The Lucky Donut

Well I’d been hoping to drop on a non-stupid fixed limit tournament all week, and it just happened.  Right there staring at me in the Poker Stars lobby the minute I signed in… way beyond my bankroll but how could I possibly say no after all this time?

$215 Weekly Fixed Limit Hold’em – Late Reg.

Baptism of fire coming right up then.  It’s 10/20 blinds in level one with 2500 starting chips, and I got a walk on my first big blind.  Definitely playing a little tighter than I’ve seen before.

EDIT: Nitted my way to 14th place.  Wasn’t as useful as it should have been for $215 but at least I did pick up some things to consider.  Probably post some hands tomorrow.

If you have no interest in things so geeky as DIY database queries and obscure poker statistics, look away now.  Next time I’ll probably post a picture of one of my socks or something to bring a little balance back into the world.

So then, how many times have I flopped a set, and how much did I win?  From just over 50,000 hands in my database now, we’d expect to see about 3000 pocket pairs dealt (1 in every 17 hands).  The odds of flopping a set are are 7.5-1, so somewhere in the region of 350 flopped sets would be about right.  I had no idea what to expect as a win rate for these hands alone.

If you could care less about SQL, just skip past the code section, otherwise here’s what I did.  I failed to find any more elegant way of checking for a flopped set than simply comparing the card rank of one of the hole cards with each card on the flop in turn.  This query will actually count flopped quads as well as flopped sets, if I have any.

select count(*) as freq,
       sum(gp.total_won) – sum(gp.total_bet) as won
from game_players gp, game g
where gp.player_id = 11
and gp.game_id = g.game_id
and gp.pair_hand = 1
and substring(gp.hole_card_1 from 1 for 1) in (
         substring(g.flop_1 from 1 for 1),
         substring(g.flop_2 from 1 for 1),
         substring(g.flop_3 from 1 for 1))

Let the results speak.  I’ve seen 317 flopped sets (close enough to what I was expecting) and overall made $2884.50 profit from those hands.  That’s just about $9.10 – 9 big bets at 50NL – each time.  Look at it another way: just over every five flopped sets, I’m winning a full $50 stack from another player.  That sounds OK.

Look at it yet another way: compare to pocket aces.  Poker Tracker tells me that I’ve been dealt AA 253 times and that I won an average of 4.86 big bets each time.  Three hundred of anything is no huge sample size, but it’s a start in the right direction.  We could say that flopped sets are, roughly, twice as profitable as pocket aces.  That’s definitely nothing to complain about.

The difference, of course, is that pocket aces always start off as the best hand – and by a long way too.  What’s been killing me lately is the abundance of set-over-set action, with me on the wrong end of it.  I’m sure I’m remembering things worse than they actually were, so what I want to find out is both how and when I’ve been getting beat.

Adding the condition

and gp.total_won < gp.total_bet

to the end of the query will find only those hands where I flopped a set and did not win.  This includes split-pot situations, as the rake taken from the pot will make the amount returned less than the amount I put in the pot.  I used this condition to find only losses (no splits), although it assumes there are two players at showdown.  Not strictly true, but close enough, and do it any more accurately would create a monster query.

and gp.total_won < gp.total_bet – (g.rake / 2)

The results: 32 losses, 8 splits.  87% of my flopped sets won at showdown (compare: I won with 89% of my pocket aces).  I can live with that – there’s still two cards to come, after all.  The combined loss on those 32 hands was $984.15, or $30.75 per hand.  A big chunk for sure, but almost certainly I was ahead when my money went in.  If only there was an easy way to check that…

The best I’ve come up with is to find all my set-over-set confrontations.  To do this I have to find the times both I and my opponent have a pair that matches one of the flop cards.  The query looks a little something like this.

