{
  "id": 567309,
  "title": "March 9 update: crash from 149 to 1529 position",
  "url": "/competitions/jane-street-real-time-market-data-forecasting/discussion/567309",
  "author_name": "",
  "post_date": "2025-03-09T15:30:11.087712800Z",
  "votes": 4,
  "comment_count": 8,
  "views": 0,
  "content": "<p>Are there many people in a similar situation?</p>",
  "messages": [
    {
      "id": "3145255",
      "postDate": "03/09/2025 15:30:11",
      "content": "<p>Are there many people in a similar situation?</p>",
      "rawMarkdown": "Are there many people in a similar situation?",
      "votes": null
    },
    {
      "id": "3145416",
      "postDate": "03/09/2025 19:00:43",
      "content": "<p>Same here. 42th to 1329th</p>",
      "rawMarkdown": "Same here. 42th to 1329th",
      "votes": null
    },
    {
      "id": "3145472",
      "postDate": "03/09/2025 21:03:43",
      "content": "<p>The drop in ranking has been bigger than in score, it seems that there are good public kernels that have been widely used. For example, score 0.008004: 178-561 places, score 0.007704: 956-1146 places and so on.</p>\n<p>From my submissions (both offline) I would rule out (I think but…) any problem with the data or the rescoring process. One of them (only nnet) has remained stable in scoring and the other, which added xgboost to form an ensemble, has dropped. So it seems that it was the performance of xgboost that caused the drop.</p>\n<p>Good luck in what remains!</p>",
      "rawMarkdown": "The drop in ranking has been bigger than in score, it seems that there are good public kernels that have been widely used. For example, score 0.008004: 178-561 places, score 0.007704: 956-1146 places and so on.\n\nFrom my submissions (both offline) I would rule out (I think but...) any problem with the data or the rescoring process. One of them (only nnet) has remained stable in scoring and the other, which added xgboost to form an ensemble, has dropped. So it seems that it was the performance of xgboost that caused the drop.\n\nGood luck in what remains!",
      "votes": null
    },
    {
      "id": "3145748",
      "postDate": "03/10/2025 08:17:48",
      "content": "<p>In my case, an offline tree model stayed stable while my online learning NN started to perform poorly. It may be due to an issue during online learning or some of my assumptions like fixed number of timestamps.</p>",
      "rawMarkdown": "In my case, an offline tree model stayed stable while my online learning NN started to perform poorly. It may be due to an issue during online learning or some of my assumptions like fixed number of timestamps.",
      "votes": null
    },
    {
      "id": "3173674",
      "postDate": "04/08/2025 07:54:43",
      "content": "<p>I don't know what is happening but I am back to life.</p>",
      "rawMarkdown": "I don't know what is happening but I am back to life.",
      "votes": null
    },
    {
      "id": "3173682",
      "postDate": "04/08/2025 08:03:26",
      "content": "<p>Probably you predicted the Trump tariffs months ago? If so, will not be surprised if you climb to top in following updates.😂</p>",
      "rawMarkdown": "Probably you predicted the Trump tariffs months ago? If so, will not be surprised if you climb to top in following updates.😂",
      "votes": null
    },
    {
      "id": "3173706",
      "postDate": "04/08/2025 08:26:35",
      "content": "<p>Yeah, I saw it and was pleasantly surprised because it gives me hope too!</p>\n<p>My rank hasn't increased much, but I'm not far from being reborn in terms of score. My xgboost component grew well, and so did my nnet, although less so.</p>",
      "rawMarkdown": "Yeah, I saw it and was pleasantly surprised because it gives me hope too!\n\nMy rank hasn't increased much, but I'm not far from being reborn in terms of score. My xgboost component grew well, and so did my nnet, although less so.",
      "votes": null
    },
    {
      "id": "3173987",
      "postDate": "04/08/2025 14:58:59",
      "content": "<p>I think the tariffs effects should be reflected in the next update</p>",
      "rawMarkdown": "I think the tariffs effects should be reflected in the next update",
      "votes": null
    },
    {
      "id": "3174007",
      "postDate": "04/08/2025 15:21:32",
      "content": "<p>Yes, but during this last update I believe there has also been movement (volatility+) as a result of rumors about tariff policy.</p>",
      "rawMarkdown": "Yes, but during this last update I believe there has also been movement (volatility+) as a result of rumors about tariff policy.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3145416,
      "author_name": "aerdem4",
      "author_url": "",
      "post_date": "03/09/2025 19:00:43",
      "content": "<p>Same here. 42th to 1329th</p>",
      "votes": null,
      "replies": [
        {
          "id": 3145472,
          "author_name": "coreacasa",
