PDA

View Full Version : Partypoker STT Tournament Detection and ICM Adj



user324234
06-03-2020, 10:42 AM
Hi,

I would like to start a discussion on partypoker STT tournament detection, win/loss graph and ICM adjusted EV line. I am aware that many problems were caused by partypoker's new handhistory policy. I am simply trying to get the ball rolling here, to slowly get the essential features back, that many STT players desperately need.

1. Win/loss graph: Winnings in HM3 are far off. There seems to be a systematic error, by which HM3 detects WAY too many losses. This results in the overall graph dropping by tens- to hundredthousands of dollars. However, HM3 appears to detect SOME winnings correctly, which can be seen by the small "upwards" bumps in the graph (I have attached a win/loss graph).

2. ICM Adj graph: Again, much like the actual win/loss graph, the ICM Adj Winnings are far below reality. There is an obvious difference between the ICM and win/loss graph, so SOME correct calculation must be going on. But the overall bias towards losses messes everything up.

After digging through the list of tournaments, I have noticed that a large chunk of the losses comes from Double-Or-Nothing (DON) SNGs. These are 6-max STTs with top3 players paid. The handhistory naturally ends at 3 players, in which case Hero successfully cashed. But HM3 awards $0 in winnings here. It seems that the partypoker handhistory offers no information about the payout structure of the tournament, and thus HM3 assumes every 6-max STT to have the "standard" 63/35 payout structure. Am I correct in this assumption? This could be solved by allowing the user to manually set a payout structure for a tournament (this was possible in PT4).

3. Number of Entrants: HM3 detects the wrong number of entrants for a lot of STTs. Most are correctly detected at 6. Many are incorrectly detected as 7. Some are incorrectly detected as 8,9,10,11,12 or 13. I have looked through the handhistories of tournaments, where the wrong number of entrants was detected, but I could not find anything which might have caused this wrong detection. For example, one tournament where HM3 detected 13 entrants, literally says "Total number of players : 6/6" in the HH of the first hand.



Best regards

user324234
06-03-2020, 10:44 AM
Hi,

I would like to start a discussion on partypoker STT tournament detection, win/loss graph and ICM adjusted EV line. I am aware that many problems were caused by partypoker's new handhistory policy. I am simply trying to get the ball rolling here, to slowly get the essential features back, that many STT players desperately need.

1. Win/loss graph: Winnings in HM3 are far off. There seems to be a systematic error, by which HM3 detects WAY too many losses. This results in the overall graph dropping by tens- to hundredthousands of dollars. However, HM3 appears to detect SOME winnings correctly, which can be seen by the small "upwards" bumps in the graph (I have attached a win/loss graph).

2. ICM Adj graph: Again, much like the actual win/loss graph, the ICM Adj Winnings are far below reality. There is an obvious difference between the ICM and win/loss graph, so SOME correct calculation must be going on. But the overall bias towards losses messes everything up.

After digging through the list of tournaments, I have noticed that a large chunk of the losses comes from Double-Or-Nothing (DON) SNGs. These are 6-max STTs with top3 players paid. The handhistory naturally ends at 3 players, in which case Hero successfully cashed. But HM3 awards $0 in winnings here. It seems that the partypoker handhistory offers no information about the payout structure of the tournament, and thus HM3 assumes every 6-max STT to have the "standard" 63/35 payout structure. Am I correct in this assumption? This could be solved by allowing the user to manually set a payout structure for a tournament (this was possible in PT4).

3. Number of Entrants: HM3 detects the wrong number of entrants for a lot of STTs. Most are correctly detected at 6. Many are incorrectly detected as 7. Some are incorrectly detected as 8,9,10,11,12 or 13. I have looked through the handhistories of tournaments, where the wrong number of entrants was detected, but I could not find anything which might have caused this wrong detection. For example, one tournament where HM3 detected 13 entrants, literally says "Total number of players : 6/6" in the HH of the first hand.



Best regards



164343

fozzy71
06-04-2020, 09:39 AM
We have a number of open issues for our developers to work on related to the party downloaded tournament hands. Some of those issues can be resolved while others may not be able to since the HH files do not have that info, and they do not write separate tournament summary files for each tournament. We are working to try and improve our tournament detection system (https://support.holdemmanager.com/support/faqView/Holdem-Manager-3/233/Tournament-Detection-%28Automatic-%26-Manual%29) to address some of these issues, if possible.

For now you will need to manually edit the results via the Reports - Tournaments - Tournaments - right-click a tournament in the upper half the report grid - Edit - Tournament Editor, then edit as needed and click 'OK' to save the edits.

Note: If you make a post on our forums and do not see it right away, please do not post it again and again. Our forums have automated moderation for possibly malicious posts and your multiple threads all went into the moderation queue to be manually approved

ZenRN
09-04-2021, 11:57 PM
Hi,
Sorry if this is dumb question, I am still trying to figure out if it is worth getting HM3 given the issues with PP anonymized downloads.
I usually only play 2 sometimes 3 tables at a time in the STT format. It is relatively easy for me to know who is sitting where and recording this information.
It is mentioned above that you can edit the results manually. Does this include replacing the 'playerx' anonymization with their actual screen name? If so, and if done consistently, would this generate useful stats on those players?
Thanks for you assistance

fozzy71
09-05-2021, 12:13 PM
Hi,
Sorry if this is dumb question, I am still trying to figure out if it is worth getting HM3 given the issues with PP anonymized downloads.
I usually only play 2 sometimes 3 tables at a time in the STT format. It is relatively easy for me to know who is sitting where and recording this information.
It is mentioned above that you can edit the results manually. Does this include replacing the 'playerx' anonymization with their actual screen name? If so, and if done consistently, would this generate useful stats on those players?
Thanks for you assistance

No, there is no way to edit player names in the HM3 user interface after they have been imported. The simplest way would be to open the hand history files in a text editor and edit the names in the files (using Find/Replace) before you import them.

It is actually possible to edit the 'players' name data in the database itself with a 3rd party database browser (https://sqlitebrowser.org/) but it is not something we generally recommend as it is possible to corrupt the database if you are not familiar with the DB Browser software and make a mistake. Here is a short video example (https://www.dropbox.com/s/7hkdepmdtrzmruz/party%20database%20edit%20players%20name.mp4?dl=0) of how to do it if you want to give it a try with one of your databases.

ZenRN
09-05-2021, 12:56 PM
Thanks so much for the quick reply, I will check out the software you mentioned.
If anyone else has a way to speed up editing software names I would appreciate it.
On of the challenges I foresee is that, since I play more than 1 STT at a time, and the hands are downloaded in chronological order, the tourneys are mixed together in the text file. Disentangling them is simple but time consuming and begging for automation. Sadly, I am not blessed with contemporary programming skills. (my last experience was with FORTRAN). Again, any suggestions would be very much appreciated.
I would also be willing to pay a small fee for the algorithm, if that would incentivize folks.
Cheers!