{
  "id": 304321,
  "title": "Using our algorithms for crypto trading",
  "url": "/competitions/g-research-crypto-forecasting/discussion/304321",
  "author_name": "",
  "post_date": "2022-01-31T17:54:26.779510Z",
  "votes": 2,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Some people may want to use their algorithms created in this competition to trade cryptocurrencies.    However, it is very unlikely that any of them will make money in the long run.  </p>\n<p>One strategy may be to figure out which coin has the greatest predicted target at each 15 minute interval and invest your money in that coin for 15 minutes.  If I did this, I would get around a .04% return on average each 15 minutes, which if compounded, would turn into a lot of money very quickly. </p>\n<p>The problem is that the transaction costs of the brokerage and the slippage fees entering and exiting the the position will likely be much greater than the .04% return which means the strategy will end up losing a bunch of money.  So, if you do try to use your algorithms, it might be a good idea to increase your time horizon to a couple hours to a day to minimize the number of transactions and their fees.  Or, you could try applying it to stocks with more liquidity than cryptocurrencies.</p>",
  "messages": [
    {
      "id": "1670645",
      "postDate": "01/31/2022 17:54:26",
      "content": "<p>Some people may want to use their algorithms created in this competition to trade cryptocurrencies.    However, it is very unlikely that any of them will make money in the long run.  </p>\n<p>One strategy may be to figure out which coin has the greatest predicted target at each 15 minute interval and invest your money in that coin for 15 minutes.  If I did this, I would get around a .04% return on average each 15 minutes, which if compounded, would turn into a lot of money very quickly. </p>\n<p>The problem is that the transaction costs of the brokerage and the slippage fees entering and exiting the the position will likely be much greater than the .04% return which means the strategy will end up losing a bunch of money.  So, if you do try to use your algorithms, it might be a good idea to increase your time horizon to a couple hours to a day to minimize the number of transactions and their fees.  Or, you could try applying it to stocks with more liquidity than cryptocurrencies.</p>",
      "rawMarkdown": "Some people may want to use their algorithms created in this competition to trade cryptocurrencies.    However, it is very unlikely that any of them will make money in the long run.  \n\nOne strategy may be to figure out which coin has the greatest predicted target at each 15 minute interval and invest your money in that coin for 15 minutes.  If I did this, I would get around a .04% return on average each 15 minutes, which if compounded, would turn into a lot of money very quickly. \n\nThe problem is that the transaction costs of the brokerage and the slippage fees entering and exiting the the position will likely be much greater than the .04% return which means the strategy will end up losing a bunch of money.  So, if you do try to use your algorithms, it might be a good idea to increase your time horizon to a couple hours to a day to minimize the number of transactions and their fees.  Or, you could try applying it to stocks with more liquidity than cryptocurrencies.",
      "votes": null
    },
    {
      "id": "1670780",
      "postDate": "01/31/2022 21:20:38",
      "content": "<p>Yeah of course</p>",
      "rawMarkdown": "Yeah of course",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1670780,
      "author_name": "pranayteaches",
      "author_url": "",
      "post_date": "01/31/2022 21:20:38",
      "content": "<p>Yeah of course</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1670645": "Some people may want to use their algorithms created in this competition to trade cryptocurrencies.    However, it is very unlikely that any of them will make money in the long run.  \n\nOne strategy may be to figure out which coin has the greatest predicted target at each 15 minute interval and invest your money in that coin for 15 minutes.  If I did this, I would get around a .04% return on average each 15 minutes, which if compounded, would turn into a lot of money very quickly. \n\nThe problem is that the transaction costs of the brokerage and the slippage fees entering and exiting the the position will likely be much greater than the .04% return which means the strategy will end up losing a bunch of money.  So, if you do try to use your algorithms, it might be a good idea to increase your time horizon to a couple hours to a day to minimize the number of transactions and their fees.  Or, you could try applying it to stocks with more liquidity than cryptocurrencies.",
    "1670780": "Yeah of course"
  },
  "source": "meta"
}