Theoretical run bad scenario?

Mikeisanace777

Mikeisanace777

Rock Star
Silver Level
Joined
Jan 24, 2017
Total posts
150
Chips
0
I haven't played enough poker in terms of millions of hands like you guys,but I have a theory called theoretical run bad predictability scenario. Everyone runs good,then bad how do you reduce this besides playing better? Is it possible to chart out and figure out what hand your on and the fact that a run bad is going to occur then deduce that it takes x amount of time to play these so you simply stop playing if the condition are just simply normal as excepted. To further analyze poker live,or online it comes down to a rng,heck even the shuffle master uses a computer bases rng to shuffle the cards it's not mechanical. So knowing this it comes down to science of randomness which may not be as it seems. If you can figure out that the next 5,000 hands will be even to a loss and anywhere from the 12-15k mark is in the money then stop playing correct? If your a live player this could be a month,online much less time,but it still does occur so might as well let those hands occur as if you were there but you ain't so when you return your in the red zone.
 
B

braveslice

Pull-ups!
Silver Level
Joined
Dec 9, 2013
Total posts
1,988
Chips
0
Sorry this would require predicting future, in a case where you can't predict even the next hand. Inventing time machine might solve the problem though. You wouldn't be able to jump back 5k hands unfortunately, because what ever you do in a new time line changes the future, but you probably would be able to jump to the previous hand, effectively doing the same.
 
Last edited:
Stuey

Stuey

Enthusiast
Silver Level
Joined
Sep 11, 2017
Total posts
76
Chips
0
When talking about running bad you talk about the past. Each new hand is in the present and has nothing to do with past events. It may seem to you that you are running bad but all you do is look back in time.

All is nonsense
 
Mikeisanace777

Mikeisanace777

Rock Star
Silver Level
Joined
Jan 24, 2017
Total posts
150
Chips
0
No guys your not getting it.

The poker game is a never ending semi cyclical of random events dumped out by a machine the rng. Past result do indicate in a fashion how the game operates just like video poker has a cycle of 438.4 hands of the 4 of a kind. Analyze 1 million hands and that cycle always holds true because it's a machine operating on a rng just like poker. In any event the cycle of you as the winning hand in nl could be one in 11.4 hands as in no matter what if you just rand the cards out. Now you have to factor in bluffing and hands you didn't play that would have won but it probably still holds true. Over time this can fluctuate as dead zones trust me I have analyzed video poker with a computer program in that one million because sometimes you go 1,500 hands with nothing then it seems to hit hard like 3-4 quads it the next 300 hands,it's never ending and so is poker live, online wont matter.Running bad is simply dead zones in the poker game of your 11.4 hand cycle of the winning run out hand that seems to go in the hundreds of junk and bad beats because it wasn't your hand. Then of course you run hot as it seems as the cycle seems to magically come alive and your winning every 3 hands or so in your game hitting cards cause.. The game is cyclical from past events in the true cycle of the 11.4 hit cycle, again like video poker and the quad reference being it's cycle as 438.4 or so it often goes dead for 1,500 to 2500 hands track this and you know your hot in the next 300 hands Iv'e tried this it happens I hit 2-5 of them in that short time period. This is all nearly predictable from past data cause if the game is cruising along as predicted then it's doing as it should,if your in the dead zone then a hot occurrence will occur sooner or later,and of course if your winning more than expected than it only has one place to go and thats down to even out that 11.4 hand win cycle of never ending RNG activity.











poker Hand Payoff % of Return % Contribution Appeared Frequency Cycle
ROYAL FLUSH 4000 1.95% 1.92% 62.30 0.00% 41714.2
FOUR of a KIND 400 19.15% 18.87% 6128.76 0.24% 424.1
STRAIGHT FLUSH 250 0.55% 0.55% 283.89 0.01% 9154.7
FULL HOUSE 40 9.33% 9.19% 29848.89 1.15% 87.1
FLUSH 30 6.78% 6.68% 28941.34 1.11% 89.8
STRAIGHT 20 5.19% 5.11% 33232.19 1.28% 78.2
THREE of a KIND 15 22.55% 22.21% 192431.00 7.40% 13.5
TWO PAIR 5 13.03% 12.83% 333568.68 12.83% 7.8
JACKS or BETTER 5 21.45% 21.13% 549170.85 21.13% 4.7
GARBAGE 0 0.00% 0.00% 1425292.10 54.84% 1.8
 
Last edited:
D

Dogwisddom

Rising Star
Bronze Level
Joined
Nov 27, 2015
Total posts
16
Chips
0
Volume exposes if you are running bad or just playing bad.

