{
  "id": 125843,
  "title": "Would've got a gold medal if I didn't select the submissions for final score",
  "url": "/competitions/ashrae-energy-prediction/discussion/125843",
  "author_name": "",
  "post_date": "2020-01-14T03:36:23.361791900Z",
  "votes": 6,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Lesson learned. I will not select submissions for private scoring again. Now that all submissions are rescored, I just noticed that one of my kernels got a private score that would award a gold medal:\n- <a href=\"https://www.kaggle.com/wittmannf/blender-leak-validation-inception\">https://www.kaggle.com/wittmannf/blender-leak-validation-inception</a></p>\n\n<p>The solution is almost the same that <a href=\"https://www.kaggle.com/c/ashrae-energy-prediction/discussion/123288\">I described here</a>. The only difference is that I did a second blending of the output from the first blending and the same submissions. It got a private score of 1.243 (between 13th and 16th position). Anyone else in a similar situation? </p>\n\n<p>PS: I am still happy with my position anyway.</p>\n\n<h3>UPDATE</h3>\n\n<p>I was actually not interpreting correctly how the submissions were selected:</p>\n\n<blockquote>\n  <p>If 2 submissions are not selected, they will be automatically chosen based on your best submission scores on the public leaderboard.</p>\n</blockquote>\n\n<p>Since the public score was worse, that kernel would not be selected anyway. </p>",
  "messages": [
    {
      "id": "718129",
      "postDate": "01/14/2020 03:36:23",
      "content": "<p>Lesson learned. I will not select submissions for private scoring again. Now that all submissions are rescored, I just noticed that one of my kernels got a private score that would award a gold medal:\n- <a href=\"https://www.kaggle.com/wittmannf/blender-leak-validation-inception\">https://www.kaggle.com/wittmannf/blender-leak-validation-inception</a></p>\n\n<p>The solution is almost the same that <a href=\"https://www.kaggle.com/c/ashrae-energy-prediction/discussion/123288\">I described here</a>. The only difference is that I did a second blending of the output from the first blending and the same submissions. It got a private score of 1.243 (between 13th and 16th position). Anyone else in a similar situation? </p>\n\n<p>PS: I am still happy with my position anyway.</p>\n\n<h3>UPDATE</h3>\n\n<p>I was actually not interpreting correctly how the submissions were selected:</p>\n\n<blockquote>\n  <p>If 2 submissions are not selected, they will be automatically chosen based on your best submission scores on the public leaderboard.</p>\n</blockquote>\n\n<p>Since the public score was worse, that kernel would not be selected anyway. </p>",
      "rawMarkdown": "Lesson learned. I will not select submissions for private scoring again. Now that all submissions are rescored, I just noticed that one of my kernels got a private score that would award a gold medal:\n- https://www.kaggle.com/wittmannf/blender-leak-validation-inception\n\nThe solution is almost the same that [I described here](https://www.kaggle.com/c/ashrae-energy-prediction/discussion/123288). The only difference is that I did a second blending of the output from the first blending and the same submissions. It got a private score of 1.243 (between 13th and 16th position). Anyone else in a similar situation? \n\nPS: I am still happy with my position anyway.\n\n### UPDATE\nI was actually not interpreting correctly how the submissions were selected:\n&gt; If 2 submissions are not selected, they will be automatically chosen based on your best submission scores on the public leaderboard.\n\nSince the public score was worse, that kernel would not be selected anyway.",
      "votes": null
    },
    {
      "id": "718563",
      "postDate": "01/14/2020 14:32:15",
      "content": "<p>I had a few submissions that would be in the top 100, and I ended up getting 1000+, but I couldn't predict that they would be better. If I had not worked in the last week of the competition, I probably would have picked up silver medal.</p>\n\n<p>PS: I'm not happy hahaha :)</p>",
      "rawMarkdown": "I had a few submissions that would be in the top 100, and I ended up getting 1000+, but I couldn't predict that they would be better. If I had not worked in the last week of the competition, I probably would have picked up silver medal.\n\nPS: I'm not happy hahaha :)",
      "votes": null
    },
    {
      "id": "721808",
      "postDate": "01/17/2020 18:04:48",
      "content": "<p><a href=\"/wittmannf\">@wittmannf</a> </p>\n\n<p>Exactly the same for me !\nI could have been in Silver zone... but NO !</p>\n\n<p>I did not select, and never done except once. how can can you choose a submission with 2 decimal places when the final countdown keeps 4 decimal places ?</p>\n\n<p>Ok, I was a bad guest.</p>",
      "rawMarkdown": "wittmannf \n\nExactly the same for me !\nI could have been in Silver zone... but NO !\n\nI did not select, and never done except once. how can can you choose a submission with 2 decimal places when the final countdown keeps 4 decimal places ?\n\nOk, I was a bad guest.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 718563,
      "author_name": "clappis",
      "author_url": "",
      "post_date": "01/14/2020 14:32:15",
      "content": "<p>I had a few submissions that would be in the top 100, and I ended up getting 1000+, but I couldn't predict that they would be better. If I had not worked in the last week of the competition, I probably would have picked up silver medal.</p>\n\n<p>PS: I'm not happy hahaha :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 721808,
      "author_name": "pfichou",
      "author_url": "",
      "post_date": "01/17/2020 18:04:48",
      "content": "<p><a href=\"/wittmannf\">@wittmannf</a> </p>\n\n<p>Exactly the same for me !\nI could have been in Silver zone... but NO !</p>\n\n<p>I did not select, and never done except once. how can can you choose a submission with 2 decimal places when the final countdown keeps 4 decimal places ?</p>\n\n<p>Ok, I was a bad guest.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "718129": "Lesson learned. I will not select submissions for private scoring again. Now that all submissions are rescored, I just noticed that one of my kernels got a private score that would award a gold medal:\n- https://www.kaggle.com/wittmannf/blender-leak-validation-inception\n\nThe solution is almost the same that [I described here](https://www.kaggle.com/c/ashrae-energy-prediction/discussion/123288). The only difference is that I did a second blending of the output from the first blending and the same submissions. It got a private score of 1.243 (between 13th and 16th position). Anyone else in a similar situation? \n\nPS: I am still happy with my position anyway.\n\n### UPDATE\nI was actually not interpreting correctly how the submissions were selected:\n&gt; If 2 submissions are not selected, they will be automatically chosen based on your best submission scores on the public leaderboard.\n\nSince the public score was worse, that kernel would not be selected anyway.",
    "718563": "I had a few submissions that would be in the top 100, and I ended up getting 1000+, but I couldn't predict that they would be better. If I had not worked in the last week of the competition, I probably would have picked up silver medal.\n\nPS: I'm not happy hahaha :)",
    "721808": "wittmannf \n\nExactly the same for me !\nI could have been in Silver zone... but NO !\n\nI did not select, and never done except once. how can can you choose a submission with 2 decimal places when the final countdown keeps 4 decimal places ?\n\nOk, I was a bad guest."
  },
  "source": "meta"
}