Next SnG tool - the Luckometer

[KJ o]

Here's that final(?) SnG tool I'd like to see as a complement to PokerTracker, SnG Tracker and eastbay's pushometer: The Luckometer.

The Luckometer would meassure your luck in $ in hands starting from the point where everyone remaining in the hand except one is all-in.

Example:
Four players left in a $10+1, everyone has 2000. UTG folds, I have KTs on the button with blinds 200/400. I push and get called by A8o from the SB.

I have (roughly) 46% chance of winning, leaving me with a $EV of 40$, and a 54% chance of going broke, leaving me with $0, if MY ICM calculation is correct. Thus, my $EV is $18.40.

If I win, luck netted me $21.60, and if I lose, bad luck netted me -$18.40.

If I make this calculation for all hands in a tourney and sum the resulting "luck", I think this will give a reasonable aproximation of my actual luck, with two important caveats.

1. This measures only luck after I am actually called. One could argue that if my opponents both had crapy hands and folded, I'd be even more lucky. That is of course a very important factor, but I think that meassuring one luck factor, while imperfect, is significantly better than messuring none.

2. It is quite possible for me to be more unlucky than the money at stake overall, which is of course simply wrong. Without doing any kind if research, I would guess that that isn't a very important factor over a large set of tournaments.

Has anyone done this analysis over a fairly large set of tourneys? Is it meaningfull? My guess is that this figure together with ROI figures would yield a significantly more reliable result after any x tourneys than just the ROI figure.

Any takers to implement this in an automated manner?

[dfscott]

I am so disappointed. I thought you had created a tool to measure how lucky I am. I was really excited about this. I could check it before I sat down -- if I was very lucky, I would play the $200s, if I was just kinda lucky, I'd play my regular buy-in. And if I was unlucky, I would just download porn.

[Phil Van Sexton]

I have written this already. I ran it for your screenname.

The answer is 42.

[the shadow]

Already been done. See Poker Academy Pro (http://www.poki-poker.com/pro_upgrade.php).

I like the bot simulations (I'm still -EV to a couple after 10k+ hands), but the luckometer is of little use, at least to me.

The Shadow

[spentrent]

[ QUOTE ]
I have written this already. I ran it for your screenname.

The answer is 42.

[/ QUOTE ]

Did you see the movie? I saw it last night and left feeling a mixture of satisfaction and dismay. The "So Long and Thanks for All the Fish" song was perfect.

Did you notice the BBC TV version of Marvin lurking in line on Vogsphere?

[Scuba Chuck]

What is the pronunciation of the luckometer?

Is it luck' ah me ter

Or luck O' Meee ter


Seriously.
Scuba

[Phil Van Sexton]

[ QUOTE ]
[ QUOTE ]
I have written this already. I ran it for your screenname.

The answer is 42.

[/ QUOTE ]

Did you see the movie? I saw it last night and left feeling a mixture of satisfaction and dismay. The "So Long and Thanks for All the Fish" song was perfect.

Did you notice the BBC TV version of Marvin lurking in line on Vogsphere?

[/ QUOTE ]

Haven't seen it yet, so I may have to wait for dvd. I don't get to the movies often and when I do its usually with the wife. Sci-fi is a tough sell with her (clearly she will be overruled for Sith, of course).

[sofere]

potayto potahto

[Ryendal]

Did you find something useful in POKER ACADEMY PRO ?

[KJ o]

From what I understand, that doesn't do what I want at all.

I want to run a hand history or thousand through a piece of software that will tell me how lucky I was in my all-in confrontations. Is the software you suggested doing anything like that or do you have commission plugging their stuff? /images/graemlins/tongue.gif

[eastbay]

Already did something like this awhile back. I called it "EV vs AV" (expected value vs. actual value).

I used it briefly but found soon that it was kind of a super techno pacifier that I would suck on when taking a lot of bad beats. "Oh yeah, look, I am running bad. But I already knew that."

I think it would be very difficult to try to tie these luck metrics to your ROI to try to figure out where your "true" ROI actually was as a function of recent luck. That would be pretty cool if it could be done, but I think it may be more trouble than its worth.

There's also lots of different ways you can get unlucky. Getting good preflop hands. Getting good preflop hands when they are useful (for example, picking up AA when you have 5 chips left isn't very useful). Getting good preflop hands that don't run into better preflop hands (running KK into AA). Getting good preflop hands when others get second best preflop hands (KK vs. QQ). Getting good flops. Getting good flops that aren't second best. Getting river luck when you're all-in. The list goes on and on.

eastbay

[sofere]

The most frustrating luck factor I'm running into is Mr. Hangeronner guy staying alive with his all-ins on the bubble when I'm second shorty. Why won't he die???

Of course that same situation flips to good luck when I'm big stack.

[the shadow]

1. I get no commission. I recommend what I like.

2. While you can't load your results into Poker Academy's utility, I found the utility to be essentially worthless. Maybe that's relevant to whether you might find a similar utility to be worth it.

