How many hands do you import a second?

  • 4 replies
    • PokerTracker
      PokerTracker
      Gold
      Joined: 07.06.2011 Posts: 643
      Originally posted by Soccout8
      Im currently at around 75/s. Is this too slow or nah?

      Im not sure if i should improve it
      .75 seconds is very reasonable. PokerTracker must perform a tremendous amount of post-import processing, this is necessary to populate data such as autonotes. Import is optimised to work ideally during live play, the # of hands/second typically is only noticed during the first import - from then on its not something that will affect you.
    • andreibalint
      andreibalint
      Bronze
      Joined: 11.04.2009 Posts: 872
      I was running at 4hands/sec from a HDD and 40hands/sec from a SSD, so if it's too slow a SSD really is 10x faster.
    • MrMardyBum
      MrMardyBum
      Bronze
      Joined: 14.03.2009 Posts: 2,206
      70 - 200 hands per second... Dependant on format - Best/Quickest way if you already have them in a database and wish to move to another is to export them, then reimport to where you want them. As it's then saved in a format that your tracker can scan/import data quicker (at least for me).

      I always export my hands (every 10k or so) because poker trackers in general are far from optimal and often get corrupted and/or need to reimport hands due to misinformation etc... Sometime I export them then import them back to the same database and the results change... So I dunno, lets hope the PT and HM merger is a success then we won't have to worry about reliability as much hopefully.

      Also, same databases show different winrates for me in PT4 and HM2... I am 3 BI better off according to PT4 over only a 40k sample. And I have no way of knowing which one is correct. Unless I read each HH individually and then cross reference via my trackers... But there is no way I care that much...

      Bring back the elephant imo!!! :s_evil:
    • PokerTracker
      PokerTracker
      Gold
      Joined: 07.06.2011 Posts: 643
      Originally posted by MrMardyBum
      70 - 200 hands per second... Dependant on format - Best/Quickest way if you already have them in a database and wish to move to another is to export them, then reimport to where you want them. As it's then saved in a format that your tracker can scan/import data quicker (at least for me).
      Every site requires different processing time, tournaments take much more time than cash games. The more players at the table, the more processing time is required, and if there are all-ins then PT4 must do some serious number crunching. All this adds up to more time. As poker becomes more advanced, processing time has actually increased, not decreased.... the only serious speed bump in modern times was the advent of the SSD drive and more memory.... new processors have only a small effect on hand import speeds.

      I always export my hands (every 10k or so) because poker trackers in general are far from optimal and often get corrupted and/or need to reimport hands due to misinformation etc... Sometime I export them then import them back to the same database and the results change... So I dunno, lets hope the PT and HM merger is a success then we won't have to worry about reliability as much hopefully.


      As new updates are released, we sometimes advise deleting and then re-importing specific hands. Our staff will always advise when re-importing is necessary - but this rarely affects win rates; usually it is due to improvements in the methods used to calculate a stat.

      Tournaments are different from cash games however, if there is no tournament summary available then we must rely on our tournament detection engine which estimates winnings and prize pools... but this database is not foolproof. As new database entries are added, you might discover these differences after a reimport. That is why we always advise manually checking your tournament results after every session to make sure the database is correct, and if it is not you can manually edit the results to match reality.

      Also, same databases show different winrates for me in PT4 and HM2... I am 3 BI better off according to PT4 over only a 40k sample. And I have no way of knowing which one is correct. Unless I read each HH individually and then cross reference via my trackers... But there is no way I care that much...


      See above. Also we should note that HM2 and PT4 does not calculate expectation in the same way, there are some differences in results. In future versions I assume we will unify this as we integrate back-end technologies.