id": "610672", "postDate": "08/29/2019 00:06:44", "content": "Biggest movers were me and another team who had access to the leaked file! (because we didn't choose for final sub)
\n\nCongratulations to everyone.
", "rawMarkdown": "Biggest movers were me and another team who had access to the leaked file! (because we didn't choose for final sub)\n\nCongratulations to everyone.", "votes": null }, { "id": "610673", "postDate": "08/29/2019 00:06:49", "content": "Yes, I never saw such stable public/private LB like this. Great dataset, love it!
", "rawMarkdown": "Yes, I never saw such stable public/private LB like this. Great dataset, love it!", "votes": null }, { "id": "610680", "postDate": "08/29/2019 00:08:58", "content": "I wish I could spend more time on this competition. Congratulations to you btw. Another gold for you!
", "rawMarkdown": "I wish I could spend more time on this competition. Congratulations to you btw. Another gold for you!", "votes": null }, { "id": "610681", "postDate": "08/29/2019 00:09:06", "content": "This means that public/privated were balanced well. Thanks to the hosts and kaggle team!
", "rawMarkdown": "This means that public/privated were balanced well. Thanks to the hosts and kaggle team!", "votes": null }, { "id": "610682", "postDate": "08/29/2019 00:09:59", "content": "Thanks! It was quite an eventful last day! Very stable indeed, though it was to be expected as it was hard to overfit without noticing in this competition.
", "rawMarkdown": "Thanks! It was quite an eventful last day! Very stable indeed, though it was to be expected as it was hard to overfit without noticing in this competition.", "votes": null }, { "id": "610692", "postDate": "08/29/2019 00:13:36", "content": "congratulations @milosh. 4th place in your first competition. That's impressive!
", "rawMarkdown": "congratulations @milosh. 4th place in your first competition. That's impressive!", "votes": null }, { "id": "610694", "postDate": "08/29/2019 00:14:44", "content": "I agree. It makes everyone happier. LANL EQ comp was like a nightmare...
", "rawMarkdown": "I agree. It makes everyone happier. LANL EQ comp was like a nightmare...", "votes": null }, { "id": "610698", "postDate": "08/29/2019 00:19:23", "content": "Yes amazing, I stand the same across both.
", "rawMarkdown": "Yes amazing, I stand the same across both.", "votes": null }, { "id": "610704", "postDate": "08/29/2019 00:24:43", "content": "congratulations
", "rawMarkdown": "congratulations", "votes": null }, { "id": "610710", "postDate": "08/29/2019 00:28:59", "content": "Thank you! We did work very hard from the start. Frankly, I admire @robin111 who entered late, and solo, and managed to finish at #6 with so few submissions!
", "rawMarkdown": "Thank you! We did work very hard from the start. Frankly, I admire @robin111 who entered late, and solo, and managed to finish at #6 with so few submissions!", "votes": null }, { "id": "610738", "postDate": "08/29/2019 00:45:47", "content": "You did an awesome job. Usually it takes one or two competitions to get the hang on kaggle but started like a fire from your very first competition. Hope to see u in next comps as well
", "rawMarkdown": "You did an awesome job. Usually it takes one or two competitions to get the hang on kaggle but started like a fire from your very first competition. Hope to see u in next comps as well", "votes": null }, { "id": "610982", "postDate": "08/29/2019 03:45:13", "content": "i only messed around with this contest a little but came back to see the results. (months later) and wow yes. i think every private and public result match with in .01 it would of course be far more variant if my score was -2 or something but still. its refreshing to see such consistency.
", "rawMarkdown": "i only messed around with this contest a little but came back to see the results. (months later) and wow yes. i think every private and public result match with in .01 it would of course be far more variant if my score was -2 or something but still. its refreshing to see such consistency.", "votes": null }, { "id": "610984", "postDate": "08/29/2019 03:47:49", "content": "Indeed. I really needed to see this to trust models again 😂 in most competitions people overfit public LB so much that their models don’t generalize at all and it makes this stigma that are ml models any good whatsoever? Take two sigma stock market prediction for an example.
", "rawMarkdown": "Indeed. I really needed to see this to trust models again 😂 in most competitions people overfit public LB so much that their models don’t generalize at all and it makes this stigma that are ml models any good whatsoever? Take two sigma stock market prediction for an example.", "votes": null }, { "id": "611245", "postDate": "08/29/2019 07:23:26", "content": "Thanks @milosh. Congrats on you amazing finish! That was quite a battle for fourth place on the last day between your team and the DL guys.
