{
  "id": 187756,
  "title": "When you risk everything or the most epic failure in kaggle competitions (Fron top 14% to to 72%)",
  "url": "/competitions/landmark-recognition-2020/discussion/187756",
  "author_name": "",
  "post_date": "2020-09-30T07:12:57.969157600Z",
  "votes": 4,
  "comment_count": 6,
  "views": 0,
  "content": "<p>Hi to all! Be very careful when choosing final submissions.<br>\nMy teammate and I decided to take a risk and chose our models as the final for this competition, ignoring our work with the base model. <br>\nAs a result, we dropped to the top 74%. For me, this is the worst result in kaggle competitions. I broke my medal streak at a competition consisting of three competitions in a row. But you know what's funny, during this competition I learned a lot more than in the previous three :-).<br>\nPlease write whether this is a failure of the year, or someone was like that?<br>\nAlways make your choice correctly and good luck to you all !!!😄😄😄</p>",
  "messages": [
    {
      "id": "1032393",
      "postDate": "09/30/2020 07:12:57",
      "content": "<p>Hi to all! Be very careful when choosing final submissions.<br>\nMy teammate and I decided to take a risk and chose our models as the final for this competition, ignoring our work with the base model. <br>\nAs a result, we dropped to the top 74%. For me, this is the worst result in kaggle competitions. I broke my medal streak at a competition consisting of three competitions in a row. But you know what's funny, during this competition I learned a lot more than in the previous three :-).<br>\nPlease write whether this is a failure of the year, or someone was like that?<br>\nAlways make your choice correctly and good luck to you all !!!😄😄😄</p>",
      "rawMarkdown": "Hi to all! Be very careful when choosing final submissions.\nMy teammate and I decided to take a risk and chose our models as the final for this competition, ignoring our work with the base model. \nAs a result, we dropped to the top 74%. For me, this is the worst result in kaggle competitions. I broke my medal streak at a competition consisting of three competitions in a row. But you know what's funny, during this competition I learned a lot more than in the previous three :-).\nPlease write whether this is a failure of the year, or someone was like that?\nAlways make your choice correctly and good luck to you all !!!😄😄😄",
      "votes": null
    },
    {
      "id": "1032447",
      "postDate": "09/30/2020 07:58:42",
      "content": "<p>Out of interest, given you can choose 2 entries for your final score, why did you decide not to hedge your bets with one baseline submission and one of your own models?</p>\n<p>Agreed, this competition was a good one for learning new techniques, which (to me at least) is the more important thing.</p>",
      "rawMarkdown": "Out of interest, given you can choose 2 entries for your final score, why did you decide not to hedge your bets with one baseline submission and one of your own models?\n\nAgreed, this competition was a good one for learning new techniques, which (to me at least) is the more important thing.",
      "votes": null
    },
    {
      "id": "1032460",
      "postDate": "09/30/2020 08:08:30",
      "content": "<p>haha we r in the same boat ! Like U, I learn alot from this competition (knowledge gained worth as much as a medal) 😏</p>",
      "rawMarkdown": "haha we r in the same boat ! Like U, I learn alot from this competition (knowledge gained worth as much as a medal) 😏",
      "votes": null
    },
    {
      "id": "1032463",
      "postDate": "09/30/2020 08:10:54",
      "content": "<blockquote>\n  <p>Out of interest, given you can choose 2 entries for your final score, why did you decide not to hedge your bets with one baseline submission and one of your own models?</p>\n</blockquote>\n<p>We chose two different models in the hope that one would work. Negative experience is also an experience :-)</p>",
      "rawMarkdown": "> Out of interest, given you can choose 2 entries for your final score, why did you decide not to hedge your bets with one baseline submission and one of your own models?\n\n\nWe chose two different models in the hope that one would work. Negative experience is also an experience :-)",
      "votes": null
    },
    {
      "id": "1032935",
      "postDate": "09/30/2020 14:22:22",
      "content": "<p>Its was right that the organizer baseline was enough but unfortunately i failed to believe that any of the model would work.Sorry to him a lot <a href=\"https://www.kaggle.com/aikhmelnytskyy\" target=\"_blank\">@aikhmelnytskyy</a> </p>",
      "rawMarkdown": "Its was right that the organizer baseline was enough but unfortunately i failed to believe that any of the model would work.Sorry to him a lot @aikhmelnytskyy",
      "votes": null
    },
    {
      "id": "1033002",
      "postDate": "09/30/2020 15:33:35",
      "content": "<p>I was about to say that what <a href=\"https://www.kaggle.com/andypenrose\" target=\"_blank\">@andypenrose</a> said. </p>\n<p>But again, it's all about learning. <br>\nBetter next time 👍</p>",
