PokerTracker: The Next Generation (NEW SOFTWARE DISCUSSION)

[PokerAce]

[ QUOTE ]
[ QUOTE ]
All of these situations are basically neglible in the grand scheme of your stats. That is why I prefer to stick with the simple rule that PT currently uses. Won money, won hand. Lose money, lost hand.

[/ QUOTE ]


alright, well thanks for asking for opinions when you already had your mind made up.

[/ QUOTE ]

The decision hasn't been made. Aperfect10 and I don't agree in this matter. He thinks lose money = lose hand. I believe that the hand was still won, though you lost money. That's the reason I posted the poll.

[yellowbastard]

How about cumulative aggression factor stats and other stats in the "Summary" tab to be able to better compair the play at different sites? The current version only gives PFR, VP$IP, and BB/100.

[tiger_hall]

be able to combine alias' as i cannot find how to do it...
make it easy

[CudjoeBill]

tiger_hall,

In PT -> Ring Game Player Statistics, the General Info tab. Use the Aliases button to define them.

CB

[HighSteaks]

I don't know if this had been mentioned but some kind of miniview that includes important info would be useful for playing and also for video. Whenever someone makes a video they generally go through a clumsy segment at the end getting up stats for dicussion and hands for replaying. So perhaps around the size of 1/4 of the screen so you can play 3 tables and have it in the other. Perhaps it could identify your tables as 1,2 and 3 and you can click 1,2 or 3 to view info on that table. Might be able to make the pop up stats redundant also.

[pankwindu]

[ QUOTE ]
The decision hasn't been made. Aperfect10 and I don't agree in this matter. He thinks lose money = lose hand. I believe that the hand was still won, though you lost money. That's the reason I posted the poll.

[/ QUOTE ]
Seems to me it's a matter of philosophy - are you tracking your PLAY or are you tracking your MONEY? IMO it's the former, and if the play of the hand resulted in you being awarded part of the pot, it's a won hand regardless of ancillary matters like the exact amount of money and/or rake. Actual "lost money" hands can be obtained/counted by the user in other ways (or at least should be able to).

Edit: Also, under the "lose money = lose hand" philosophy, the game level and rake structure can cause the exact same hands/cards/action to count as different outcomes in different games. That doesn't make sense to me. According to the rules of poker you were awarded a pot, to me that's a won hand no matter the environment. The money aspect belongs in a different column or stat or calculation or whatever.

[w_alloy]

I would like PT3 to have more advanced statistical tools. Some things I have in mind (disclaimer: I don't know much about advanced statistics but it seems like this stuff wouldn't be that hard to implement from my limited knowledge):

-Calculating confidence intervals, especially with win rate, but also with things like VPIP. Some research or consulting work might need to be done to find the exact right math for the shape of the binomial distribution. It may or may not help to obtain large 10mil+ dbs from 3rd parties and do research on the shapes of distributions. It is my *impression* that simple standard deviation is not enough unless very large sample sizes are used, because most poker stats are not normally distributed. Even if, for example, the $ won from samples of 100 hands is normally distributed (the only stat that std deviation is currently given for in pt, it should be at least close to normally distributed because of central limit theorem, right?), the population of players' winrates is not, which affects any confidence intervals. Hopefully you can see what I am getting at? Parts of this were discussed in Math of Poker.

-Optionally incorperating a Bayesian "certainty" rating in the HUD so that users would have a better idea when and how much to trust certain stats. This wouldn't be too usefull for experienced users (at least ones who have correct ideas about these numbers already) but would help new users not as familiar with statistical significance. One thing that would be particularly useful with this is figuring out a simple way to graphically represent the significance of a stat. There may be better uses for color (like how it is currently used to help quickly differentiate between tight and loose) but things like the slant of letters (the degree of italics), or different fonts, or transperancy, etc. could be implemented to quickly inform the user and save screen space. IDK how you want to calculate this confidence, but I'm sure it would relate to what I wrote above.

-(less important suggestion) You could use these tools, combined with some kind of semi-random sampling (probably of consecutive hands on a certain table and consecutive hands across all tables), to give a measure for the degree an opponent is likely varying his play across a specific game. For example, if a player plays 300 hands at 20 vpip, and then 400 hands at 30 vpip, his variation rating will be high. Of course, this could also be variance, but if you have a good understanding of how likely this is to be caused by variance, over largish sample sizes, players who spend effort to change gears and adapt could be easily singled out. To compensate for the time it would take to calculate this for each player, you could have it be an optional computation and/or set a minimum number of hands needed to trigger this.

-At the very least, include a short (or even long) statistics tutorial for people who have not taken or have forgotten the relevant parts of college level statistics (which is a requirement to get the max out of the program, imo).

-Given calculations from my first suggestion (this would be especially useful if a distribution of player winrates was discovered), calculate risk of ruin of continuing to play a specific game/level given past performance and a user input bankroll (and, obviously, the stipulation that said player does not move down, which the user would have to be aware of).

There are other improvements along the lines of some I suggested to move PT into the "poker career management" software category, which is something that is not presently available to this extent (to my knowledge) and judging from posts on 2p2 there is certainly a market for it. PT already helps decide when to move up, or down, or when to keep playing, etc., but it does this in the most basic way possible and does not help one interpret results (which are often very misleading due to people not understanding variance).

The last thing I want to suggest is possibly changing your business model and adding to your team as others have suggested. I know I would be willing to dish out 10 times or more what I paid for PT to have the functionality I want (a lot of other things needed improved have been detailed in this thread), and I would like this much sooner than seems possible with a 3 person team. I'm sure there are others who are like me, and some have posted in this thread.

I only wrote out this post because I had nothing to do while I was importing my hands into HM beta. I love PT and PA but currently as promised HM is a huge improvement. Implementing some of these suggestions would leapfrog HM and make PT3 vital for serious players.

[baronzeus]

if you wont put them in a separate tab can you at least keep a COUNT of split pots somewhere, and read them all as won pots? this would make things a lot easier to deal with. that way i wouldnt have to manually remember all the post that got split, whether i won or lost money, and then remove those from my showdown numbers

[ReptileHouse]

I would love to see a more normalized (from the technical definition of the term) database. Right now, things are a bit scattered, making it difficult to interface to the database.

Standalone hand importer for Postgres databases. I like to run PT on a different computer from the one I'm using to play. Right now, I use a network share of the HH directory to let PT see the HH files and import them, but that's awkward. I'd prefer to have a standalone importer utility that I can run. This would obviously be less feature rich than the built-in importer (e.g., no auto-rate while importing).

Similarly, for PAHUD, it's a bit of a pain right now to set up to connect to a remote database. It works great once configured, but getting to that point is not easy.

My ideal setup to have a DB server running remotely. I play poker with an importer and a HUD running. At the end of the session, I fire up PT itself to review the session. While I'm playing, I don't need to have PT running at all.

[Bill Haywood]

If we could rearrange the stat layout right at the table we are playing on, that would be awesome. Failing that, if the table in the layout manager were the same size as the one you are playing on, it would make it MUCH easier to arrange things where you want them. Usually the scale is off, so things are forever not in just the right place.