3. The statistic of how lucky you were in the ST in your all-in confrontations is irrelevant. The meaningful question is how lucky you should have been -- that is, were the odds favorable when the money went into the pot? That you can do with Poker Stove (http://www.pokerstove.com/) or Two Dimes (http://www.twodimes.net/poker/).

The Shadow (who, in the interests of full disclosure, reveals that he gets a 10% commission on the sale of all Poker Stove and Two Dimes programs /images/graemlins/wink.gif)

[KJ o]

[ QUOTE ]
Already did something like this awhile back. I called it "EV vs AV" (expected value vs. actual value).

[/ QUOTE ]
This doesn't surprise me. You da man! /images/graemlins/smile.gif

Any chance of sharing? Or including it in the pushometer?

[ QUOTE ]
I used it briefly but found soon that it was kind of a super techno pacifier that I would suck on when taking a lot of bad beats. "Oh yeah, look, I am running bad. But I already knew that."

[/ QUOTE ]

Well, is your gut feeling always aligned with the truth? Since you are a much better player, that may be the case, but I have learnt *not* to trust my gut feeling.

Your pushometer has taught me dozens of cases where I used to do the wrong thing, but now do the Right Thing(TM). If I hadn't had a tool for that, I wouldn't have been able to learn, or rather I would attribute everything to luck.

[ QUOTE ]
I think it would be very difficult to try to tie these luck metrics to your ROI to try to figure out where your "true" ROI actually was as a function of recent luck. That would be pretty cool if it could be done, but I think it may be more trouble than its worth.

[/ QUOTE ]

Agreed that it is difficult. We don't know if it's worth it until we tried, however, do we?

[ QUOTE ]
There's also lots of different ways you can get unlucky. Getting good preflop hands. Getting good preflop hands when they are useful (for example, picking up AA when you have 5 chips left isn't very useful). Getting good preflop hands that don't run into better preflop hands (running KK into AA). Getting good preflop hands when others get second best preflop hands (KK vs. QQ). Getting good flops. Getting good flops that aren't second best. Getting river luck when you're all-in. The list goes on and on.

[/ QUOTE ]

True. But removing one unknown at a time seems easonable and wothwhile anyway.

[eastbay]

[ QUOTE ]

3. The statistic of how lucky you were in the ST in your all-in confrontations is irrelevant. The meaningful question is how lucky you should have been -- that is, were the odds favorable when the money went into the pot?

[/ QUOTE ]

It's not always irrelevant. It depends on what kinds of questions you're asking.

It would be very useful to determine if a current downswing is the result of bad luck or bad play. If you could quantify luck and compare it against a baseline, this would go a long way towards answering that question.

eastbay

[KJ o]

[ QUOTE ]
3. The statistic of how lucky you were in the ST in your all-in confrontations is irrelevant. The meaningful question is how lucky you should have been -- that is, were the odds favorable when the money went into the pot? That you can do with Poker Stove (http://www.pokerstove.com/) or Two Dimes (http://www.twodimes.net/poker/).

[/ QUOTE ]

I don't agree that it's useless, at least not before I actually see the figures.

I used to have significantly more wins than seconds, and I always felt very lucky. Recently, I have more seconds than first and I suspect bad luck.

If I could get confirmed that this was bad luck, no worries. But I fear that it's really a leak where I'm playing worse. The Luckometer would be one way of finding out which is true.

[Apathy]

I really don't think your Ev vs AV even comes close to measuring what you are saying. Luck can't just be determined by how often you win an a showdown compared to how often you should win. What if you were constantly (more then normal) getting dominating hands like AK vs AQ in the end game in situations where if the hands had been reversed you still would have gotten the chips in, but were losing more often with these dominating hands then you should be, the program might conclude you are unlucky, when that could completely not be the case.

Your program only measures how often they win, and makes the assumption that everyone is dealt hands equally. That's a strange assumption when trying to quantify luck.

[Apathy]

[ QUOTE ]
If I could get confirmed that this was bad luck, no worries. But I fear that it's really a leak where I'm playing worse. The Luckometer would be one way of finding out which is true

[/ QUOTE ]

A luckometer would be nice for this reason, plus it would probably be a weird comfort to mediocre players which would allow them to tilt less during a bad run, but a luckometer like the one talked about in this thread wouldnt be effective or accurate, even HU.

[the shadow]

Maybe I'm wrong, but it seems to me that a downswing can be due to any combination of "bad luck" and "bad play." Because luck is perfectly independent of play, having "bad luck" does not tend to make the existence of "bad play" more or less likely. So isn't having "bad luck" irrelevant to determining whether someone's play is good or bad?

