|
|
#21
10-19-2005, 09:59 PM
|
|||
|
|||
Re: Help with Ace 4 hands in Limit O8
[ QUOTE ]
[ QUOTE ] [ QUOTE ] [ QUOTE ] [ QUOTE ] I need 2 other things to make A4 playable in limit, a deuce and a trey. [/ QUOTE ] Then you are a pussy. You might want to consult with the fella that puts the O8poker.com website together. His advice on A-4 might make you some money. [/ QUOTE ] Copy and pasting what other people write, then putting it on your own website, does not make it "your advice". [/ QUOTE ] Why care whose advice it is? Good advice is good advice. My interest is in becoming a better Omaha player and making more money. Even your advice is good advice, on occasion. But as Professor Buzz shows above, saying that the only good A-4 is A-4-3-2 is ridiculous. [/ QUOTE ] Of course it's ridiculous, It's pretty blatantly obvious I wasn't being serious when I said it. You need to get your humour chip installed in your robotic brain. [/ QUOTE ] Then I owe you an apology. Usually "humorous" things result in laughter. I didn't laugh. But then again, I am a robot. I am a robot. I am a robot. 
|
|
#22
10-20-2005, 04:24 AM
|
|||
|
|||
|
Re: Supplemental simutation data, A4XY hands, 9 handed
[ QUOTE ]
The only caution I’d toss out is that these are no-fold-em sims. [/ QUOTE ] Hi Greg - Good point. But no-fold-em sims do show relative potential value of various starting hands. For example, A45Q-double suited is quite a bit better as a starting hand than AJQK-rainbow, and I think A45Q-double suited is even better than AJQK-double suited, although it doesn't scoop quite as much. It's a moot point because both A45Q-double suited and AJQK-double suited are very playable starting hands. hand......high....low.....scoop....total....adjust ed A45Qd.....421.....739.....803.....1963 .....1527 AJQKn.....342.....0.....758.....1100 .....986 AJQKd.....662.....0.....876.....1538 .....1317 (The new numbers to the right of each hand are possible adjusted values, giving scoops greater weight in the totals - what do you think?). The simulator counts: • each untied win for high, where somebody else gets a share of low as 0.5, • each tie for high where nobody makes low as 0.5, • each tie for high with one opponent, where Hero or somebody else gets a share of low as 0.25, • each tie for high with two opponents, where Hero or somebody else gets a share of low as 1/6, and • each tie for high with three opponents, where Hero or somebody else gets a share of low as 1/8. Thus the "421" in the high column for A45Qd represents at least twice that many high wins. Similarly, the “739” in the low column for A45Qd represents at least twice that many low wins. Because tying for low is more common than tying for high, and because we're not concerned with tying differences (as we are in high hands), we can estimate how many low wins are actually represented by “739.” Probably the number of low wins plus low ties is in the neighborhood of 2460. But it doesn’t really matter. “739” is the accurate equivalent of whole pots for simulator purposes. A flaw, as pointed out clearly by Chaos in an earlier post, is that I shouldn’t just be adding the equivalent of whole pot wins for high to the equivalent of whole pot wins for low to the number of scoops to get what I’m calling “total.” What no-fold-'em sims show is the <font color="white">_</font>potential to win the pot. The <font color="white">_</font> potential for scooping is actually worth substantially more than twice the <font color="white">_</font> potential of winning half the pot. I could get a better approximation of the true potential of a starting hand by going back and multiplying the high column by a factor, maybe ~4/5, and multiplying the low column by another factor, maybe ~3/5, to make the totals better reflect the actual potential of the hands. The way it is, I'm clearly not giving enough emphasis to the potential for scooping. One difficulty is the actual factors that should be used for each hand are a bit different for each type of hand (the A4 in A48K is not quite as likely to make a winning low as the A4 in A49K). At any rate, although you didn’t specifically mention this particular dilemma, I’m aware that adding high + low + scoop to get a grand total for each hand is flawed. A subtle difference between the potential for scooping and the potential for