Bankroll requirements for PP STT's

[SNOWBALL138]

I'm trying to become a single table tourneyment specialist on Party poker. What do people think the bankroll requirements are for various STT levels? My assumption so far has been that for 10-1 STT's $100 is good, and for $20, 200 is good, etc. Is this enough to cover standard fluctuations?

I turned $50 into 450 on the low level STT's, but then I set my bankroll on fire at the $50 STT's and had to go back down to lower levels.

I'm the kind of guy that doesn't like to reload my account on party poker. I always prefer to have one way transactions (i.e. cashouts).

[Sidekick]

I'll try and give you a reasonable answer before you possibly get torched for not searching on this topic (as it has been covered hundreds of times /images/graemlins/smile.gif ).

The general consensus here on these boards seems to be you need at least 30 buy ins to weather the inevitable downswings. That means if you are playing the Party $10+$1 SNGs then you need a minimum of $330 as a starting bank roll. This number assumes that you are a solid winning player as well. If you aren't a winning player, then no BR is safe.

I personally am in the 50 buy in camp myself and I've seen some posters espouse even higher BRs.

No matter how good you are, you are going to have some major downswings from time to time. The less skilled you are the greater the downswings will be.

Some very good players who post on these boards have gone bust. Expect large swings in your BR, especially when you are just starting out. Don't assume that because you have won 'X' amount of dollars in your first 50-75 SNGs that you will continue to do so because you have proved that you can win over a large number of SNGs. If you have played less than 200 SNGs your sample size is little more than a blip and any ROI that you have so far is subject to drastic change.

Hope this helps.

[ericlambi]

I really question all the posters on here that believe 200 SNGs is a blip and may not represent your true capability. I have been playing SNGs for a while, but just created a spreadsheet to actually track my results this weekend. I've played 43 SNGs this weekend, here are some statistics:

Profit/SNG: $30.67
Std. Deviation: $76.63
t-value (95% confidence interval, DOF=42): 2.018

True profitability at SNGs with 95% confidence:

$7.09 <= Profit/SNG <= $54.26

If you see something wrong with my statistical analysis, please feel free to post. If you think statistics are some kind of hokum, and that this can't possibly be correct just because you don't believe it - don't bother to respond.

[stupidsucker]

[ QUOTE ]
I really question all the posters on here that believe 200 SNGs is a blip and may not represent your true capability.

[/ QUOTE ]

This month I have played 283 SnGs all at the 30s level.
The first 112 gave me 44.64% ITM with a 41.23% RoI
The next 171 was 32.16 % ITM with 0.48% RoI

Over 2kplus tournyes at this same level I am a 25% roi player.

It isnt fiction my friend. A small sample size will always and forever be just a small sample size.

But as for bankroll (The real thread)

30x buy in is a pretty good mark to shoot for from the 10s-30s. It isnt foolproof even for a winning player, but comes with a low ror if you know how to play.

[Pokerscott]

[ QUOTE ]
I really question all the posters on here that believe 200 SNGs is a blip and may not represent your true capability.

[/ QUOTE ]

Depending on how good you performed and how good you think you are, 200 can give you pretty decent evidence of your skill level.

See this thread (and linked threads) for details...

ROI significance levels (http://forumserver.twoplustwo.com/showflat.php?Cat=&Number=1570111&page=6&view=colla psed&sb=5&o=14&fpart=1)

Pokerscott

[AleoMagus]

Really, I thought about the many ways I could answer this post for a while. I even started typing a few different responses, but in the end I think all that there is to say is:

Play 200 more and then tell us what you think of your first 43 tourney sample.

Regards
Brad S

PS - What stakes are you playing to get those numbers?

[byronkincaid]

[ QUOTE ]
This month I have played 283 SnGs all at the 30s level.


[/ QUOTE ]

You had a couple of weeks off or something? Wondered why I didn't see you or ben on the empire leader board anymore.

[ericlambi]

Take a statistics class. 200 samples of normally distributed random variable is not a small sample.

[AleoMagus]

You are assuming the the posters on this board do not understand statistics, when in fact, you might be surprised how intimate a knowledge they have of it's real meaning.

First of all, while 200 SNGs may not be a small sample when calculating something like winning confidence, it is very small when trying to approximate what you describe in this statement

[ QUOTE ]
I really question all the posters on here that believe 200 SNGs is a blip and may not represent your true capability.

