Apophenia in Slots

Quote: DoubleGold

…
Could it be you have apophenia?
…
link to original post

I do! Episodes well up especially when I am playing. The best I can do is recognize this weakness and try to not let it influence my actions.

Most, if not all people have this condition. It is a natural trait of the human mind. Read Thinking, Fast and Slow by Daniel Kahneman if you would like to better understand how the human mind works. The whole book is an excellent read, with sections addressing such things as our innate inability to properly process small sample statistics.

Nice try at a veiled ad hominem attack. I’ll be getting my coat …

I brought up the possibility that we are just being trolled, so for me to continue to respond means I am considering the possibility that you are not doing so. I am done with trying to point out some points you are not accepting

Instead, I am going to point out some other things

*you are not convincing anyone here, and that should mean something.

*for gamblers to adhere to fallacies is a well known problem, so you need to demonstrate you are picking up on this. So far you have failed utterly, even embracing the most ridiculed fallacy out there, that streaks of, say, heads in coin tosses, make tails more likely.

*you are using gambling language oddly sometimes. RTP, for example, you are citing frequently but at odd places. This pretty much says you are not well read on the subject. Relative and absolute probability is a topic of study, alright, but you are not showing you know much about it. I don’t think it links back to Einstein. The prime number thing is grasping at straws.

*you have asked for lenience in the matter of questioning consensus. I can accept that, but I think it is pretty incumbent upon you to come up with some smoking hot evidence when pouncing on what has been called a fallacy. It is scientific to question things, but there is the scientific method to consider, and we need that. You have a hypothesis, so how about some testing? You can’t just make the assertion and leave it at that, everybody then has the right to just tell you to stfu.

*you really have to consider more the factor of the billions of gamblers who have already considered the question you are raising. How do I know it is billions? Because it occurs to every single gambler who has stayed with it, excluding perhaps those few who stay casual players. None of those billions of people has cracked the problem. If you are going to, you better get cracking. I wouldn’t waste time debating in such a silly manner on the internet.

Quote: BleedingChipsSlowly

Mods: Can this thread be moved to the Betting Systems sty?
link to original post

Indeed, it has been so stymied.

Quote: BleedingChipsSlowly

Quote: DoubleGold

…
Could it be you have apophenia?
…
link to original post

I do! Episodes well up especially when I am playing. The best I can do is recognize this weakness and try to not let it influence my actions.

Most, if not all people have this condition. It is a natural trait of the human mind. Read Thinking, Fast and Slow by Daniel Kahneman if you would like to better understand how the human mind works. The whole book is an excellent read, with sections addressing such things as our innate inability to properly process small sample statistics.

Nice try at a veiled ad hominem attack. I’ll be getting my coat …
link to original post

I'm trying to read it now, TY.

I question if any of us and all of us have apophenia.

I question if you have apophenia, even after admitting it.

But if apophenia is real and I have it, I am willing to attempt to recognize it and correct it.

In regards to this thread placement, I wouldn't object to another placement.

In regards to your coat, I appreciate your feedback.

Quote: odiousgambit

I brought up the possibility that we are just being trolled, so for me to continue to respond means I am considering the possibility that you are not doing so. I am done with trying to point out some points you are not accepting

Instead, I am going to point out some other things

*you are not convincing anyone here, and that should mean something.

*for gamblers to adhere to fallacies is a well known problem, so you need to demonstrate you are picking up on this. So far you have failed utterly, even embracing the most ridiculed fallacy out there, that streaks of, say, heads in coin tosses, make tails more likely.

*you are using gambling language oddly sometimes. RTP, for example, you are citing frequently but at odd places. This pretty much says you are not well read on the subject. Relative and absolute probability is a topic of study, alright, but you are not showing you know much about it. I don’t think it links back to Einstein. The prime number thing is grasping at straws.

*you have asked for lenience in the matter of questioning consensus. I can accept that, but I think it is pretty incumbent upon you to come up with some smoking hot evidence when pouncing on what has been called a fallacy. It is scientific to question things, but there is the scientific method to consider, and we need that. You have a hypothesis, so how about some testing? You can’t just make the assertion and leave it at that, everybody then has the right to just tell you to stfu.