[Apathy]

[ QUOTE ]
Maybe I'm wrong, but it seems to me that a downswing can be due to any combination of "bad luck" and "bad play." Because luck is perfectly independent of play, having "bad luck" does not tend to make the existence of "bad play" more or less likely. So isn't having "bad luck" irrelevant to determining whether someone's play is good or bad?

[/ QUOTE ]

it is just as irrelevant as using results to make conclusions about your play, but the problem is that people want to correlate results with how they were playing, and if they have bad results they want to correlate how unlucky they were despite their good play to acheive those results.

[eastbay]

[ QUOTE ]
I really don't think your Ev vs AV even comes close to measuring what you are saying.


[/ QUOTE ]

You might want to read my other posts in this thread.

eastbay

[eastbay]

[ QUOTE ]
Maybe I'm wrong, but it seems to me that a downswing can be due to any combination of "bad luck" and "bad play."


[/ QUOTE ]

Follow you so far.

[ QUOTE ]

Because luck is perfectly independent of play, having "bad luck" does not tend to make the existence of "bad play" more or less likely. So isn't having "bad luck" irrelevant to determining whether someone's play is good or bad?

[/ QUOTE ]

Only when correlated with results does a luck quantifier indicate something about play.

Losses can only happen through some combination of bad luck and/or bad play. If I am getting no bad luck, the only way I can lose is through bad play.

Losing more than the rake with no bad luck is proof that I am playing badly. Without that quantifier, it can be very difficult to tell if the bad results are the result of bad luck or bad play.

Whether it's possible to really nail down and quantify all the components of luck is another question.

eastbay

[Slim Pickens]

[ QUOTE ]
I used it briefly but found soon that it was kind of a super techno pacifier that I would suck on when taking a lot of bad beats. "Oh yeah, look, I am running bad. But I already knew that."

[/ QUOTE ]

Like a dope, I did this by hand. It served to pacify me during a big downswing. That's about all. stupidsucker proposed something similar ("So you want to know if your good" I think was the post title) but no one really got what he was saying.

Slim

[eastbay]

[ QUOTE ]
A luckometer like the one talked about in this thread wouldnt be effective or accurate, even HU.

[/ QUOTE ]

Given what I've said elsewhere in the thread about the many components of luck, I will say this:

I think you would be very surprised at the strength of the correlation between river luck and bankroll swings. It is very strong.

eastbay

[KJ o]

[ QUOTE ]
I think you would be very surprised at the strength of the correlation between river luck and bankroll swings. It is very strong.

[/ QUOTE ]

Then I *really* think you should add it to the Pushometer! I think a lot of people would like proof that they are unucky (independently of whether they actually are or not).

[Scuba Chuck]

I really think it should be called the luck ah me ter. It sounds much more professional than the other, if it really is going to be used.

[tminus]

probably one of the funniest posts ive seen at this site

[tminus]

dont need it...
just play tight and consult a 1-900 psychic when you get heads up

[burningyen]

Is it bad luck for your KK to run into AA?

[KJ o]

Well, duh! I specifically mentioned that meassuring the luck in all-in situations after the hole cards are known is only one aspect of luck, but suggesting that it is uninteresting just because it's not 100% complete is silly bordering on moronic.

The pushometer will tell you that getting your chips in the middle with KK is a good idea. The luckometer will tell you that running into AA and losing is unlucky, but only slightly so. Together, they will tell you that you played well but was unlucky. Change KK, AA and losing to T4o, K9s and winning and they will tell the opposite story.

[burningyen]

[ QUOTE ]
silly bordering on moronic

[/ QUOTE ]
Wow, never had an idea of mine called that before. I must finally be doing something right.

From your original post:
[ QUOTE ]
I think that meassuring one luck factor, while imperfect, is significantly better than messuring none.

[/ QUOTE ]
I suggest you look up the concept known as the "theory of the second best."

[ QUOTE ]
The pushometer will tell you that getting your chips in the middle with KK is a good idea. The luckometer will tell you that running into AA and losing is unlucky, but only slightly so. Together, they will tell you that you played well but was unlucky.

[/ QUOTE ]
Wait, so now my silly and moronic idea is being incorporated into your Luckometer? Do I get a silly and moronic royalty for that?

[burningyen]

Also, consider the statistical likelihood of your KK running into AA. Why is that only slightly unlucky?

[KJ o]

[ QUOTE ]
I suggest you look up the concept known as the "theory of the second best."

[/ QUOTE ]
I suggest you tell us why that theory would be applicable here, instead. The theory talks about what happens when one condition necessary to achieve Pareto-optimality is missing. Please explain what this has to do with the situation at hand.

(I'm not saying you are not right, it's just not ovious to me what you are aiming for.)

[KJ o]

[ QUOTE ]
Also, consider the statistical likelihood of your KK running into AA. Why is that only slightly unlucky?

[/ QUOTE ]

I was referring to the fact that losing with KK vs. AA is slightly unlucky. Running into AA with KK is very unlucky, expecially short-handed, of course.