      "rawMarkdown": "I was about to say that what @andypenrose said. \n\nBut again, it's all about learning. \nBetter next time 👍",
      "votes": null
    },
    {
      "id": "1033131",
      "postDate": "09/30/2020 17:24:26",
      "content": "<p>Keep your head up. Two years ago I dropped from 107 -&gt; 1207 in TalkingData. The most important part is A) take the KO like a champ and just get back up, don't let it bruise your ego, and B) learn from the L.</p>",
      "rawMarkdown": "Keep your head up. Two years ago I dropped from 107 -> 1207 in TalkingData. The most important part is A) take the KO like a champ and just get back up, don't let it bruise your ego, and B) learn from the L.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1032447,
      "author_name": "andypenrose",
      "author_url": "",
      "post_date": "09/30/2020 07:58:42",
      "content": "<p>Out of interest, given you can choose 2 entries for your final score, why did you decide not to hedge your bets with one baseline submission and one of your own models?</p>\n<p>Agreed, this competition was a good one for learning new techniques, which (to me at least) is the more important thing.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1032463,
          "author_name": "aikhmelnytskyy",
          "author_url": "",
          "post_date": "09/30/2020 08:10:54",
          "content": "<blockquote>\n  <p>Out of interest, given you can choose 2 entries for your final score, why did you decide not to hedge your bets with one baseline submission and one of your own models?</p>\n</blockquote>\n<p>We chose two different models in the hope that one would work. Negative experience is also an experience :-)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1033002,
          "author_name": "rahim3",
          "author_url": "",
          "post_date": "09/30/2020 15:33:35",
          "content": "<p>I was about to say that what <a href=\"https://www.kaggle.com/andypenrose\" target=\"_blank\">@andypenrose</a> said. </p>\n<p>But again, it's all about learning. <br>\nBetter next time 👍</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1032460,
      "author_name": "sidneyng",
      "author_url": "",
      "post_date": "09/30/2020 08:08:30",
      "content": "<p>haha we r in the same boat ! Like U, I learn alot from this competition (knowledge gained worth as much as a medal) 😏</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1032935,
      "author_name": "morizin",
      "author_url": "",
      "post_date": "09/30/2020 14:22:22",
      "content": "<p>Its was right that the organizer baseline was enough but unfortunately i failed to believe that any of the model would work.Sorry to him a lot <a href=\"https://www.kaggle.com/aikhmelnytskyy\" target=\"_blank\">@aikhmelnytskyy</a> </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1033131,
      "author_name": "authman",
      "author_url": "",
      "post_date": "09/30/2020 17:24:26",
      "content": "<p>Keep your head up. Two years ago I dropped from 107 -&gt; 1207 in TalkingData. The most important part is A) take the KO like a champ and just get back up, don't let it bruise your ego, and B) learn from the L.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1032393": "Hi to all! Be very careful when choosing final submissions.\nMy teammate and I decided to take a risk and chose our models as the final for this competition, ignoring our work with the base model. \nAs a result, we dropped to the top 74%. For me, this is the worst result in kaggle competitions. I broke my medal streak at a competition consisting of three competitions in a row. But you know what's funny, during this competition I learned a lot more than in the previous three :-).\nPlease write whether this is a failure of the year, or someone was like that?\nAlways make your choice correctly and good luck to you all !!!😄😄😄",
    "1032447": "Out of interest, given you can choose 2 entries for your final score, why did you decide not to hedge your bets with one baseline submission and one of your own models?\n\nAgreed, this competition was a good one for learning new techniques, which (to me at least) is the more important thing.",
    "1032460": "haha we r in the same boat ! Like U, I learn alot from this competition (knowledge gained worth as much as a medal) 😏",
    "1032463": "> Out of interest, given you can choose 2 entries for your final score, why did you decide not to hedge your bets with one baseline submission and one of your own models?\n\n\nWe chose two different models in the hope that one would work. Negative experience is also an experience :-)",
    "1032935": "Its was right that the organizer baseline was enough but unfortunately i failed to believe that any of the model would work.Sorry to him a lot @aikhmelnytskyy",
    "1033002": "I was about to say that what @andypenrose said. \n\nBut again, it's all about learning. \nBetter next time 👍",
    "1033131": "Keep your head up. Two years ago I dropped from 107 -> 1207 in TalkingData. The most important part is A) take the KO like a champ and just get back up, don't let it bruise your ego, and B) learn from the L."
  },
  "source": "meta"
}