*you really have to consider more the factor of the billions of gamblers who have already considered the question you are raising. How do I know it is billions? Because it occurs to every single gambler who has stayed with it, excluding perhaps those few who stay casual players. None of those billions of people has cracked the problem. If you are going to, you better get cracking. I wouldn’t waste time debating in such a silly manner on the internet.
link to original post

I think you're fishing for information, for example, a troll image, etc.

Like wanting me to claim I am not a troll, or any of your other fishing expeditions.

But, you are allowing me to consider concepts, so I appreciate your feedback.

I think part of the issue is in my writing what I am trying to convey.

It could be most of us agree, just the wording is conveyed differently.

Quote: rsactuary

What the law of large numbers is: The next 1000 flips of that coin are expected to come 500 heads and 500 tails, meaning that after 1020 flips, you can expect to be at 513 heads which is about 50.3%. That is how regression to the mean works.
link to original post

So if I plan to flip a coin 1000 times, and after 500 flips I am at 400 heads and 100 tails, shouldn't I expect there to be a much larger number of tails in the next 500 flips?

Eventually, the relative probability will be exactly the same as the absolute probability (50% or 85% or 90%, etc.).

Without seeing any internal raw data, I envision angular trending (time series) back and forth over the absolute line.

But some machines might be major losers and some might be major winners that offset each other.

Quote: TigerWu

Quote: rsactuary

What the law of large numbers is: The next 1000 flips of that coin are expected to come 500 heads and 500 tails, meaning that after 1020 flips, you can expect to be at 513 heads which is about 50.3%. That is how regression to the mean works.
link to original post

So if I plan to flip a coin 1000 times, and after 500 flips I am at 400 heads and 100 tails, shouldn't I expect there to be a much larger number of tails in the next 500 flips?
link to original post

Assuming the coin is fair, no you shouldn't.

Quote: rsactuary

Quote: TigerWu

Quote: rsactuary

What the law of large numbers is: The next 1000 flips of that coin are expected to come 500 heads and 500 tails, meaning that after 1020 flips, you can expect to be at 513 heads which is about 50.3%. That is how regression to the mean works.
link to original post

So if I plan to flip a coin 1000 times, and after 500 flips I am at 400 heads and 100 tails, shouldn't I expect there to be a much larger number of tails in the next 500 flips?
link to original post

Assuming the coin is fair, no you shouldn't.
link to original post

How does this not contradict your earlier point:

"....after 1020 flips, you can expect to be at 513 heads."

After 500 flips, I'm at 400 heads. So, according to what you said earlier, I should only be expecting about 100 more heads over the next 500 flips according to "regression to the mean."

Quote: TigerWu

Quote: rsactuary

What the law of large numbers is: The next 1000 flips of that coin are expected to come 500 heads and 500 tails, meaning that after 1020 flips, you can expect to be at 513 heads which is about 50.3%. That is how regression to the mean works.
link to original post

So if I plan to flip a coin 1000 times, and after 500 flips I am at 400 heads and 100 tails, shouldn't I expect there to be a much larger number of tails in the next 500 flips?
link to original post

Nope, you would average 250 heads and 250 tails.

Quote: DRich

Quote: TigerWu

Quote: rsactuary

What the law of large numbers is: The next 1000 flips of that coin are expected to come 500 heads and 500 tails, meaning that after 1020 flips, you can expect to be at 513 heads which is about 50.3%. That is how regression to the mean works.
link to original post

So if I plan to flip a coin 1000 times, and after 500 flips I am at 400 heads and 100 tails, shouldn't I expect there to be a much larger number of tails in the next 500 flips?
link to original post

Nope, you would average 250 heads and 250 tails.
link to original post

So that's a total of 650 heads and 350 tails, right? Tails is still behind by 300. Since everything should be "regressing to the mean," at some point I should expect tails to "catch up" to heads, right? So over the long run, they will both be about 50%, right?