select me.hole_cards, opp.hole_cards,
       me.total_won – me.total_bet as won
from game_players me, game_players opp, game g
where me.player_id = 11
and me.game_id = g.game_id
and me.pair_hand = 1
and substring(me.hole_card_1 from 1 for 1) in (
         substring(g.flop_1 from 1 for 1),
         substring(g.flop_2 from 1 for 1),
         substring(g.flop_3 from 1 for 1))
and opp.player_id != me.player_id
and opp.game_id = g.game_id
and opp.pair_hand = 1
and substring(opp.hole_card_1 from 1 for 1) in (
         substring(g.flop_1 from 1 for 1),
         substring(g.flop_2 from 1 for 1),
         substring(g.flop_3 from 1 for 1))

It’s ugly for sure, and still a little flawed too.  I can’t even start to get my head around how this might cope with three players all flopping sets at the same time.  This query just dumps out a list of every time I was in a pot with a flopped set against another flopped set, rather than counting wins and losses – or times I flopped the bigger set or smaller set.  It’s only a small sample size – 15 confrontations – but it proves a point.

Won 3, lost 12.  Overall, I’m down $386 from those hands.  A loss of $32.17 each time I got shafted is understandable, but I’m down an average of $25.70 – nearly 26 BB/hand – across all set over set scenarios, simply because I never seem to be on the right end of them.  I would expect that these are usually breakeven situations over the long term: the times you’re ahead you expect to get paid off, and the times you’re behind you’re almost certainly going to get stacked.

Boy I am owed.  And if I talked to you about how poker’s been going for me lately and spouted some far fetched story about how bad I’m running… just look.  Told you so!

The royal flush was an anomaly.

When you run bad you start doubting things.  You start to ask yourself things like whether flopping a set really is that good at all.  Because yesterday, I flopped four sets, and every one of them lost, and every one of them cost me my entire stack.

OK, as you insist, I’ll indulge you with a bad beat story. 

I have pocket fives on a king-five-deuce flop, all different suits.  It all looks so gorgeous.  The pre-flop raiser fires on the flop, I call and the other player – who it turns out couldn’t get away from pocket queens – also calls.  Magic.  The turn brings a second king and he leads out again, but for a much smaller bet this time.  Does he hate the king, or does he have AK and feel like he just got invincible?  Well, I can’t put him on K5 or K2 and I don’t expect he’d have played KK so strong on the flop.  Right now is where I’m going to make him pay, so I move all in.  Pocket queens feels like he’s pot committed and comes along for the ride…

"I have AA", he types in the chat box.  I don’t know why, because almost half a second later he’s called anyway.  Considering I was hoping that he would have ace-king, I’m especially pleased.  Just four cards left in the deck – the two remaining aces and two kings – can improve his hand to beat me now, whereas AK would have had 3 aces, 1 king and 3 deuces for the win.

But when you run bad, you don’t just get beat, you get beat in the cruelest possible ways.  A third king on the river gives us both kings full, his aces both playing but my hand being reduced to what may as well have been a five and a joker.

On days like this you start to wonder whether it’s just a matter of karma.  If you somehow did wrong by someone, but it surely couldn’t be that when I hadn’t left the house all day, and the only phone call I answered was from the bank, telling me that a cheque I dropped through the letterbox at the weekend wasn’t signed.  I now know I must try harder to check these things to make the nice lady’s job so much easier.  She ended the call by asking, with a sigh, whether I wanted her to send it back to me.  Clearly, some of their customers are indifferent about money.

And so, finally, to the business of reassuring myself that flopped sets are actually good hands.  Of course I know this really, but it’s a great excuse to conjure up some stats that Poker Tracker can’t provide by itself.  So I’ve started to construct a few queries that can be run against a PostgreSQL Poker Tracker database to analyse set performance.

You lucky lucky people have a stat-packed post, full of SQL code and probably just proving something we alreayd know, to look forward to – hopefully tomorrow.

Royal flush, baby!

It’s been a while.  Definitely over a year as I’ve not had a picture to post since I started the blog.  I so definitely would, as you can now see…!

Possibly the only thing more awesome than the royal itself is to have hit it backdoor style.

I hope you enjoy the picture as much as I did.  Click on it if you would like to join me in basking in its full size glory. 🙂

I’ve decided to fudge my poker graphs a little, so here’s where I fess up to what I’ve done.

