PLO hand ranking system

Hey guys,
Been working on some interesting poker projects lately. To get a better understanding of hand values in PLO and work on my programming, I've tried to write a program that evaluates the 'strength' of a PLO hand preflop by giving it a total score based on 4 different components that are calculated independently (although the value of those components are not independent). This is specifically to help me evaluate hands to play from different positions preflop, ie. I could say that hands that have a score of 1000 or more I open UTG, 920 I open UTG + 1 etc. I want to post hands and hand rank orders here periodically and see if you guys agree with the computer to tweak parameters. Programming was done in R, which is free to download for anybody. Once it gets closer to being done I'll probably be happy to give the code to people here if they think they'd find it useful; not sure if it really will be or not yet.

I've also been working on some bankroll management and backing models. I think the standard 'play until doubled and take x% of profits' is outdated and leaves a lot to be desired, so I've been working with creating a backing model that is based more on the continuous time finance stuff I've been studying (or in the case of poker, a discreet model that updates player/backer equity after the smallest unit of time, an individual hand). Might make a post about that a bit in the future, although few people would find that interesting.

A couple assumptions:
This 'strength rating' is only an approximation for the amount of value one would expect when playing the hand. Hand values can change depending on how PF raising action went (ie. very weak aces are better than the best wraps if you can get allin preflop), but this does not take that into account. The strength doesn't necessarily say what your action should be at a specific time, just whether you should fold or not.

There are certain hands that gain more value from position that others; right now that isn't taken into account, as I don't know enough about what parameters would need to change.

Here is first hand set, as evaluated by Mr. Computer. What do you guys think about the order?

> > > > > > > > >

I would say seems too high, and seems too low. Honestly the program probably does about as well evaluating hands now as I do (and the parameters are all set by me, so I guess that's not a surprise). There are some things I see that look like obvious mistakes, but then in other ways it's probably better.

Here's another set

> > > > > > > > >

looks good cept id reverse the last 2 at 1st glance

well maybe not didnt see it was double suited; hu that J764ds id rather have in ring the suited ace of AJ85 is useful but id fold it generally cept like open buttons and defend bb

Thanks for feedback Trax.

In this stage it is supposed to be a tool to determine whether a hand should be played in a 6-max game. So the value of having a flush draw to an ace is seen as about twice as value able as a flush draw to the 9. This would probably be one of those cases where in earlier position the value of the nut draw is worth more than low double suited draws, but then shifts as we get closer to the button if nobody has opened. In a more advanced version I could have different value matrices for each position, but at this point will stay position independent.

I don't think it is valuing rundown type hands enough (which is one reason why we don't see many rundowns, as I had it output the and rank the top 10 hands out of a draw of 50). I am making an additional modifier for rundowns, and will post another set.

Modified some parameters and added another component for playability advantage in rundowns.

> ??

It sees these as virtually even but the hand with the ace gets a very slightly higher score.

whether it's heads up or multiway prob matters a bit in some of these?

like the A876ds vs 8765ds, the nut suit one prob better multiway and 8765ds better HU

and AAAA > 226Trainbow in HU probs lol