Quote: TigerWu

Quote: DRich

Quote: TigerWu

Quote: rsactuary

What the law of large numbers is: The next 1000 flips of that coin are expected to come 500 heads and 500 tails, meaning that after 1020 flips, you can expect to be at 513 heads which is about 50.3%. That is how regression to the mean works.
link to original post

So if I plan to flip a coin 1000 times, and after 500 flips I am at 400 heads and 100 tails, shouldn't I expect there to be a much larger number of tails in the next 500 flips?
link to original post

Nope, you would average 250 heads and 250 tails.
link to original post

So that's a total of 650 heads and 350 tails, right? Tails is still behind by 300. Since everything should be "regressing to the mean," at some point I should expect tails to "catch up" to heads, right? So over the long run, they will both be about 50%, right?
link to original post

You did regress to the mean. You were at 80% heads at 500 flips, you're at 65% heads at 1000 flips. Variance may ultimately let tails catch up, but you can't assume that.

Quote: TigerWu

Quote: rsactuary

What the law of large numbers is: The next 1000 flips of that coin are expected to come 500 heads and 500 tails, meaning that after 1020 flips, you can expect to be at 513 heads which is about 50.3%. That is how regression to the mean works.
link to original post

So if I plan to flip a coin 1000 times, and after 500 flips I am at 400 heads and 100 tails, shouldn't I expect there to be a much larger number of tails in the next 500 flips?
link to original post

If the flipper keeps flipping as they have been, I wouldn't expect 400 tails.
I would expect either about 100 tails or about 250 tails.
I also wouldn't go carelessly betting on coin flips with whoever is flipping the coin.

I think there's a scientific law (as long as the initial imbalance is not artificial) that not only it can catch up, but it must catch up.

Provided time is not limited and a few other params, like doing the actual flipping, having no bias, etc.

For slot machines, the casinos have to report regularly to the authorities, so they have to average out at least annually.

Quote: DoubleGold

I think there's a scientific law (as long as the initial imbalance is not artificial) that not only it can catch up, but it must catch up.


link to original post

Please provide evidence of that law. This contradicts the law of large numbers.

Quote: rsactuary

Quote: DoubleGold

I think there's a scientific law (as long as the initial imbalance is not artificial) that not only it can catch up, but it must catch up.


link to original post

Please provide evidence of that law. This contradicts the law of large numbers.
link to original post

It's common sense.

A coin flip's absolute probability is 50%.

It must hit that number or it's not random.

It would be a scam.

The relative probability can be varied.

Therein lies the profit potential.

The difference of the relative probability and the absolute probability.

Quote: DoubleGold

Quote: rsactuary

Quote: DoubleGold

I think there's a scientific law (as long as the initial imbalance is not artificial) that not only it can catch up, but it must catch up.


link to original post

Please provide evidence of that law. This contradicts the law of large numbers.
link to original post

It's common sense.

A coin flip's absolute probability is 50%.

It must hit that number or it's not random.

It would be a scam.

The relative probability can be varied.

Therein lies the profit potential.

The difference of the relative probability and the absolute probability.
link to original post

I'm sorry that you don't understand the law of large numbers. My work here is done.

Quote: DoubleGold

It's common sense.

A coin flip's absolute probability is 50%.

It must hit that number or it's not random.

It would be a scam.

The relative probability can be varied.

Therein lies the profit potential.

going along with your theory that the coin takes active measures to erase anomalies, the profit potential would not be within the relative probability, but in the absolute one. A player would only know that if he waited long enough, and kept betting, before the universe came to an end somewhere this activity would take place.

But I subscribe to the other theory, that the law of large numbers continues to be 50% heads, 50% tails, on average, and will simply overwhelm the anomaly. It will not be told it has to do anything. Note that neither circumstance is helpful to the gambler, yours, or this latter one. Umm, in any practical way , that is.

Quote:

The difference of the relative probability and the absolute probability.
link to original post

I think you have gotten a little tipsy absorbing this bit about relative/absolute.

Quote: rsactuary

Quote: DoubleGold

Quote: rsactuary

Quote: DoubleGold

I think there's a scientific law (as long as the initial imbalance is not artificial) that not only it can catch up, but it must catch up.


link to original post

Please provide evidence of that law. This contradicts the law of large numbers.
link to original post

It's common sense.

A coin flip's absolute probability is 50%.

It must hit that number or it's not random.

It would be a scam.

The relative probability can be varied.

Therein lies the profit potential.

The difference of the relative probability and the absolute probability.
link to original post

I'm sorry that you don't understand the law of large numbers. My work here is done.
link to original post

If a number was not random, as in a slot machine, then it's a scam.