", "rawMarkdown": "Thanks @milosh. Congrats on you amazing finish! That was quite a battle for fourth place on the last day between your team and the DL guys.", "votes": null }, { "id": "611281", "postDate": "08/29/2019 07:44:34", "content": "Yes, that's really interesting. Even bronze positions remained quite the same!\nIt was really a nice competition :)
", "rawMarkdown": "Yes, that's really interesting. Even bronze positions remained quite the same!\nIt was really a nice competition :)", "votes": null }, { "id": "611716", "postDate": "08/29/2019 11:56:52", "content": "This was a really well run competition. No leaks, organisers were around the whole time, and no crazy deviation between train & test.
\n\nI wish they could all be like this!
", "rawMarkdown": "This was a really well run competition. No leaks, organisers were around the whole time, and no crazy deviation between train & test.\n\nI wish they could all be like this!", "votes": null }, { "id": "612043", "postDate": "08/29/2019 15:19:53", "content": "
Is the private leaderboard based on other 70% testing data? But I did not submit prediction for that testing set.
", "rawMarkdown": "Is the private leaderboard based on other 70% testing data? But I did not submit prediction for that testing set.", "votes": null }, { "id": "612181", "postDate": "08/29/2019 16:34:11", "content": "I believe the way it works is that your submissions are for both the public and private test sets. The thing is, you don't know which rows are for which test set. So you actually did submit predictions for the private LB dataset every time you uploaded a submission...
", "rawMarkdown": "I believe the way it works is that your submissions are for both the public and private test sets. The thing is, you don't know which rows are for which test set. So you actually did submit predictions for the private LB dataset every time you uploaded a submission...", "votes": null }, { "id": "612384", "postDate": "08/29/2019 18:53:45", "content": "@abendorf what you believe is correct.
", "rawMarkdown": "abendorf what you believe is correct.", "votes": null } ], "comments": [ { "id": 610672, "author_name": "julianmukaj", "author_url": "", "post_date": "08/29/2019 00:06:44", "content": "Biggest movers were me and another team who had access to the leaked file! (because we didn't choose for final sub)
\n\nCongratulations to everyone.
", "votes": null, "replies": [] }, { "id": 610673, "author_name": "khahuras", "author_url": "", "post_date": "08/29/2019 00:06:49", "content": "Yes, I never saw such stable public/private LB like this. Great dataset, love it!
", "votes": null, "replies": [ { "id": 610680, "author_name": "mhviraf", "author_url": "", "post_date": "08/29/2019 00:08:58", "content": "I wish I could spend more time on this competition. Congratulations to you btw. Another gold for you!
", "votes": null, "replies": [] } ] }, { "id": 610681, "author_name": "sorrow321", "author_url": "", "post_date": "08/29/2019 00:09:06", "content": "This means that public/privated were balanced well. Thanks to the hosts and kaggle team!
", "votes": null, "replies": [ { "id": 610694, "author_name": "mhviraf", "author_url": "", "post_date": "08/29/2019 00:14:44", "content": "I agree. It makes everyone happier. LANL EQ comp was like a nightmare...
", "votes": null, "replies": [] } ] }, { "id": 610682, "author_name": "milosh", "author_url": "", "post_date": "08/29/2019 00:09:59", "content": "Thanks! It was quite an eventful last day! Very stable indeed, though it was to be expected as it was hard to overfit without noticing in this competition.
", "votes": null, "replies": [ { "id": 610692, "author_name": "mhviraf", "author_url": "", "post_date": "08/29/2019 00:13:36", "content": "congratulations @milosh. 4th place in your first competition. That's impressive!
", "votes": null, "replies": [] }, { "id": 610710, "author_name": "milosh", "author_url": "", "post_date": "08/29/2019 00:28:59", "content": "Thank you! We did work very hard from the start. Frankly, I admire @robin111 who entered late, and solo, and managed to finish at #6 with so few submissions!
", "votes": null, "replies": [] }, { "id": 610738, "author_name": "mhviraf", "author_url": "", "post_date": "08/29/2019 00:45:47", "content": "You did an awesome job. Usually it takes one or two competitions to get the hang on kaggle but started like a fire from your very first competition. Hope to see u in next comps as well
", "votes": null, "replies": [] }, { "id": 611245, "author_name": "robin111", "author_url": "", "post_date": "08/29/2019 07:23:26", "content": "Thanks @milosh. Congrats on you amazing finish! That was quite a battle for fourth place on the last day between your team and the DL guys.
", "votes": null, "replies": [] } ] }, { "id": 610698, "author_name": "drpatrickchan", "author_url": "", "post_date": "08/29/2019 00:19:23", "content": "Yes amazing, I stand the same across both.