You can click the thumbnails for larger versions if you want, but it’s pretty clear what happened from these tiny versions anyway.

My doomswitch was flipped a few days ago.  Everything I touched turned to shit, and the graph I posted on Friday was only the start of it.  I’d already sunk four buy-ins on Poker Stars and I thought a change of scenery might do me good.  Especially as I’d already achieved Gold Star status on Stars for next month but was miles away from the next level, earning a few more FPPs wasn’t really going to matter.

I found a $25 freebie in my Party Poker account and a $100 reload bonus on Empire.  Often I’d just blitz through these on $1/$2 limit, but as I have a no-limit goal right now, what harm could it do to carry on playing for my NL50 target on here?  Quite a bit, as it happened.

Five more buy-ins dribbled away before I’d unlocked the bonuses.  $250 down for $125 of bonus – not good at all.  Though I got lucky on Party to start with, 250 hands wasn’t enough to realise that this game plays rather different to what I’d gotten used to.  By the end of 1000 hands on Empire, I’d just about figured that out and almost stopped spewing.  Almost.

So I’ve decided to write off the hands on Party and Empire.  Quite naughty I know, but it makes my graph look so untidy otherwise, and a little deception is worthwhile if it helps to keep me motivated.  My original plan was, after all, to see whether I could beat the NL50 game on Stars in particular so I can tell myself somewhat convincingly that it’s not too bad to discard some bad results.  OK it’s still cheating, but I’m not counting rakeback or bonuses (so far worth at least $600) in my profit figures, so it all cancels out in the end.

The comeback begins now, with only a four buy-in mountain to climb to get back on track, not nine.  Suddenly that seems much more achievable.

Time for a quick poker lesson in how to represent a hand, courtesy of Rickib78.

a/s/l?  We assume from the username late twenties, and probably male but using a cool or deliberately feminised form of Richard.  PokerStars tells me he’s from Clayton, Victoria.  G’day, Dick.

Representing a hand is simple, apparently.  Just bet and raise wildly and then your opponent has to assume that you currently hold the best hand available with the community cards that are showing.

Here’s the example :-

Poker Stars – No Limit Hold’em Cash Game – $0.25/$0.50 Blinds – 9 Players – (LegoPoker HH Converter)

SB: $49.45
Hero (BB): $71.40
UTG: $39.95
UTG+1: $28.65
MP1: $16.90
MP2: $21.05
MP3: $34.10
CO: $48.00
BTN: $51.25

Preflop: Hero is dealt K T (9 Players)
UTG calls $0.50, 5 folds, BTN calls $0.50, SB calls $0.25, Hero checks

Flop: ($2) 9 K 8 (4 Players)
SB checks, Hero bets $1.50, UTG calls $1.50, 2 folds

Turn: ($5) 7 (2 Players)
Hero bets $2.50, UTG raises to $10.00, Hero calls $7.50

River: ($25) 2 (2 Players)
Hero checks, UTG bets all-in for $27.95, Hero folds

So I folded the best hand.  He proudly showed QJo, thinking he’d pushed me off a big hand and not, in fact, the weak pair and busted draw that I had.

luckydonut said, "wow u misread ur hand"
Rickib78 said, "looool"
Rickib78 said, "did u have a straight?"
luckydonut said, "you think i folded a straight?"
luckydonut said, "lol"
Rickib78 said, "loool"
luckydonut said, "looooooool"
luckydonut said, "whats this, lol envy?"
Rickib78 said, "well i represented a flush !!!!"
luckydonut said, "you did?"
Rickib78 said, "3 spades on the table"
luckydonut said, "oh so you represented what on the turn, before you decided to represent the flush, exactly?"
Rickib78 said, "on the turn it was a straigh"
Rickib78 said, "looooooooooooool"

Looooooooooooooooooooooooooooool.  I win.

But really, I don’t know what’s so funny.  Clearly I have a lot to learn to take on a player who is good enough to represent a different hand on each and every betting round.

I think I should probably just fold every hand.