The answer is to put in more hands and play your A-game.
 
A

AlexTheOwl

Visionary
Silver Level
Joined
Aug 22, 2017
Total posts
860
Chips
0
You are not making yourself clear.

This forum doesn't charge extra for punctuation marks, or for using the enter key.

To the extent that I can make any sense of this, you seem to be talking about regression to the mean and the gambler's fallacy.

In a large enough sample size, random events occur with the frequency that math predicts. For example, the chance of being dealt AA is 1 in 220, or 0.45%. In a very large sample of randomly dealt starting hands, about 0.45% of those hands will be AA.

However, that doesn't mean that every 220 hand slice of that large sample will contain one and only one AA. Some will contain several AA's. Some will contain none.

Let's say we take a sample size of one hundred thousand randomly dealt hold'em hole cards.

Before we deal the cards, Regression to the Mean validly predicts that no matter how AA's cluster in any given slice of our sample, by the end we are likely to have about (100,000 * .045) = 4500 AA's dealt.

The "gambler's fallacy" comes into play when we assign causality to regression to the mean.
Regression to the mean does not predict the number of events in any given slice of our sample.

In our sample of one hundred thousand hands, we can predict that 90% of the expected 4500 AA hands will be dealt when 90% of the total sample is complete. So when we reach 90,000 random hands, we should expect (4500*.90) = 4050 AA hands to be dealt.

What if only 4000 AA hands have been dealt? What is the most accurate prediction for the number of AA hands that will be dealt in the last 10,000 hands of our sample?
It is simply (10,000*.045) = 450 AA hands.

We should not adjust our prediction for the number of AA hands in the last 10,000 hands of our sample based on the prediction we made before the hands were dealt.
Instead, we should adjust our prediction for the number of AA hands in the whole 100,000 hand sample based on the number of AA hands that have been dealt so far.
The cards have no memory.
They do not know they are part of a 100,000 hand experiment.
They will just continue to be random.

Even though we started the process with a valid expectation that about 4500 AA hands would be dealt, Regression to the Mean DOES NOT predict that 500 AA hands will be dealt in these last 10,000 hands in order to "catch up to" our original expectation.
We are not "due" for an above average number of AA hands dealt in this last 10,000 hand slice of our sample. Believing that we are is the "gambler's fallacy".

TL;DR:
Random events, by definition, do not occur in cycles. They are usually not distributed evenly across a sample. They occur in clusters. Clusters are not necessarily part of a cycle, and for truly random events they never are cyclical. Cycles are predictable, and random events, by definition, are not.
 
B

braveslice

Pull-ups!
Silver Level
Joined
Dec 9, 2013
Total posts
1,988
Chips
0
I'm getting that he means that random generators are not truly random, this used to be practical problem not long ago. In environments which are static this can be still a problem I would guess, but not in environment which includes true randomness where seed can be generated easily, like poker involving human action, latency and so on.

Edit, just googled this: "RANDOM.ORG is a true random number service that generates randomness via atmospheric noise. This page explains why it's hard (and interesting) to get a computer to generate proper random numbers."
 
Last edited:
A

AlexTheOwl

Visionary
Silver Level
Joined
Aug 22, 2017
Total posts
860
Chips
0
I'm getting that he means that random generators are not truly random, this used to be practical problem not long ago. In environments which are static this can be still a problem I would guess, but not in environment which includes true randomness where seed can be generated easily, like poker involving human action, latency and so on.

Maybe. He talks a lot about RNGs, and mentions that even ShuffleMasters use them.

As you say, software RNGs use a seed.
The series of "random" numbers generated will always be the same if the seed is the same.
But it's always been so easy to keep changing the seed in an unpredictable way.
In the old days this was usually based on the current time when a new random sequence is generated.
A new sequence would be generated after a certain count of random numbers was used, and the time of day that would occur would depend on user actions. Users would "use up" a sequence at an unpredictable pace.

The random sequences are applied across various tables.
The player has no way of knowing the current seed, the sequence generated by it, which section of the sequence is currently being generated, or which number from the sequence will generate their hole cards or the board cards.

So from the player's point of view, there is no predictability, no cycle, even in these old software-only RNGs.
 
B

braveslice

Pull-ups!
Silver Level
Joined
Dec 9, 2013
Total posts
1,988
Chips
0
So from the player's point of view, there is no predictability, no cycle, even in these old software-only RNGs.


Yes I agree.

Even the best case scenario is too impractical to generate more than marginal, non-meaningful difference. It's never good to ask mathematician about practical matters.
 
Top