[/ QUOTE ]

Ok, lets examine a typical SNG player after he has played 200 SNGs. Lets assume he has so far acheived a ROI of 25% and that his SD is 1.7 buy-ins. So, at 50+5, this would give him a profit per tourney of $13.75 and a SD of $93.5.

What about his 'true capability' and what stats can tell us?

well, he can be 98% confident that he is a winning player, but just what kind of winning player? well, after this many tourneys, he can be 95% confident of a profit per tourney of $1 to $26 per tourney!

This means after 200 SNGs, he might have an ROI of anywhere from 1% to 47%! If that is what you define as a good representation of 'true capability' then fine. whatever.

Your sample of 43 is even more subject to statistical fallacy.

You have been running well. If you do not think that you have been, you are seriously in for dissapointment. I have tried to figure how you could be getting $30 per SNG with only a $76 SD and to me, it can only mean a huge and unsustainable ROI.

Sure, your stats tell you that you can be very confident that you are a winning player, but make no mistake, that is all that they are telling you. Do not be surprised if after 200, you are winning significantly less.

The worst part of doing these stats with a huge (unsustainable) ROI, is that you need to weigh in other's experiences when approximating in your own mind if you are actually equally likely to be earning $7 as $54. This is to say that there is a very small chance that you are earning even as much as you currently think that you are.


[ QUOTE ]
Take a statistics class

[/ QUOTE ]

comments like this... sheesh

Regards
Brad S

PS - I'm still very interested in what stakes you ahve been playing to get those numbers.

[ilya]

SNOWBALL --

If you don't like to reload, use 60xbuy-in and drop down if you fall below 30xbuy-in.

[bball904]

[ QUOTE ]
Profit/SNG: $30.67
Std. Deviation: $76.63
t-value (95% confidence interval, DOF=42): 2.018

True profitability at SNGs with 95% confidence:

$7.09 <= Profit/SNG <= $54.26

[/ QUOTE ]

You believe that you can have 95% confidence that your long term profit is in that range after only 43 tournaments? How did you know that I needed a good laugh?

I took a few statistics classes on the way to my statistics degree and you, my friend, are stoned.

[Voltron87]

Bankroll requirements for SNGS, any level>>>>>what you think.

[KenProspero]

For the mathmeticians amongst us --

Wouldn't the necessary bankroll depend on what long term %ROI, a player would wind up at?

For example, let's say that if a player should have a +10% ROI in the long run. It seems to me that this player would need a bigger bankroll to account for variation than a player who should have a 20% ROI long run.

Whenever I see rules of thumb (e.g., the 300 Big Bet rule for limit), I wonder what ROI is being tested for to make it statistically valid.

[ericlambi]

I think everyone has missed the point of my post. I didn't post because I wanted to brag about my weekend results or get affirmation from anyone that I am a profitable player. The point of the post was to put to rest the fallacy that 200 SNGs is the minimum sample size required to determine if you are a profitable player by posting real actual MATH.

Apparently though, no one is interested in real math and would prefer to either talk about their 200 SNG bad streak like it was all the random number generator's fault, or to laugh at my analysis while claiming to have a statistics degree, but failing to point out where the analysis goes wrong.

Of course I realize that all the confidence interval says is that there is a 95% chance my true mean is between $7 and $54, or whatever the #s were. But that still leaves a high certainty of profitable play with N=43, which is the point I was trying to express.

[AleoMagus]

You are very correct, and this drives me nuts too.

the reason people throw the 20-30x Buy-in around is that this gives about a 1% ROR for 20%-30% ROI. This is a bad assumption though, and the real guideline with the actual math looks more like this:

For a 1% ROR,
ROI Buy-ins
05% - 133.1
10% - 66.5
15% - 44.4
20% - 33.3
25% - 26.6
30% - 22.2
35% - 19.0
40% - 16.6

Note: These figures assume a SD of 1.7 Buy-ins. This is of course, an approximation, albeit a useful one that will approximate most players actual SD.

The actual calculations to determine a specific bankroll requirement or a specific ROR are:

B=-(SD^2/2W)LN(R)

r=EXP(-2WB/SD^2)

where,
W is your average profit per tourney ($)
SD is your standard deviation per tournament ($)
R is your desired risk of ruin
B is your bankroll ($)

Regards
Brad S

[Sidekick]

Hmmm... I think several people are misreading a few things (slightly, but that can sure change the context...).

No one has said that you cannot have a reasonable degree of confidence that you are a winning player after 200 SNGs. What I (and if I read the other responses correctly anyways) and others are saying is that you cannot be confident that your true long term ROI is really close to what you have actually posted over a 200 SNG set.

