ARISE! (If you play in my/our weekly/monthly homegame, feel free to go find something sparkly to distract yourself!
SpaceLord, in the Original Post, wrote:This is a thread regarding strategy and poker styles.
I'm going to begin a new conversation. I'm not sure there's a point to what I'm doing, though I'd love to hear what the random firing of other folks' neurons would have to discuss on this topic (i.e., don't censor yourself).
I've spent a bit of time playing live poker (I live in the US) this spring and summer, and have naturally gravitated towards multi-table tournaments. I gave the $2-$100 cash game a try in the spring and while I think I ended up getting fairly unlucky (on a short bankroll), I also never got the feeling that I had a massive advantage in that game. So I've moved on. (Were I able to play on large online sites, I think I'd be able to leverage my math analysis background into a consistent, if-not-small, winner. But that's a different topic for a different hypothetical time.)
As the summer progressed, I played in a local card room that regularly runs $60-$100 buy-in tournaments. Blinds progress at 20 minute increments, at a reasonable progression... to a point. From my memory, the structure is (T8000 in starting chips):
(break, T25 chips removed)
(break, T100 and T500s removed, beginning of silly season)
3000/6000 (the huge jump!)
...and pretty irrelevant after this point.
For 90% of my summer, that was the structure. It's been (meaningfully) modified since then, but for the sake of this conversation I'm going to ignore that for now.
Games have run from 1-table through 5-tables, with a pretty heavy emphasis on 3-4 tables. I've probably played in 15 of these tournaments of late. I've made it to the final table no fewer than 11 of those 15 times. Typically the payout structure is that for each table you start with, one player will get paid, with no fewer than three folks getting paid. So 1-table, 2-tables, or 3-tables begin = 3 payouts; 4-tables = 4 payouts, and 5-tables = 5 payouts. Additionally, it's not uncommon (75% of the time) for the folks at the tables to make a side agreement that whomever would bust out just outside the money payments (so the person who came in 4th in a 3-table tournament), would get money out of the prize pool covering their buy-in.
Of those eleven times I've made the final table, I've been that "bubble boy" twice, and I've finished one spot better (4th in a 4-table tournament) once. The other 8 times I've finished one-off the bubble six times, and have finished worse the other two times. The TL;DR version is that I'm going very deep in tournaments but have not even been close to profitable in terms of results.
I've had my share of "bad beats" (me being the favorite when all the chips go into the pot, but not winning the hand), but I'm not going to recount those. There have been a few doozies. I've made some mathematically justifiable plays which haven't worked out (i.e., I had a hand that the math says was likely to be the favorite, but then have happened to run into an even bigger hand) on a couple of occasions. I have come into the final table short stacked and been forced to play "faster" (taking aggressive actions with cards worse than desired based on my tenuous position vs. the rising structure) than I'd like, though that was more an issue earlier in the spring/summer than of late.
I'm pretty convinced that I've been unlucky. Getting unlucky is certainly within the realm of possibility for this sample size. Of course, just about any analytically-minded poker player in a similar situation is going to say the same thing. I have not
felt outclassed at the tournament tables (far far far from it), but have arrived at the conclusion that at the point where the game turns into that of "a bingo card," (in the local parlance) I've not had my numbers come up in the hopper. During the more skillful segments of the tournaments, my performance speaks for itself (I generally finish around 15%th when the top 10% gets paid).
But despite that, there's no reason for me to not brush up on my preflop all-in skills (which is what the game turns into when you are deep and the structure dominates your decisions). This is the time of a poker tournament when quality mathematical decisions dominates the thinking process. Don't take that as me thinking the people-skills and reading abilities are not important, it's just that they are reduced in importance compared to earlier in the tournament when you will be playing hands for more bets after the flop has come out. When you're at the shove-or-fold stage, you've basically got one decision to make each hand rather than upwards of ten per hand when the tournament is first starting out.
I've got a tool which helps perform these sorts of analyses, called "PkrCruncher." Given inputs in terms of specific hands or hand ranges, the tool generates your winning percentage. (It does much more than this, but those are outside the realm of the shove-or-fold game.)
Best to show the power of the tool given an example. I'll underline
the inputs I've dumped into the analysis tool.
Number of players: five
Player One: hand range bottom 85% (not: AA-77, AKs-A7s, KQs-K9s, QJs-QTs, JTs, AKo-ATo, KQo-KTo, QJo), action: fold
Player Two: hand range top 20% (AA-66, AKs-A4s, KQs-K8s, QJs-Q9s, JTs-J9s, T9s, AKo-A9o, KQo-KTo, QJo-QTo, JTo), action: shove all-in
Player Three: hand range bottom 90% (not: AA-77, AKs-ATs, KQs, AKo-AJo, KQo), action: fold
Player Four: hero (see below)
Player Five: hand range: any two random cards
The scenario is meant to look at what two cards the person in the SB would need to call an all-in from the player one off the button. A big part of that equation would be ICM (explained upwards in this thread) which would incorporate how many chips you have compared to the other players at the table, and the specifics of the payouts. I'm going to ignore that here in this post, but instead just talk in terms of your chances of winning the hand at showdown. (Mine doesn't answer the ultimate question of "is it profitable to call?" but gets us meaningfully closer to the answer to that question.)
Obviously the hand ranges, above, are totally arbitrary. The Player Two all-in is not someone who's shoving every hand trying to dominate a bunch of weakies at the table, but a more reasoned player that I find to be more typical at the final table. The Player One and Player Three folds are also my reasoned ranges of typical players in this spot. (I could argue that Player Three isn't quite tight enough, but those ranges are entered into the simulator merely for card-removal effects, which are admittedly pretty minor.) Dealing with Player Five is a bit trickier, as you'd almost rather run two separate simulation scenarios: one where you enter the range of hands with which she'd overcall, and one with the opposite of that range. But for now I'm going to simplify things and assume she plays perfectly (or perfectly random) and that she's only in the hand were she to ultimately win.
So my first question is "what's my gut feeling for hands which'd be >50% to win at showdown" in this scenario. Note that 50% is super arbitrary as the amount of money in the blinds, the sizes of the relative stacks, and the details of the prize pool would all be necessary to decide whether calling would be a profitably play, but mostly I'm doing this to try to get a better gut feeling for how good are certain hands under these conditions.
Without running numbers, I'd guess that 99+, AJs+, KQs, AQo+ would be >50% in this scenario. How'd I do?
KQs: 35.9% (whoops -- I was way off)
AQo: 47.7% (again, whoops)
ATs is 41.1% (so I got that breakpoint correct), and it turns out that 88 is 52.3% while 77 is 49.0% (so my guess of 99 was close). AKo is also good at 51.5%.
Turns out that AA is 90.0%, 72o is 16.4%, and that 32o is 16.0%.
Sort of an interesting experiment in my mind.
One thing that is important when it comes to ranges is that for a tight range (typically tighter than Player Two's top 20%), the presence of all-pairs vs. more high-cards when making that definition. A Top 10% of AA-77, AKs-A9s, KQs-KTs, QJs-QTs, AKo-AJo, KQo performs differently than a top 10% of AA-22, AKs-ATs, KQs, AKo-AQo, KQo, even though they basically have the same number of random cards. This isn't my discovery but something to keep in mind when working with tight ranges and these sorts of analyses. At the top-20% the two ranges (high-card heavy vs. pair-heavy) isn't nearly as important.
Anyway, I expect that I'm going to be running a sequence of numbers like these over the days ahead and was interested in seeing if a discussion of such would give me any additional insights. I'm not committing to sharing anything here in the forums, but would be more likely to do so if there's a back-and-forth conversation taking place.