What is your longest downswing/upswing due to variance?

[trader01]

Here's the question: variance obviously plays a big part in texas hold'em, with players getting "hot" or "cold" for periods of time, but I haven't seen much talk about the maximum number of hands that anyone can honestly attribute to variance.

For example, if someone gets "hot" for 100 hands, that's clearly variance and cannot be attributed to that poker player's skill - clearly anyone can get a good run of cards over only 100 hands. Similarly, if a player's profit graph is straight down for 100,000 hands, that is clearly more than just "bad luck" or variance - if someone is going straight down for that many hands, it's because they are not playing the proper strategy or don't know how to play.

So the question is, at what point (in terms of number of hands) can we say that a player's skill truly starts to show and that the slope of the graph is not due to variance? Similarly, how many hands would constitute the maximum reasonable run of "abnormal" good (or bad) luck that a player can expect to see from time to time due to variance?

I'm interested to hear any stories...

[Eric Stoner]

[ QUOTE ]
Here's the question: variance obviously plays a big part in texas hold'em, with players getting "hot" or "cold" for periods of time, but I haven't seen much talk about the maximum number of hands that anyone can honestly attribute to variance.

For example, if someone gets "hot" for 100 hands, that's clearly variance and cannot be attributed to that poker player's skill - clearly anyone can get a good run of cards over only 100 hands. Similarly, if a player's profit graph is straight down for 100,000 hands, that is clearly more than just "bad luck" or variance - if someone is going straight down for that many hands, it's because they are not playing the proper strategy or don't know how to play.

So the question is, at what point (in terms of number of hands) can we say that a player's skill truly starts to show and that the slope of the graph is not due to variance? Similarly, how many hands would constitute the maximum reasonable run of "abnormal" good (or bad) luck that a player can expect to see from time to time due to variance?

I'm interested to hear any stories...

[/ QUOTE ]

Because we don't know.

Not to be flip about it, but the reason why variance is not written about (with regards to your specific question) is that we can't go back in time to analyze over a x number of hand sample what is "variance" or "bad luck" (or good luck) or what is bad play or what player is going to lose.

Similarly, we can't surmise that well since I have had a 10,000 swing downward of 300BB, I figure I am "due" for an upswing just as we can't surmise that I have not had pocket Aces in 219 hands so therefore that next hand will be aces.

That introspection is left for each individual player.

With that said, I've had two instances of 10,000 hands where I have swung downward of 150BB. However, over 50,000 hands I have been a 1BB/100 winner - not great but not bad. There may have been instances within both of those downswings where I've played poorly and where luck did not go my way.

In either case, it still was not fun.

There have been players who have endured a 500 + downswing - solid, professional players with lots of experience as well as others who have had multiple drops of 300BB. From reading their posts in the other forums, they were/are solid players, but only they can tell you if the downswing was purely variance purely bad play, or a combination of both.

And even then, it's still not that cut and dry.

It is understood though that if a player has a 3BB+/100 hand sample over a short period of time (up or down) that luck possibly played a factor in those results.

It's far better to focus on the hand-by-hand decisions at the table(s) without worrying about the fluctuations of variance. It's a very deadly and delusional trap.

[trader01]

I guess what I am getting at is trying to figure out whether my recent play against certain bad players is really just variance or not... obviously I have no doubt that I can consistently make money from these people because right now I am just killing them, but the question is, how much of my current return is variance and how much can I expect to have it continue?

I just started on a new site where the players are complete garbage, and my return over the first 18,000 hands is on average 9.55 BB/100... and that includes a long period where quite frankly I was playing too loose. If I just take the last 7000 hands where I have limited my play to peak times, switched tables at the right time, and honestly just played better, my return is around 17 BB/100, with one period of 2000 hands where it was 33 BB/100.

Now, I am pretty sure that the 2000 hand spurt of 33 BB/100 was variance, and I am also pretty sure that over 17,000 hands I can count on around 10 BB/100 (since that seems like too many hands for simple statistical deviation), but I'm wondering if 7000 hands is enough to determine the new (improved) rate of 17 BB/100, where I have limited my playing to peak times and improved my strategy. I can also say that I have had quite a number of bad beats that made me want to throw a shoe through my monitor, so it's not like I have been getting perfect cards this entire time.

Am I correct in thinking this or is there a nasty 25,000 hand downswing on the way? That's kind of what I was thinking about this... I haven't played that many hands online, so someone who has played millions of hands would have a better idea of this type of thing...

Edit: The more I think about it, the more these BB/100 numbers seem out of whack in comparison to what others are talking about - you mentioned 1 BB to 3 BB/100, and I am talking about 5-10 times that amount - yet that is what Poker Office is reporting... anyone else use that program and are these numbers correct?

[mbillie1]

Upswing: 5 grand playing drunk one weekend. Downswing: 4 grand playing drunk later that week.

And no that is not a joke. I play 100 and 200 NL, but I was playing 1000 NL and ran hot for a while, thought I was good, then blew most of what I won. In short, it's bad to play above your skill level lol.

[Palomino]

[ QUOTE ]
I guess what I am getting at is trying to figure out whether my recent play against certain bad players is really just variance or not... obviously I have no doubt that I can consistently make money from these people because right now I am just killing them, but the question is, how much of my current return is variance and how much can I expect to have it continue?

I just started on a new site where the players are complete garbage, and my return over the first 18,000 hands is on average 9.55 BB/100... and that includes a long period where quite frankly I was playing too loose. If I just take the last 7000 hands where I have limited my play to peak times, switched tables at the right time, and honestly just played better, my return is around 17 BB/100, with one period of 2000 hands where it was 33 BB/100.