Say you post a 25% ROI over 200 SNGs. You can be confident that you are a winning player, but your ROI can be off by more than 20% either way over such a small sample. Now to me (and I suspect to many other posters here) having a true ROI that could be anywhere from 5% to 45% means that I cannot have any real certainty what my true long term ROI is.

The one point to your original post that I would argue is your assumption that SNG results are on a normal random distribution. Other than that I don't find any problems with your math, but even your posted results indicate a very large range in possible ROI.

[meatballAK]

what does 1% ROR mean? 1% chance for a month, year, lifetime, infinity?

[AleoMagus]

[ QUOTE ]
The point of the post was to put to rest the fallacy that 200 SNGs is the minimum sample size required to determine if you are a profitable player by posting real actual MATH.


[/ QUOTE ]

That may have been the point of the post, but it's not how it came across. You seemed to be saying a lot more than this.

Sidekick said:

[ QUOTE ]
If you have played less than 200 SNGs your sample size is little more than a blip and any ROI that you have so far is subject to drastic change.

[/ QUOTE ]

to which you said:

[ QUOTE ]
I really question all the posters on here that believe 200 SNGs is a blip and may not represent your true capability.

[/ QUOTE ]

Well, call me crazy, but it seems to me that you were suggesting that after 200 SNGs, sidekicks comments about ROI still being subject to drastic change were wrong.

After 200 SNGs, ROI is subject to drastic change and I demonstrated that with

[ QUOTE ]
real actual MATH.

[/ QUOTE ]

Stupidsucker posted a real life sample which demonstrated the same, and you didn't seem to like this.

All that said, if your real intent was to show that you can be confident to 2 SD that you are a winner after 43 SNGs, well OK. I'll grant you that but barely, as after 43 tourneys, SNG results only start to approximate a normal distribution.

I'll also point out that we know this already.

Regards
Brad S

[AleoMagus]

1% Forever, if you never cashed out any winnings. In reality, you will cash out winnings, so eventually, even a small ROR will be realized.

if you want to cash out your winnings, your ROR will be effectively be 'reset'. It really doesn't matter though as anything you have cashed off can effectively still be considered bankroll in a way.

I hate bankroll threads. It's all just a psychological perspective at the low stakes games and doesn't matter at all in my opinion.

From now on, when anyone asks about required bankroll, my standard answer will be 1 buy-in, and I will mean it.

Regards
Brad S

[meatballAK]

No, keep posting. The math is interesting for those of us without a statistics background.

Thanks.

[eagle]

I think part of the problem here is that some of the posters keep forgetting that these projections all have an assumption in them that the "player" will continue to play at the same "skill level" ( Not buy-in level) into the future. This is what causes the difference between what the stat guys project and what the "winning posters" warn about when relying on statistics.

( I'm sure someone has already explained this in excruciating detail in the archives given some of the insightful posts I've read during my first two weeks here)

To put it simply you cannot rely on past performance in poker like you can rely on the dice in Craps. So if we were robots the stat guys would be right, but were not so I will side with the more experienced players and their insight.

And if you want to further complicate the analysis, the skill level of the other players that you play against on average also varies( i.e the $11 players may get better or worse as group and/or adapt to your style of play making it less profitable). Remember, the statistics cannot predict the future they can only make assumptions about it. See the robot analogy above.

So this is another point that says LISTEN to the "long-term winning players" when they say that 200 SNGs is a low number.

As far as bankroll, listen to the winning posters (30-50 Units) and don't worry so much about the statistics.

Also, I agree with Aleo. The stats people on this sight do know their stuff.

[fiyah]

To Ericlambi, not to sound mean or come of rude but I must say one thing "Small Sample Size", honestly I was/am guility like you when I started playing SnGs exclusively. I have about 2000 SnGs tracked from 10s to 50s and I must say that my progression was rapid and I felt as you do that this is a very profitable game. However, lately I've been seeing the "variance" factor kick in, for example AA on the bubble all in getting cracked constantly, dominating hands losing to dominated hands. All this is part of poker and its the reason that there is variance. The thing is since I've taken these losses its made me re-evaluate my game. Really made me look back at when I was "winning" and seeing how often I was winning cuz I was in coin flip situations and I just got the best of it and when I didn't need to put my stack in these situations and what not. I guess really all I'm saying is don't mistake running good over a short period of time as crushing the game.