early flips

[schwza]

according to ICM, doubling up on the first hand only gives you .184 of the prize pool compared to .10 before doubling, so it would appear that taking an early flip for your stack is horrible. but if you're trying to maximize $/hr instead of ROI (it's obvious that you should be, imo), you need to know how much time doubling adds to your game, on average. does anyone know this information, or at least have some impression?

it could matter a lot - here are some completely made up numbers.

suppose i one-table 11's (one-tabling does not affect the argument, just makes it simpler). my ROI is 20% (so my expected payout is 1.2 * 11 = 13.2) and my average stt takes 40 minutes. my hourly rate is $2.20/40 min = $3.30/hr.

if i double up on the first hand, assume my expected payout is multiplied by 1.84 (could be higher or lower in reality), so it's 24.3, for a profit of $13.3. also assume that it takes 50 minutes on average if i double immediately, for a rate of $16/hr.

thus, passing on a flip has an hourly EV of $3.30. taking a flip has an hourly EV of:

.5 (16) + .5 (-11 + 3.30) = 8 - 3.85 = 4.15 [-11 for busting, + 3.30 for his EV in the next one]

so with these numbers that i made up, it looks like this guy should take a flip on the first hand.

i'm not saying this proves anything at all; i just wanted to point out that some of 2+2's horror at playing in the first level can be traced to incorrectly focusing on ROI instead of $/hr. if someone has some better guesses for actual numbers, i'm all ears.

[gumpzilla]

Interesting. However, something nags me a little bit about the numbers. I think one issue is that you're looking at ROI in one case, but also subtracting 11 dollars when you bust. But ROI already has the buyin subtracted, since it's profit above the buyin. So I think the two things you're comparing shouldn't be compared. Of course, the correction to this would lead to an even more pronounced version of the same effect, since it leads to an increase in the EV of flipping.

Also, I suspect things get more complicated if you're getting opportunities to double up with flips not on the first hand, but, say, 10 minutes in.

[Bigwig]

I didn't get enough sleep to properly analyze your numbers. But my personal belief is that I don't mind getting into flip situations early if I was the one raising all-in. This obviously changes as the tournament advances.

[Moonsugar]

I try to get all my money in with AK as often as possible in the beginning of SnGs. I read somewhere about the math a couple of years ago. $/hr it makes sense.

[Bigwig]

[ QUOTE ]
I try to get all my money in with AK as often as possible in the beginning of SnGs. I read somewhere about the math a couple of years ago. $/hr it makes sense.

[/ QUOTE ]

You are talking about being the pusher, correct?

[schwza]

sorry, the math on that was all screwed up. that was pretty sloppy. i think this is proper.

now i think for this guy taking a flip is bad, which makes more sense. the point i was originally trying to make in the first post is that taking a flip reduces ROI from 20% to 10.5%, which looks like a disaster. but it only reduces $/hr from $3.30 to $2.76, which is not nearly so bad. of course, the issue is usually not taking a literal flip, but something stickier like "should i limp 45s in MP2?" people seem to usually overlook the fact that it's much better to bust on the first hand than bust in 4th 40 minutes later.

pass on flip

buy-in = $11, expected cash-out = $13.20, time = 40 minutes, so hourly rate is $3.30. ROI = 20%

take flip

win flip
buy-in = $11, expected cash = $24.30, time = 50 minutes.

lose flip
buy-in = $11, expected cash = 0, time = 0 minutes.

total after taking flip
assume you take two flips and instantly lose the first and win the second.

buy-in = $22, cash = $24.30, time = 50 minutes, rate = $2.76/ hr. ROI = 10.5%

[gumpzilla]

Here's another issue that occurred to me.

You're using ICM to adjust the equity as a result of doubling up. This is all well and good, BUT we're already assuming that ICM isn't telling us the whole story about the equity of this player, otherwise his ROI would be -9% like Joe Public. Your assumption of scaling up the equity as ICM says is certainly a reasonable choice, but it seems to me like there's a fair amount of wiggle room conceivable.

[schwza]

i agree completely.

[ QUOTE ]
if i double up on the first hand, assume my expected payout is multiplied by 1.84 (could be higher or lower in reality

[/ QUOTE ]

when i got to that point, i said, "um.... multiplying by 1.84 is not totally crazy, right?" but yeah, pretty arbitrary. doesn't obscure the main point that focusing on ROI instead of $/hr leads to wrong conclusions though.

[citanul]

aside from the already pointed out problem of ICM being both inaccurate for situations where there are many players, and pretty inaccurate for situations where a player plays more skillfully than his opponents, i have another issue with this line of thinking. note: i have no problem with pointing out that hourly rate is considerably more important in many situations than ROI.

here's my problem: i think that people sometimes take thinking of hourly rate as more important than ROI too far. particularly if you're still a learning player and still in the lower stakes games, trying to max out your hourly rate instead of learning to play as well as possible is probably not the greatest idea in the world. situations that are going to be assured early flips at worst against bad competition are going to become worse and worse spots as the competition gets harder. not that that last sentence is exactly on the same point as my real point. =P but yeah, basically, taking flips may or may not increase your hourly rate earlier in games, but it certainly doesn't teach you to play great poker.

when i consider the sngs, i consider ROI the main statistic, even though i know there's hourly rate, and all sorts of more advanced statistics to delve deeper into shapes and powers of your results. i feel that ROI gives a more accurate assessment of "how badly you are beating the game" than the other statistics. hourly rate can be manipulated and befuddled by all sorts of things, notably adding more tables or playing rolling games instead of sets of games.

the point that you make, which was sort of made in that disgusting "which would you prefer, 3rd or 10th" thread, that hourly rate is worthwhile to consider especially when you are considering situations where you might bust out early is clearly a very good one. i just beg all people who start to think about this to remember that each time you bust out early you lose a buyin, and at a 25% ROI, that takes 4 games to make up. stuff like that.

alright, before i start to ramble like skipper in his sleep,

citanul

[wiggs73]

[ QUOTE ]
[ QUOTE ]
I try to get all my money in with AK as often as possible in the beginning of SnGs. I read somewhere about the math a couple of years ago. $/hr it makes sense.

[/ QUOTE ]

You are talking about being the pusher, correct?

[/ QUOTE ]

If the answer is yes, then I also take this line. I will very rarely call in all-in in the first level, but with substantial action pre-flop (but not to the point I assume someone has AA or KK), I don't mind pushing with it. I think this is a much better line, BTW, in the 11s and 22s where I play rather than higher buy-ins.