winning half the pot is you're not necessarily risking more than the amount you can lose in the first betting round. What? You didn’t get that? [img]/images/graemlins/smile.gif[/img] Well... honestly I don’t think it was very clear. Let me try to explain. For simplicity, let's assume no raise on the first betting round and one bet with no raise on each succeeding betting round. Let one chip represent one small bet and two chips represent one big bet. If you see the showdown with a hand, your total investment in the hand will be 6 chips. Thus if there are 24 chips in the pot at the showdown, 6 of those will be yours. • When you win the whole pot, you'll get 24 chips (six of them from your own original stack). • When you win half the pot, you'll get 12 chips (six of them from your own original stack). • When you lose, you’ll lose 6 chips. However, in real life you don't usually <font color="white">_</font>lose 6 chips!! Assuming you play the game well, usually you fold immediately after the flop when you're headed for a loser, Or perhaps you fold after seeing the turn when you're headed for a loser. Thus on the first betting round, it's not a matter of risking 6 chips to be awarded 24 when you scoop or risking 6 chips to be awarded 12 chips when you split the pot evenly with one opponent. You’re only risking the one chip!!! If you see the flop for one bet in a game where there commonly are 24 chips in the pot at the showdown, (6 of which will be from your original stack), your implied odds are 18 to 1 for scooping and only 6 to 1 for winning half the pot. So in terms of your pre-flop odds, scooping in this particular instance is worth almost three times as much as getting awarded half the pot. There’s no way to generalize this discrepancy. What I mean is when you’re looking at your own cards before the flop, the potential to scoop is not always worth three times as much as the potential to win half. It’s simply different for each pot, depending on you, your cards, your opponents, and their cards. Yes, it's true that you get your own chips back when you invest 6 chips in the pot and win either the whole pot or half the pot and it's true those chips are the same as any other chips won - but it's not true that you'll necessarily <font color="white">_</font>invest 6 chips if you’re not going to win the pot. There are so many variables.... We always want to simplfy, to come up with a simple rule. Some two card combinations are easy to generalize about. We almost always should want to see the flop when we’re dealt an ace and a deuce. We almost never should want to see the flop when we’re dealt a seven and an eight. If I listed all the possible hands with an ace and a deuce, almost all of them would be colored black (my coding for playable). If I listed all the possible hands with a seven and an eight, almost all of them would be colored red (my coding for trash). But it’s hard to generalize about hands with an ace and a four. Some of them are playable, some are marginal, and some are trash. There are always the other two cards and the suitedness of the hand to consider. And that’s why I listed and colored all those different ace-four hands, and why I listed them in the order I listed them.... so that the different possibilities and the patterns would be clearer. [ QUOTE ] In actual play, your results will be significantly worse, because 1) you will be forced to fold on the flop many of those hands that would have improved on the turn and river. [/ QUOTE ] True, but also true of any starting hand. Okay, less so for premium starting hands since they are more likely to catch the nuts or a nut draw on the flop. But unless you’re a beginner trying to learn the game inexpensively, you can’t wait for premium hands! Most of the ace-four hands, especially if suited and missing middle cards as the associated two cards, are bona fide contenders for the pot, with very favorable pre-flop implied odds. I don't want to seem defensive here. I actually feel very objective about the data and am seeking as correct an interpretation as possible. [ QUOTE ] 2) You will not have 8 other people sticking in all that money thru to the river. In real life many of the crappiest hands will get folded, not played, and with 8 others in you are losing equity to the 1-3 others who have A2/A3, but are gaining