New Software + Central Player Database (site by site basis)

Two_Slick · #1 07-06-2005, 01:05 AM

I'm working on some new software for PP + skins that does a lot of what PT does real-time (hoping to expand to other systems later). In addition to statistical analysis, it will flag questionable plays by other players and display them. I've even considered modifying the program to offer real-time advice on hands based on the stake, number of players, and player tendencies (although this would be quiet an undertaking). The idea has occurred to me to offer a centralized database for the major sites offering mined data for a very small monthly fee. The data would obviously be more comprehensive than any 1 player's mined database.

I'm mainly looking for:

1) Feedback on what you think would be the most useful (what's not already being done). I have tried and am familiar with some of the software that's out there, but there's so much for so many different sites that it's hard to keep it all straight.

2) Potential help with either programming or graphics in exchange for a cut of the profits. I only have limited time to work on this as my day job is Sr. Developer on a multi-million $ eCommerce site. I have very little experience with graphics so I could use help here as far as laying out an interface that meshes well with PP + skins. As for programming I am doing the software in VB.Net and would tie to a SQL 2K backend for the central data mining database. I've developed some base classes and currently have the program reading hands and some opponent's tendencies (not storing this info anywhere atm).

Thoughts???

- Andy

PokerAce · #2 07-06-2005, 01:20 AM

We developers need to be careful what capabilities we give to our software. Anything that specifically instructs the user what they should do, or what the optimal move would be is basically cheating, and chances are the sites will frown upon it.

[ QUOTE ]
I've even considered modifying the program to offer real-time advice on hands based on the stake, number of players, and player tendencies

[/ QUOTE ]

You realize this is one step away from a bot, don't you? All you would need is code to click the buttons.

There are plenty of applications out there that offer odds, what hands opponents could have, etc. Chances are, the sites are going to start cracking down on software that does this. Giving advice on a hand would be like someone whispering in your ear about a hand at the table. Think that would fly?

So my advice is be careful what features you add to your software. Avoid anything that specifically advises the user of the best action. Giving information about the hand, like previous actions in the hand, or stats on a user is one thing. Telling them there's x% change of making this hand versus this hand, and that their opponents could have this hand, could be crossing into dangerous territory.

Two_Slick · #3 07-06-2005, 01:36 AM

Your point is well-taken. The idea in my head had only been thought through from a "doability" stand point rather than an ethical and/or legal standpoint. You're right, this probably does cross the line. Perhaps an offline version that attempts to teach and analyze play would be a safer route. From a programming standpoint this is a fascinating project because of the AI involved, but I think it would take too long and be too difficult in a reasonable amount of time (especially with little interest in an offline version, but who knows?).

wdbaker · #4 07-06-2005, 02:52 PM

I may or may not understand what you are proposing....

I would give my left nut if someone would build an add on to PT that would scan through the days events and pick out the mistakes and badbeats etc... for review... and then actually give some advice based on the best you can do with software.

I personally have to go through each session, find the large net loss hands and examine them. This helps, but I'm sure that I'm missing a range of hands that have mistakes that I won anyway, that I really should review and understand said mistakes.

In Poker, just because you won doesn't mean that you didn't make mistakes...

One Street at a Time
wdbaker Denver, Co

Two_Slick · #5 07-06-2005, 03:14 PM

Yes, I actually have been thinking about it more since posting this last night and I think that software that scans hand histories offline and flags potential mistakes, and makes suggestions based on "text-book" plays may be more valuable than I originally thought. The only problem I think is that it has to be limited to one type of play (at least to start) like short-handed small stakes, full ring small stakes, etc. Writing software that did a worthwhile job of this would also be a large task. I'm definitely going to explore this some more though. Particularly if I'm successful in finding development help. Thanks for the feedback!