Looks to me like Holmes just got lucky in the 3456 hand. I have Wilson ... haven't touched it in about 6 years but I remember passing time in some cross-country flights challenging old Sherlock.

OP, you should consider changing the lineups in your game. I think you can also change the number of opponents. Holmes is playing a set strategy against a bunch of stiffs seeing the flop 70% of the time. See how he does against villains seeing 33% of their flops. Holmes won't change, but you will. I suspect that Holmes will be break-even at best against tighter opponents. His AI is set up to presume that he's playing the Villains that Wilson was playing in the late 90s when he wrote the program.

In addition to having some stock characters that you can use in addition to the ones in play (and you can face 9 of the same villain if you want), you can create your own. None of them are going to have imagination or flair, but having to face tighter and more aggressive opponents, even in AI, is going to be more realistic and help you get more from the experience. In my experience, even up to 30/60, a great many opponents IRL are going to be using 1st level thinking only, which isn't that different from AI.

Jeff Hwang actually stated this a time or two in his book. I don't really disagree -- you probably can be a marginal winner in a typical loose low-stakes live game just by playing good hands and not being an idiot postflop. When it's 7 to each flop, your will have a few opportunities for +EV fancy plays but not nearly as many as we'd probably like to believe.

Obviously this forum isn't really for people who are satisfied aiming to grind away the next 30 years with a marginal edge.


Tighter opponents may not be more realistic. OP seems to be playing this as an academic exercise, not in preparation for any particular game. If he were training to take on the Greektown 5/10, optimizing play for opponents playing 33% VP$IP would be senseless.

AKQJ10:

I recognize the error in being too results oriented. However, there comes a point where you have to concede that an opponent (especially "Sherlock") might not be just lucky. Sherlock, despite being a software creation, just might be a damn good player. I have now played 62 of these 100-hand challenge matches versus Sherlock. Early on Sherlock was outchipping me by a ratio of greater than 2.2 to 1. Over the last 10-15 matches, I've managed to whittle that ratio down to somewhere in the neighborhood of 1.9 to 1, (so I'm getting better), but Sherlock is still one very tough opponent. (Of course, it's also possible that I'm not a very good Omaha/8 player, but like you said: "We're all learning.")

I'm personally aware that Bob Wilson ran (literally) thousands [tens of thousands] of million-hand simulations testing just about every imaginable scenario that can come up at the table. The point count system that Bob developed evolved from all those simulations. I did database programming work years ago, but I can't begin to imagine how challenging it must be to create a software opponent that plays like Sherlock. (I call Sherlock a "beast" - and that's a compliment!)

The Wilson software is not perfect. Occasionally you'll see the Wilson software (like any complex software application) do something very "un-human like" and make a play that you would never see from a human player. (Interestingly, I have seen these "hiccups" more often in the various Wilson hold 'em simulations rather than the Omaha program.) Even in the Omaha/8 software, the Wilson program makes bets and raises that leave me scratching my head, but I'm (slowly) starting to understand why the software is making some of these plays. It's a major miscalculation to think that the Wilson program plays unimaginatively ... that's certainly not the case with Sherlock!

I know some folks discount rudimentary point count systems (and I did myself for a long time), but there does seem to be a definite statistical correlation between hands with high point counts and hands that win pots. Also, nearly all of the Omaha/8 books repeatedly stress the importance of starting hand selection. Both Super System 2 and Lou Krieger's "Mastering Omaha/8 Poker" stress that good starting hand selection is the "most important" skill (and the key) to playing winning Omaha. Lou goes so far as to say that if the only Omaha skill you learn is good starting hand selection, that will probably be enough for you to be a winning player. The experience I am having with these challenge matches versus Sherlock tends to confirm this.

To the extent that point count systems improve your starting hand selection, they can't be bad - although the more detailed point count systems are not "simple" or easy-to-implement - especially when you're at a live table. At some point you have to be able to "count" without actually counting - if that makes sense.

Former DJ

I don't think you should get too caught up in an exact point system. It's a good guide for a beginner or relative novice, but not a steadfast rule.

I'm not a great FLO8 player and rarely play it these days, but I'll ante my two cents here.

First, it depends on your opponents. Second, it depends on how you will tend to play the hand post-flop in this given situation.

I see some good reasons to raise here, esp. if you are likely to see this hand thru showdown in many cases:

- you can buy the button and assure yourself of last to act
- you can win the blinds, although expecting three players left to act and three limpers to all fold may be a bit much to ask in most cases
- you can force out other Axxx hands which might hit two pair on an Axx flop, promoting your AK to best high
- you can force out smaller pairs that might hit a set to beat your KK or AAK (if A comes on flop), or random hands that might hit trips or str8 to beat your high hand
- you can force out other low draws which might promote your poor low to best low if remaining players have no low or are counterfeited
- you can give yourself better odds to get to showdown with an iffy high hand, although this is a double edged sword (end up staying in with iffy hand that you would have folded without PF raise)

What are the reasons not to raise? Well, it's not the greatest hand ever, you might force out lower flushes that would pay you off when your nut club flush hits, and you might be more likely to stay in with an iffy hand and lose a bigger pot.

I would say the reasons to raise generally outweigh the reasons not to raise. However, it again depends on your opponents and how you will play out the hand. If you're going to fold to any bet on the flop unless it's QJT, there's a K, or 2+ clubs, then it's probably better to just call and play fit or fold. If you intend to play the hand further, often until showdown, even on marginal flops, then I think a raise is good here.