For example, promising all slot players they will lose 14% but the math shows slot players lost 16% in a year (for a specific denomination).

They could lose their license.

Because the RNG is supposed to be random.

But it could be outright theft and the RNG works perfectly.

I too vow to quit participating for a while. I think this has to be stated or the vow gets violated , ha!

I'll come back if there is an elimination of the circular nature of the debate. It is *very* circular, something else that points to trolling .

It will be helpful if the OP does something towards proving his assertions ... or saying he understands he has indulged in some fallacies, and what they are

Quote: TigerWu

Quote: DRich

Quote: TigerWu

Quote: rsactuary

What the law of large numbers is: The next 1000 flips of that coin are expected to come 500 heads and 500 tails, meaning that after 1020 flips, you can expect to be at 513 heads which is about 50.3%. That is how regression to the mean works.
link to original post

So if I plan to flip a coin 1000 times, and after 500 flips I am at 400 heads and 100 tails, shouldn't I expect there to be a much larger number of tails in the next 500 flips?
link to original post

Nope, you would average 250 heads and 250 tails.
link to original post

So that's a total of 650 heads and 350 tails, right? Tails is still behind by 300. Since everything should be "regressing to the mean," at some point I should expect tails to "catch up" to heads, right? So over the long run, they will both be about 50%, right?
link to original post

It will but not necessarily under a specific number of spins.

If you did a million additional spins and it came up 500,000 Heads and 500,000 tails your totals would now be 500,400 Heads and 500,100 tails. that would be 49.99% Tails and 50.01% Heads. That is very close to 50/50 and there were no more Tails than Heads in the final 1,000,000 flips.

I can't define trolling, but I know it when I see it.

Quote: DoubleGold

Quote: rsactuary

Quote: DoubleGold

I think there's a scientific law (as long as the initial imbalance is not artificial) that not only it can catch up, but it must catch up.


link to original post

Please provide evidence of that law. This contradicts the law of large numbers.
link to original post

It's common sense.

A coin flip's absolute probability is 50%.

It must hit that number or it's not random.

It would be a scam.

The relative probability can be varied.

Therein lies the profit potential.

The difference of the relative probability and the absolute probability.
link to original post

"It must hit that number or it's not random."

This is the incorrect part.

Quote: Mental

I can't define trolling, but I know it when I see it.
link to original post

Yes. There can be NO DOUBT that the forum is being trolled in this thread by the OP. The only thing in doubt is when an administrator has the cajones to ban the OP.

Quote: DoubleGold

I think there's a scientific law (as long as the initial imbalance is not artificial) that not only it can catch up, but it must catch up.

Provided time is not limited and a few other params, like doing the actual flipping, having no bias, etc.

For slot machines, the casinos have to report regularly to the authorities, so they have to average out at least annually.
link to original post

Outcomes from coin flips, slot machines, roulette wheels, and dice are independent trials processes. Each event is independent of every other event. In the vernacular, coins, slot machines, roulette wheels, and dice have no memory.

However, playing cards dealt from a partially depleted deck are a dependent trials process. Each card dealt from a partially depleted deck is dependent upon the composition of the preceding cards removed from the deck. Cards do have a memory.

Quote: DoubleGold

Identifying the historical loser machines for users are also the historical money makers for the casinos, since they're the counter-party.

When I write historical loser machines, I am referring to those machines during the reporting cycle that likely have a wide spread from the average.

Those statistics are available internally, I'm pretty sure, but wouldn't want to do anything illegal nor unethical.

To identify one externally would be very difficult historically, so would likely have to focus on $25 a credit and higher machines.

So assign one person in each high limit room to report hand payouts of each high limit machine.

Employ about 25 and provide them a cell phone.

Track each machine by ID and only play the ones that have never hit big, whatever amount is decided upon.
link to original post

That sounds like an expensive complicated endeavor just to test a farfetched theory based on a condition well associated with gamblers fallacy.

With that being said, wouldn't it be better if the machines weren't truly random and there was actually something to identify? Take your theory on the road, and by road, I mean online.

There(online) you might actually have a chance to both test your theory and have a good chance at it actually working and making money.

