Hello @jetakow
\nI think most of the public forum codes including my work elicit the scores before the submissions were reopened. As an example, my baseline inference kernel shows up a score of 0.611 (this was the score before the submissions were disabled).
\nConsidering that the yester scores are now superfluous, can we somehow display the new scores against these notebooks, at least for the resubmitted ones? It will be highly useful for reference in my opinion.
Best regards.
", "messages": [ { "id": "2712195", "postDate": "03/23/2024 11:20:07", "content": "Hello @jetakow
\nI think most of the public forum codes including my work elicit the scores before the submissions were reopened. As an example, my baseline inference kernel shows up a score of 0.611 (this was the score before the submissions were disabled).
\nConsidering that the yester scores are now superfluous, can we somehow display the new scores against these notebooks, at least for the resubmitted ones? It will be highly useful for reference in my opinion.
Best regards.
", "rawMarkdown": "Hello @jetakow \n\nI think most of the public forum codes including my work elicit the scores before the submissions were reopened. As an example, my baseline inference kernel shows up a score of 0.611 (this was the score before the submissions were disabled). \nConsidering that the yester scores are now superfluous, can we somehow display the new scores against these notebooks, at least for the resubmitted ones? It will be highly useful for reference in my opinion.\n\nBest regards.", "votes": null }, { "id": "2712204", "postDate": "03/23/2024 11:33:08", "content": "Yes, scores of older public notebooks are misleading now. Even if we push a new version it will score much worse, so won't affect the old non-reproducible score that is shown under the notebook name.
", "rawMarkdown": "Yes, scores of older public notebooks are misleading now. Even if we push a new version it will score much worse, so won't affect the old non-reproducible score that is shown under the notebook name.", "votes": null }, { "id": "2712221", "postDate": "03/23/2024 11:46:33", "content": "@kononenko I asked this question for this reason precisely. I think the new scores are almost going to be lower than the old scores, unless one decides to copy and edit the kernel. This will mislead the viewers and participants in the forum as the public scores may not be reflective of the new submissions.
", "rawMarkdown": "kononenko I asked this question for this reason precisely. I think the new scores are almost going to be lower than the old scores, unless one decides to copy and edit the kernel. This will mislead the viewers and participants in the forum as the public scores may not be reflective of the new submissions.", "votes": null }, { "id": "2712407", "postDate": "03/23/2024 14:42:41", "content": "I apologize for the inconvenience, we didn't think of this possibility. I understand that it is counterproductive. I am not sure if it is possible to fix it in the amount of time we have, I will get in touch with Kaggle staff. Thanks for raising this issue.
", "rawMarkdown": "I apologize for the inconvenience, we didn't think of this possibility. I understand that it is counterproductive. I am not sure if it is possible to fix it in the amount of time we have, I will get in touch with Kaggle staff. Thanks for raising this issue.", "votes": null }, { "id": "2712499", "postDate": "03/23/2024 15:55:28", "content": "Thanks for the information @jetakow.
\nI will then encourage people to kindly make new kernels to explicitly cite activity before and after the submission break. Also, do not filter kernels by public score while looking for code materials.
Yes, scores of older public notebooks are misleading now. Even if we push a new version it will score much worse, so won't affect the old non-reproducible score that is shown under the notebook name.
", "votes": null, "replies": [ { "id": 2712221, "author_name": "ravi20076", "author_url": "", "post_date": "03/23/2024 11:46:33", "content": "@kononenko I asked this question for this reason precisely. I think the new scores are almost going to be lower than the old scores, unless one decides to copy and edit the kernel. This will mislead the viewers and participants in the forum as the public scores may not be reflective of the new submissions.
", "votes": null, "replies": [] } ] }, { "id": 2712407, "author_name": "jetakow", "author_url": "", "post_date": "03/23/2024 14:42:41", "content": "I apologize for the inconvenience, we didn't think of this possibility. I understand that it is counterproductive. I am not sure if it is possible to fix it in the amount of time we have, I will get in touch with Kaggle staff. Thanks for raising this issue.
", "votes": null, "replies": [ { "id": 2712499, "author_name": "ravi20076", "author_url": "", "post_date": "03/23/2024 15:55:28", "content": "Thanks for the information @jetakow.
\nI will then encourage people to kindly make new kernels to explicitly cite activity before and after the submission break. Also, do not filter kernels by public score while looking for code materials.