Now, I am pretty sure that the 2000 hand spurt of 33 BB/100 was variance, and I am also pretty sure that over 17,000 hands I can count on around 10 BB/100 (since that seems like too many hands for simple statistical deviation), but I'm wondering if 7000 hands is enough to determine the new (improved) rate of 17 BB/100, where I have limited my playing to peak times and improved my strategy. I can also say that I have had quite a number of bad beats that made me want to throw a shoe through my monitor, so it's not like I have been getting perfect cards this entire time.

Am I correct in thinking this or is there a nasty 25,000 hand downswing on the way? That's kind of what I was thinking about this... I haven't played that many hands online, so someone who has played millions of hands would have a better idea of this type of thing...

Edit: The more I think about it, the more these BB/100 numbers seem out of whack in comparison to what others are talking about - you mentioned 1 BB to 3 BB/100, and I am talking about 5-10 times that amount - yet that is what Poker Office is reporting... anyone else use that program and are these numbers correct?

[/ QUOTE ]

You playing NL or limit? Over 18k hands there is no way anyone playing limit can be running at 10 BB/100 IMO. So I'm guessing you play NL.

[etxel]

I think there's got to be a number of hands played after which you can tell if your win or loss is due to variance or luck. I don't know that number but I think there must be one.

Its like for example if you flip a coin a 1000 times. There is a chance that you will always get heads, but that chance is so small that it is essentially zero. The same way, there must be a certain number of poker hands after which, the chance of your results being reflective on pure luck is essentially zero. But, I don't know what that number is. And what would determine a good result?

[Eric Stoner]

[ QUOTE ]
I guess what I am getting at is trying to figure out whether my recent play against certain bad players is really just variance or not... obviously I have no doubt that I can consistently make money from these people because right now I am just killing them, but the question is, how much of my current return is variance and how much can I expect to have it continue?

I just started on a new site where the players are complete garbage, and my return over the first 18,000 hands is on average 9.55 BB/100... and that includes a long period where quite frankly I was playing too loose. If I just take the last 7000 hands where I have limited my play to peak times, switched tables at the right time, and honestly just played better, my return is around 17 BB/100, with one period of 2000 hands where it was 33 BB/100.

Now, I am pretty sure that the 2000 hand spurt of 33 BB/100 was variance, and I am also pretty sure that over 17,000 hands I can count on around 10 BB/100 (since that seems like too many hands for simple statistical deviation), but I'm wondering if 7000 hands is enough to determine the new (improved) rate of 17 BB/100, where I have limited my playing to peak times and improved my strategy. I can also say that I have had quite a number of bad beats that made me want to throw a shoe through my monitor, so it's not like I have been getting perfect cards this entire time.

Am I correct in thinking this or is there a nasty 25,000 hand downswing on the way? That's kind of what I was thinking about this... I haven't played that many hands online, so someone who has played millions of hands would have a better idea of this type of thing...

Edit: The more I think about it, the more these BB/100 numbers seem out of whack in comparison to what others are talking about - you mentioned 1 BB to 3 BB/100, and I am talking about 5-10 times that amount - yet that is what Poker Office is reporting... anyone else use that program and are these numbers correct?

[/ QUOTE ]

My numbers are from limit. In no limit poker, using Poker Tracker terminology, a consistent winner could make between 5-7 PTBB/100 (double the big blind is a PTBB).

However, you are correct in surmising that it could be a factor of having the cards hit along with playing your cards well for maximum value.

If you are making the best decisions that you can, I wouldn't worry about the negative variance to come. Just focus on your decisions and just recognize that there will be a time when the cards will run against you.

You just can't predict the exact moment with certainty. Variance is determined by what happened and not what will happen.

[trader01]

OK, now this makes perfect sense - I am playing no limit, and I assumed that other people were too... I can see how the numbers would be totally different for limit hold'em. So actually, if we take 5-7 PTBB/100 hands and equate that to my latest figures of around 17 BB/100 (roughly 2-3 times), then that makes sense and gives me a good idea of where I should be running (which I what I was feeling anyways). Against really bad players it's not hard to do that, you just have to avoid going on tilt from all the ridiculous suckouts that happen when some idiot calls your AA all-in with 2-7 offsuit and wins. :-)

[PJS]

[ QUOTE ]
Here's the question: variance obviously plays a big part in texas hold'em, with players getting "hot" or "cold" for periods of time, but I haven't seen much talk about the maximum number of hands that anyone can honestly attribute to variance.

For example, if someone gets "hot" for 100 hands, that's clearly variance and cannot be attributed to that poker player's skill - clearly anyone can get a good run of cards over only 100 hands. Similarly, if a player's profit graph is straight down for 100,000 hands, that is clearly more than just "bad luck" or variance - if someone is going straight down for that many hands, it's because they are not playing the proper strategy or don't know how to play.

So the question is, at what point (in terms of number of hands) can we say that a player's skill truly starts to show and that the slope of the graph is not due to variance? Similarly, how many hands would constitute the maximum reasonable run of "abnormal" good (or bad) luck that a player can expect to see from time to time due to variance?

I'm interested to hear any stories...

[/ QUOTE ]

Here is a detailed thread regarding variance.

[MasterLJ]

My horrendous downswing that caused me to withdraw all my money from online poker was about 15k hands and about $5k at $1/$2 NL. I've never seen anything so disgusting in my life.

I used to think 20+ buy-in downswings were BS, but they are there. Probably 5-7 of my 25 BI downer could be attributed to bad play due to running bad, I tried and tried to overcome and not tilt, but I couldn't. Taking a few weeks off. Probably going to rebuild at $100 to get my confidence back, but I won't be on for a long while.