{
  "id": 1312,
  "title": "Trade Times",
  "url": "/competitions/benchmark-bond-trade-price-challenge/discussion/1312",
  "author_name": "",
  "post_date": "2012-01-28T18:16:44.157Z",
  "votes": null,
  "comment_count": 3,
  "views": 2434,
  "content": "<p>I have a question about the received_time_diff_last{1-10} parameters. These are the times between the actual trade time of the trade we are trying to precict, and the actual trade times of the last 10 trades. These are not trade reporting times. Is that\r\n right? In other words, to calculate the time between the the trade and any of the previous 10 trades, I only need the received_time_diff_last parameters, and I do not need to use the reporting_delay parameter. Correct?</p>\r\n<p>Out of curiosity, does anyone see much value to the reporting_delay parameter? If we are just trying to predict the trade price, what difference does it make how long it took before the trade was reported after it had already occured. I suppose it's possible\r\n there is a correlation between certain types of trades and their reporting delay, but this seems unlikely to me.</p>",
  "messages": [
    {
      "id": "8315",
      "postDate": "01/28/2012 18:16:44",
      "content": "<p>I have a question about the received_time_diff_last{1-10} parameters. These are the times between the actual trade time of the trade we are trying to precict, and the actual trade times of the last 10 trades. These are not trade reporting times. Is that\r\n right? In other words, to calculate the time between the the trade and any of the previous 10 trades, I only need the received_time_diff_last parameters, and I do not need to use the reporting_delay parameter. Correct?</p>\r\n<p>Out of curiosity, does anyone see much value to the reporting_delay parameter? If we are just trying to predict the trade price, what difference does it make how long it took before the trade was reported after it had already occured. I suppose it's possible\r\n there is a correlation between certain types of trades and their reporting delay, but this seems unlikely to me.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "8316",
      "postDate": "01/28/2012 18:33:11",
      "content": "<p>I just signed up for Kaggle and thinking about doing this competition, but I read your question. I don't have a great answer for you, but I would think reporting time would have some correlation with the size of the trade, just as you would have with stocks,\r\n but I'm not that familiar with market micro-structure of the secondary bond market.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "8325",
      "postDate": "01/28/2012 23:00:35",
      "content": "<p>You are correct. Reporting delay has nothing to do with the previous trades.</p>\r\n<p>As to the second question, market participants have 15 minutes to report trades (in general) There may be some correlation between how long they take to do so. For instance maybe after a customer trade a dealer wants to flip their purchase to another dealer\r\n and they delay reporting until the trade is made. I'm not saying this is the only or even the most common reason for a delay. I'm just offering an example of why you might find some meaning in this field.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "8402",
      "postDate": "02/02/2012 10:23:52",
      "content": "<p>I had already asked both of these <a href=\"http://www.kaggle.com/c/benchmark-bond-trade-price-challenge/forums/t/1315/clarifications-of-received-time-diff-reporting-delay-and-curve-based-price\">\r\nhere</a></p>\r\n<p>and the answers seem to be:</p>\r\n<p>&gt;Trading time differences are as defined by received_time_diff_last{1..10} parameters, and I do not need to use the reporting_delay parameter to offset them . Correct?</p>\r\n<p>Yes I believe so.</p>\r\n<p>&nbsp;</p>\r\n<p>&gt;Out of curiosity, does anyone see much value to the reporting_delay parameter? If we are just trying to predict the trade price, what difference does it make how long it took before the trade was reported after it had already occured. I suppose it's possible\r\n there is a correlation between certain types of trades and their reporting delay, but this seems unlikely to me.</p>\r\n<p>No it seems to be Just Another Feature, and a pretty insignificant, noisy and dirty one at that. I guess you could try to answer those questions by developing a submodel to predict reporting delay by trade_size, type etc., but I don't see much point in wasting\r\n effort on that.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 8316,
      "author_name": "alphag77",
      "author_url": "",
      "post_date": "01/28/2012 18:33:11",
      "content": "<p>I just signed up for Kaggle and thinking about doing this competition, but I read your question. I don't have a great answer for you, but I would think reporting time would have some correlation with the size of the trade, just as you would have with stocks,\r\n but I'm not that familiar with market micro-structure of the secondary bond market.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 8325,
      "author_name": "danglaser",
      "author_url": "",
      "post_date": "01/28/2012 23:00:35",
      "content": "<p>You are correct. Reporting delay has nothing to do with the previous trades.</p>\r\n<p>As to the second question, market participants have 15 minutes to report trades (in general) There may be some correlation between how long they take to do so. For instance maybe after a customer trade a dealer wants to flip their purchase to another dealer\r\n and they delay reporting until the trade is made. I'm not saying this is the only or even the most common reason for a delay. I'm just offering an example of why you might find some meaning in this field.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 8402,
      "author_name": "smcinerney",
      "author_url": "",
      "post_date": "02/02/2012 10:23:52",
      "content": "<p>I had already asked both of these <a href=\"http://www.kaggle.com/c/benchmark-bond-trade-price-challenge/forums/t/1315/clarifications-of-received-time-diff-reporting-delay-and-curve-based-price\">\r\nhere</a></p>\r\n<p>and the answers seem to be:</p>\r\n<p>&gt;Trading time differences are as defined by received_time_diff_last{1..10} parameters, and I do not need to use the reporting_delay parameter to offset them . Correct?</p>\r\n<p>Yes I believe so.</p>\r\n<p>&nbsp;</p>\r\n<p>&gt;Out of curiosity, does anyone see much value to the reporting_delay parameter? If we are just trying to predict the trade price, what difference does it make how long it took before the trade was reported after it had already occured. I suppose it's possible\r\n there is a correlation between certain types of trades and their reporting delay, but this seems unlikely to me.</p>\r\n<p>No it seems to be Just Another Feature, and a pretty insignificant, noisy and dirty one at that. I guess you could try to answer those questions by developing a submodel to predict reporting delay by trade_size, type etc., but I don't see much point in wasting\r\n effort on that.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "8315": "",
    "8316": "",
    "8325": "",
    "8402": ""
  },
  "source": "meta"
}