", "votes": null, "replies": [ { "id": 610704, "author_name": "mhviraf", "author_url": "", "post_date": "08/29/2019 00:24:43", "content": "congratulations
", "votes": null, "replies": [] } ] }, { "id": 610982, "author_name": "jscheibel", "author_url": "", "post_date": "08/29/2019 03:45:13", "content": "i only messed around with this contest a little but came back to see the results. (months later) and wow yes. i think every private and public result match with in .01 it would of course be far more variant if my score was -2 or something but still. its refreshing to see such consistency.
", "votes": null, "replies": [ { "id": 610984, "author_name": "mhviraf", "author_url": "", "post_date": "08/29/2019 03:47:49", "content": "Indeed. I really needed to see this to trust models again 😂 in most competitions people overfit public LB so much that their models don’t generalize at all and it makes this stigma that are ml models any good whatsoever? Take two sigma stock market prediction for an example.
", "votes": null, "replies": [] } ] }, { "id": 611281, "author_name": "raimonds1993", "author_url": "", "post_date": "08/29/2019 07:44:34", "content": "Yes, that's really interesting. Even bronze positions remained quite the same!\nIt was really a nice competition :)
", "votes": null, "replies": [] }, { "id": 611716, "author_name": "bigironsphere", "author_url": "", "post_date": "08/29/2019 11:56:52", "content": "This was a really well run competition. No leaks, organisers were around the whole time, and no crazy deviation between train & test.
\n\nI wish they could all be like this!
", "votes": null, "replies": [] }, { "id": 612043, "author_name": "frtgnn", "author_url": "", "post_date": "08/29/2019 15:19:53", "content": "
Is the private leaderboard based on other 70% testing data? But I did not submit prediction for that testing set.
", "votes": null, "replies": [ { "id": 612181, "author_name": "abendorf", "author_url": "", "post_date": "08/29/2019 16:34:11", "content": "I believe the way it works is that your submissions are for both the public and private test sets. The thing is, you don't know which rows are for which test set. So you actually did submit predictions for the private LB dataset every time you uploaded a submission...
", "votes": null, "replies": [] }, { "id": 612384, "author_name": "mhviraf", "author_url": "", "post_date": "08/29/2019 18:53:45", "content": "@abendorf what you believe is correct.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "610670": "Congratulations to all of the winners and thanks to kaggle for hosting this competition.\nThis is by far the most stable LB I've ever seen on Kaggle! I hope it means the results will be useful for the organizers as well.\n\nLooking forward to the top teams sharing their solutions.", "610672": "Biggest movers were me and another team who had access to the leaked file! (because we didn't choose for final sub)\n\nCongratulations to everyone.", "610673": "Yes, I never saw such stable public/private LB like this. Great dataset, love it!", "610680": "I wish I could spend more time on this competition. Congratulations to you btw. Another gold for you!", "610681": "This means that public/privated were balanced well. Thanks to the hosts and kaggle team!", "610682": "Thanks! It was quite an eventful last day! Very stable indeed, though it was to be expected as it was hard to overfit without noticing in this competition.", "610692": "congratulations @milosh. 4th place in your first competition. That's impressive!", "610694": "I agree. It makes everyone happier. LANL EQ comp was like a nightmare...", "610698": "Yes amazing, I stand the same across both.", "610704": "congratulations", "610710": "Thank you! We did work very hard from the start. Frankly, I admire @robin111 who entered late, and solo, and managed to finish at #6 with so few submissions!", "610738": "You did an awesome job. Usually it takes one or two competitions to get the hang on kaggle but started like a fire from your very first competition. Hope to see u in next comps as well", "610982": "i only messed around with this contest a little but came back to see the results. (months later) and wow yes. i think every private and public result match with in .01 it would of course be far more variant if my score was -2 or something but still. its refreshing to see such consistency.", "610984": "Indeed. I really needed to see this to trust models again 😂 in most competitions people overfit public LB so much that their models don’t generalize at all and it makes this stigma that are ml models any good whatsoever? Take two sigma stock market prediction for an example.", "611245": "Thanks @milosh. Congrats on you amazing finish! That was quite a battle for fourth place on the last day between your team and the DL guys.", "611281": "Yes, that's really interesting. Even bronze positions remained quite the same!\nIt was really a nice competition :)", "611716": "This was a really well run competition. No leaks, organisers were around the whole time, and no crazy deviation between train & test.\n\nI wish they could all be like this!", "612043": "", "612162": "Is the private leaderboard based on other 70% testing data? But I did not submit prediction for that testing set.", "612181": "I believe the way it works is that your submissions are for both the public and private test sets. The thing is, you don't know which rows are for which test set. So you actually did submit predictions for the private LB dataset every time you uploaded a submission...", "612384": "abendorf what you believe is correct." }, "source": "meta" }