Quote:
Originally Posted by Trev
So it doesn't use a database to keep past session stats on opponents? That renders it MUCH less useful.
I agree on that. Let me explain a few of the difficulties of implementing a persistent database.
1. Aggregated data is only meaningful if it is properly categorized; you cannot combine tournaments statistics with ring statistics (for example), because you will get an overall average that will only mislead you with false numbers. For even more accurate results, you might even have to consider blind levels, since people tend to be more aggressive late in tournaments rather than early on. If you just aggregate data together and get an average, you might have very misleading information.
2. Categorizing information in an automated manner is TRICKY. Not because it's a difficult task per-se, but because poker softwares don't give you much information to work with. Some poker clients are good at giving plenty of information in the hand histories, others don't care at all and you have to make assumptions. For example, only a handful of them give you information about the maximum number of seats available at the table. Others you can infere that information from the name of the game. Others (like SWC), don't give you anything. None of the poker clients I built a parser for give you a definite and clear description of what type of game you are playing (whether a ring, tournament, heads-up, etc.). You have to infer it from the name of the game (if that's even available).
3. Importing existing hand histories and saving statistics data in a local database. This is an easy task compared to the previous two, but it still takes some time to implement correctly.