a lot of equity from the fish with 389K-type hands. So your good results rely heavily on players 5-8 coming in – players who in real life may not be there. [/ QUOTE ] I didn't set the sims up to record money won. The data presented simply shows how many whole pot equivalents any given hand figures to win. And all the results are not “good.” Some A4XY hands are trash, especially the rainbows with one or two middle cards. Finally, if I reported the results for all hands with a seven and an eight, the results would not be good, in spite of the equity gained from fish with 389K-type hands. [ QUOTE ] Let me know if I’ve misunderstood anything on the methodology here Buzz. [/ QUOTE ] No, I think you get it. Your criticisms are valid. [ QUOTE ] That said, these make a wider subset of A4 hands look very playable (my read indicates the two most important factors as being 1) suited and 2) no middle cards) [/ QUOTE ] Bingo! That's my impression also. Thanks, as always, for your insights and input. Buzz
|
|
#25
10-20-2005, 04:45 AM
|
|||
|
|||
|
Re: Supplemental simutation data, A4XY hands, 9 handed
Buzz,
Interesting how well the A44Hd hands score. They seem to have a favorable scoop component. (I notice that your total number is a simple sum of high low and scoop. Have you ever considered weighting the high and low by dividing them by 2 to give extra emphasis to scooping.) [ QUOTE ] A4XY simulation data, 9 handed non-folding, 10000 runs each For my own usage, I have color coded playable (black), marginal (green), and trash hands (red). snip To put things into better perspective, here are the sims results for a few starting hands you’d probably want to play: hand.....high.....low.....scoop.....total A234n.....127.....1918.....351.....2396 .....(rainbow) <font color="green">A278n.....257.....866.....240.....13 63 .....(rainbow) </font> <font color="green">A288n.....194.....765.....358.....13 17 .....(rainbow) </font> AJQKd.....662.....0.....876.....1538 .....(double suited) <font color="red">AJQKs.....468.....0.....775.....1243 .....(AJ of same suit) </font> <font color="red">AJQKn.....342.....0.....758.....1100 .....(rainbow) </font> TJQKd.....650.....0.....878.....1528 .....(double suited) <font color="green">TJQKs.....550.....0.....788.....1338 .....(KT of same suit) </font> <font color="red">TJQKn.....409.....0.....721.....1131 .....(rainbow) </font> QQKKd.....681.....0.....1182.....1863 .....(double suited) QQKKs.....555.....0.....1119.....1674 .....( KQ single suited) QQKKn.....407.....0.....1125.....1532 .....(rainbow) [/ QUOTE ] Would you even want to play the red ones? Mack
|
|
#26
10-20-2005, 09:17 AM
|
|||
|
|||
|
Re: Supplemental simutation data, A4XY hands, 9 handed
[ QUOTE ]
Interesting how well the A44Hd hands score. They seem to have a favorable scoop component. [/ QUOTE ] Hi Mack - Yes. I noticed that too. Interesting. I don’t generally like playing hands with low pairs because you can take a bath when you make a full house. You’ll generally be stuck in the pot because you’ll have a hard time telling if your opponent who is betting has you beat for high. It’s been a while since I ran most of those hands. I didn’t keep track of how they fared for specific types of high hands. I’ll re-run A44Kd because it might be interesting to see how the hand does with respect to “high hands held” possibilities. Wilson lists all the various categories of high hands and how often the test hand won and lost when the best high it could make fit into a particular category. First the new totals are a bit different, as expected. hand......high.....low....scoop.....total A44Kd.....496.....604.....772.....1872 ....old A44Kd.....453.....606.....755.....1814 ....new When I re-run a particular hand, I generally average the totals for the old and the new, and that’s what I’ll do here for my main data bank. But for this post, I’m going to use the new values. What I really want is how often the best high made by A44Kd wins and loses. Here’s the data: • one pair ....... 1 win ..... 2468 losses • two pair ..... 64 wins ....3118 losses • trips ......... 149 wins, ...1280 losses • straight .... 205 wins, .....299 losses • flush ......... 803 wins, .....563 losses • boat ......... 370 wins, .....581 losses • quads ........ 92 wins, ......... 1 loss • st. flush ....... 6 wins, ......... 