          "author_url": "",
          "post_date": "03/09/2025 21:03:43",
          "content": "<p>The drop in ranking has been bigger than in score, it seems that there are good public kernels that have been widely used. For example, score 0.008004: 178-561 places, score 0.007704: 956-1146 places and so on.</p>\n<p>From my submissions (both offline) I would rule out (I think but…) any problem with the data or the rescoring process. One of them (only nnet) has remained stable in scoring and the other, which added xgboost to form an ensemble, has dropped. So it seems that it was the performance of xgboost that caused the drop.</p>\n<p>Good luck in what remains!</p>",
          "votes": null,
          "replies": [
            {
              "id": 3145748,
              "author_name": "aerdem4",
              "author_url": "",
              "post_date": "03/10/2025 08:17:48",
              "content": "<p>In my case, an offline tree model stayed stable while my online learning NN started to perform poorly. It may be due to an issue during online learning or some of my assumptions like fixed number of timestamps.</p>",
              "votes": null,
              "replies": []
            }
          ]
        },
        {
          "id": 3173674,
          "author_name": "aerdem4",
          "author_url": "",
          "post_date": "04/08/2025 07:54:43",
          "content": "<p>I don't know what is happening but I am back to life.</p>",
          "votes": null,
          "replies": [
            {
              "id": 3173682,
              "author_name": "lihaorocky",
              "author_url": "",
              "post_date": "04/08/2025 08:03:26",
              "content": "<p>Probably you predicted the Trump tariffs months ago? If so, will not be surprised if you climb to top in following updates.😂</p>",
              "votes": null,
              "replies": [
                {
                  "id": 3173987,
                  "author_name": "redfoongus",
                  "author_url": "",
                  "post_date": "04/08/2025 14:58:59",
                  "content": "<p>I think the tariffs effects should be reflected in the next update</p>",
                  "votes": null,
                  "replies": [
                    {
                      "id": 3174007,
                      "author_name": "coreacasa",
                      "author_url": "",
                      "post_date": "04/08/2025 15:21:32",
                      "content": "<p>Yes, but during this last update I believe there has also been movement (volatility+) as a result of rumors about tariff policy.</p>",
                      "votes": null,
                      "replies": []
                    }
                  ]
                }
              ]
            },
            {
              "id": 3173706,
              "author_name": "coreacasa",
              "author_url": "",
              "post_date": "04/08/2025 08:26:35",
              "content": "<p>Yeah, I saw it and was pleasantly surprised because it gives me hope too!</p>\n<p>My rank hasn't increased much, but I'm not far from being reborn in terms of score. My xgboost component grew well, and so did my nnet, although less so.</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3145255": "Are there many people in a similar situation?",
    "3145416": "Same here. 42th to 1329th",
    "3145472": "The drop in ranking has been bigger than in score, it seems that there are good public kernels that have been widely used. For example, score 0.008004: 178-561 places, score 0.007704: 956-1146 places and so on.\n\nFrom my submissions (both offline) I would rule out (I think but...) any problem with the data or the rescoring process. One of them (only nnet) has remained stable in scoring and the other, which added xgboost to form an ensemble, has dropped. So it seems that it was the performance of xgboost that caused the drop.\n\nGood luck in what remains!",
    "3145748": "In my case, an offline tree model stayed stable while my online learning NN started to perform poorly. It may be due to an issue during online learning or some of my assumptions like fixed number of timestamps.",
    "3173674": "I don't know what is happening but I am back to life.",
    "3173682": "Probably you predicted the Trump tariffs months ago? If so, will not be surprised if you climb to top in following updates.😂",
    "3173706": "Yeah, I saw it and was pleasantly surprised because it gives me hope too!\n\nMy rank hasn't increased much, but I'm not far from being reborn in terms of score. My xgboost component grew well, and so did my nnet, although less so.",
    "3173987": "I think the tariffs effects should be reflected in the next update",
    "3174007": "Yes, but during this last update I believe there has also been movement (volatility+) as a result of rumors about tariff policy."
  },
  "source": "meta"
}