{
  "id": 1040,
  "title": "RMSE",
  "url": "/competitions/AlgorithmicTradingChallenge/discussion/1040",
  "author_name": "",
  "post_date": "2011-11-13T09:40:15.490Z",
  "votes": null,
  "comment_count": 8,
  "views": 4300,
  "content": "<p>Is the RMSE calculated as</p>\r\n<p>sqrt(mean((y-p)^2)) where y is actual and p is prediction or</p>\r\n<p>sqrt(mean(mean((y-p)^2))) where we take mean of a row first?</p>",
  "messages": [
    {
      "id": "6503",
      "postDate": "11/13/2011 09:40:15",
      "content": "<p>Is the RMSE calculated as</p>\r\n<p>sqrt(mean((y-p)^2)) where y is actual and p is prediction or</p>\r\n<p>sqrt(mean(mean((y-p)^2))) where we take mean of a row first?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "6576",
      "postDate": "11/17/2011 21:14:21",
      "content": "<p>Also, I wasn't able to find any info on if these are absolute or percentage errors.&nbsp; If absolute, wouldn't that make it biased towards getting the expensive stocks correct?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "6589",
      "postDate": "11/18/2011 11:07:04",
      "content": "<p>The S in RMSE makes it absolute. RMSE is a bit flawed as 'bigger' stocks will see higher fluctuations.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "6593",
      "postDate": "11/18/2011 17:36:35",
      "content": "<p>[quote=Dirk Nachbar;6589]</p>\r\n<p>The S in RMSE makes it absolute.[/quote]</p>\r\n<p>I could argue that you could have squared percentage error and that would actually be preferable for getting a good algorithm you'd want to use in the real world.&nbsp; If I'm off by 20 on a stock that is priced at 2000, that\r\n<em>should</em> be the same as being off by 2 on a stock of 200, yet in this contest it is 10x worse.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "6595",
      "postDate": "11/18/2011 18:16:46",
      "content": "<p>Answering original questions : both formulas will produce the same result.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "6597",
      "postDate": "11/18/2011 19:58:12",
      "content": "<p>[quote=Ty S.;6593]</p>\r\n<p>I could argue that you could have squared percentage error and that would actually be preferable for getting a good algorithm you'd want to use in the real world.&nbsp; If I'm off by 20 on a stock that is priced at 2000, that\r\n<em>should</em> be the same as being off by 2 on a stock of 200, yet in this contest it is 10x worse.</p>\r\n<p>[/quote]</p>\r\n<p>It really isn't the same - which would you prefer to lose - $2 or $20?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "6606",
      "postDate": "11/19/2011 01:43:28",
      "content": "<p>It is the same if the same amount of money is invested. For investor there is no difference between one stock for 2000 and 10 stocks for 200 each.\r\n</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "6624",
      "postDate": "11/20/2011 07:26:24",
      "content": "<p>Is this how people are calculating RMSE?</p>\r\n<p>BidRMSE (row i) = sqrt(sum N = 51 to 100 (FcastBid<em>N - ActualBid</em>N) ^2)/50</p>\r\n<p>AskRMSE (row i) = sqrt(sum N = 51 to 100 (FcastAsk<em>N - ActualBid</em>N) ^2)/50</p>\r\n<p>Overall RMSE = sum (all rows) (BidRMSE &#43; AskRMSE)/2</p>\r\n<p>That is how I interpreted it i.e. average RMSE on Bid and Ask, separately.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "6625",
      "postDate": "11/20/2011 07:27:21",
      "content": "<p>Er... last line should be average not sum</p>\r\n<p>You know what I mean... the average RMSE over all rows.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 6576,
      "author_name": "trojanrabbit",
      "author_url": "",
      "post_date": "11/17/2011 21:14:21",
      "content": "<p>Also, I wasn't able to find any info on if these are absolute or percentage errors.&nbsp; If absolute, wouldn't that make it biased towards getting the expensive stocks correct?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 6589,
      "author_name": "dirknbr",
      "author_url": "",
      "post_date": "11/18/2011 11:07:04",
      "content": "<p>The S in RMSE makes it absolute. RMSE is a bit flawed as 'bigger' stocks will see higher fluctuations.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 6593,
      "author_name": "trojanrabbit",
      "author_url": "",
      "post_date": "11/18/2011 17:36:35",
      "content": "<p>[quote=Dirk Nachbar;6589]</p>\r\n<p>The S in RMSE makes it absolute.[/quote]</p>\r\n<p>I could argue that you could have squared percentage error and that would actually be preferable for getting a good algorithm you'd want to use in the real world.&nbsp; If I'm off by 20 on a stock that is priced at 2000, that\r\n<em>should</em> be the same as being off by 2 on a stock of 200, yet in this contest it is 10x worse.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 6595,
      "author_name": "ccccat",
      "author_url": "",
      "post_date": "11/18/2011 18:16:46",
      "content": "<p>Answering original questions : both formulas will produce the same result.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 6597,
      "author_name": "sirguessalot",
      "author_url": "",
      "post_date": "11/18/2011 19:58:12",
      "content": "<p>[quote=Ty S.;6593]</p>\r\n<p>I could argue that you could have squared percentage error and that would actually be preferable for getting a good algorithm you'd want to use in the real world.&nbsp; If I'm off by 20 on a stock that is priced at 2000, that\r\n<em>should</em> be the same as being off by 2 on a stock of 200, yet in this contest it is 10x worse.</p>\r\n<p>[/quote]</p>\r\n<p>It really isn't the same - which would you prefer to lose - $2 or $20?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 6606,
      "author_name": "ccccat",
      "author_url": "",
      "post_date": "11/19/2011 01:43:28",
      "content": "<p>It is the same if the same amount of money is invested. For investor there is no difference between one stock for 2000 and 10 stocks for 200 each.\r\n</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 6624,
      "author_name": "zenbot",
      "author_url": "",
      "post_date": "11/20/2011 07:26:24",
      "content": "<p>Is this how people are calculating RMSE?</p>\r\n<p>BidRMSE (row i) = sqrt(sum N = 51 to 100 (FcastBid<em>N - ActualBid</em>N) ^2)/50</p>\r\n<p>AskRMSE (row i) = sqrt(sum N = 51 to 100 (FcastAsk<em>N - ActualBid</em>N) ^2)/50</p>\r\n<p>Overall RMSE = sum (all rows) (BidRMSE &#43; AskRMSE)/2</p>\r\n<p>That is how I interpreted it i.e. average RMSE on Bid and Ask, separately.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 6625,
      "author_name": "zenbot",
      "author_url": "",
      "post_date": "11/20/2011 07:27:21",
      "content": "<p>Er... last line should be average not sum</p>\r\n<p>You know what I mean... the average RMSE over all rows.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "6503": "",
    "6576": "",
    "6589": "",
    "6593": "",
    "6595": "",
    "6597": "",
    "6606": "",
    "6624": "",
    "6625": ""
  },
  "source": "meta"
}