0 losses The total is 10,000, as it should be. The total number of wins for high is 1690. If you look several paragraphs above at the whole pot equivalent totals for the hand, it wins high 453, and scoops 755, a total of 1208. There’s an obvious discrepancy of 482. This is due to two factors. (1) When the hand ties for high, the tie is listed as a win in the data for “high hands held,” but the tie counts as a fraction in the data for “totals.” (2) When the hand doesn’t scoop, but wins the high (and may also win part of the low), this also counts as a fraction in the data for “totals.” I don’t generally like playing hands with low pairs because I don’t like losing with low full houses. For example, you make a full house with A44Kd roughly 9.5% of the time, but only win 370 while losing 562 against a full field. Losing with a low boat can be brutal because you’re generally stuck in the pot, not knowing whether the opponent who is betting is just driving with the low or has you beaten for high (or both). Meanwhile, your own low is counterfeited and you’re generally playing for only the high half the pot. Even when your boat ends up a winner for high, you’ll generally miss a bet on the end because you’ll only be playing for half the pot and can’t usually very well tell if you have a winner or not. A44Jd does end up more with a flush than a full house - and it wins more and loses less with a flush than with a full house. In addition making quads or a straight flush is obviously nice and there are some wheel and Broadway possibilities. Thus I’m going to want to see the flop with A44Kd, even though I hate that low pair. Someone might suggest folding when the flop has a four, but that’s probably not the best way to play the hand. If you’re going to play a hand with a pair of fours, then I think you have to play your low set if you flop one. But it can be a tough play! And the same goes for playing non-nut flopped lows. With a deuce or trey on the flop plus another low flop card, you’ll have the second nut low draw. That particular second nut low draw is going to encounter the nut low draw about half the time in a full game. And that makes it tough to play. I’m not saying play A44Kd or don’t play it. I’m just saying some flops can be tough to play. What to do in a real game “depends.” The game is Omaha-8, but you still have to play good poker. If anyone has difficulty playing a flopped set of fours, maybe they need to learn to play better, but maybe folding is their best course of action until they learn to do that. However, if they’re going to fold a flopped set of fours, or if they’re going to fold a flopped second nut ace-four low or second nut ace-four low draw, then A44Kd is obviously worth less as a starting hand for them than the sims indicate. [ QUOTE ] (I notice that your total number is a simple sum of high low and scoop. Have you ever considered weighting the high and low by dividing them by 2 to give extra emphasis to scooping.) [/ QUOTE ] Yes. However.... Wilson has already divided the high and low wins by at least 2 in order to convert them to the equivalent of whole pot wins, and.... I think dividing by 2 again (using a factor of 0.5) is too much. My current thinking is that maybe a factor of 0.75 or 0.8 for high and a factor of 0.6 or 0.65 for low gets closer to the correct relative value for real game play. So I’m not sure how to do it exactly, but clearly my “totals” need tweaking to give scoop possibilities the greater consideration they deserve. At any rate, thanks for your suggestions. Buzz
|
|
#27
10-20-2005, 02:32 PM
|
|||
|
|||
|
Re: Supplemental simutation data, A4XY hands, 9 handed
Buzz,
[ QUOTE ] Yes. However.... Wilson has already divided the high and low wins by at least 2 in order to convert them to the equivalent of whole pot wins, and.... [/ QUOTE ] OK. Didn't realize they had already done this, so it makes no sense to do it again. [img]/images/graemlins/smile.gif[/img] [ QUOTE ] So I’m not sure how to do it exactly, but clearly my “totals” need tweaking to give scoop possibilities the greater consideration they deserve. [/ QUOTE ] Me neither, I sometimes dump sim results into an Excel spreadsheet and graph the results. I might give this a try and see if anything interesting leaps out at me. Thanks for the interesting work, Mack
|
 |
|
|
Linear Mode