There are a ton of online casinos and a ton of different slot providers, you don't have to worry about unreliable human-logged data, there are different tools/programs you could use to aid you. Someone can code specific programs to do most of the Work. You can test so much faster online with multiple computers and accounts at the same time, different slots at the same time, different providers, and much more. You have a good chance of finding other flaws that could be exploited. You could possibly even have +EV, while testing... if done correctly, using bonuses and promotions.

I know for a fact there have been online casinos with games that were not random and highly exploitable. I'm certain there are still some out there. Needles in haystacks? I don't know, perhaps not with enough data and testing.

I actually think a project like this would be worthwhile and a good idea for many reasons that I can think of. It's just a matter of having the willingness, dedication and knowledge.



I honestly couldn't tell you how you're gift/ handicap would help since a computer or AI could probably do a much better job at detecting that sort of thing.

Sorry for the slightly off-topic post, but this got me thinking about a situation online that I put on the back burner.

If there's anyone here reading who has the time, willingness, and ability to track an online slot and do some calculations.
I think it's oftentimes +EV Just from experimenting and eyeballing it, I don't think you'll get rich, but who the heck knows, I might be underestimating it.

If you're some brilliant math and slot expert making a crap ton(I know a few but, I haven't bothered asking them ), it's probably not worth your time.

The casino has a really good reputation.

I'll fund the project. PM me if interested.

FYI this is a totally different situation than what's been talked about previously. This situation deals with different level of "must-hit" mystery progressives with huge meter movement. It could probably even be automated if proven to be +EV.

DoubleGold,

Just don't like my post, go do some testing online.

As rsactuary has been eloquently explaining, the concept of "regressing to the mean" is based on the assumption that all the subsequent trials will be random. For coin flips, that translates into an assumption that heads and tails will occur randomly in the future with a statistical expectation of a 50/50 probability, and that such an outcome over the long-term will bring you closer to the mean of 50% in the long-term.

Because "regression to the mean" is based on the concept that all subsequent flips will be random, it is impossible to use the principle of "regression to the mean" to predict some skew away from randomness in the future.

"This is the incorrect part."




I refer to being random, because without being random, we cannot use the Probability Theorem nor Statistics.

Notice the following reference use of the word "random":



Probability and Statistics, the branches of mathematics concerned with the laws governing random events, including the collection, analysis, interpretation, and display of numerical data.

https://www.britannica.com/science/probability




I refer to the 50% absolute probability because of the usage of the Probability Theorem, Statistics, and Newton's Law.

The "action" would be a flip of coins where an imbalance of heads to tails occurs.

“and to every action there is always an equal and opposite or contrary, reaction”

https://www.goodreads.com/quotes/95351-and-to-every-action-there-is-always-an-equal-and



Reasoning: If there is an imbalance of heads versus tails after a random coin flip for a time series, then it must at least reach 50% absolute probability (using infinite time).

Not just be close nor approach.

Additionally, an imbalanced set (heads and tails are not equal) must reach an opposite relative probability.

Such that the deviation must at least move through 50% absolute probability line or 0% relative probability.

The only way to do that is to offset the imbalance.

It doesn't matter how many additional balanced flips occur.

It must make up for the imbalance to reach the absolute probability.

That implies moving through the 50% threshold.

If the the 50% was never reached, then the probability is not 50%.



If I am incorrect, please correct me and show me why.






To all:

I am still under newbie post constraints for 30 days, so thank-you for your patience.

I replied to this one because of the person's admin status and because I had a post available.

Quote: gordonm888

As rsactuary has been eloquently explaining, the concept of "regressing to the mean" is based on the assumption that all the subsequent trials will be random. For coin flips, that translates into an assumption that heads and tails will occur randomly in the future with a statistical expectation of a 50/50 probability, and that such an outcome over the long-term will bring you closer to the mean of 50% in the long-term.

Because "regression to the mean" is based on the concept that all subsequent flips will be random, it is impossible to use the principle of "regression to the mean" to predict some skew away from randomness in the future.
link to original post

Coin flip probability is 50%.

That implies balance of heads and tails.

If a series of flips gets out of whack by two, an offset the other direction by the same magnitude has to occur to get to 50% average.

But I am not using a time limit nor count limit.

If there was either, then that could prevent the balancing.

Anything less than balancing will not reach the 50% probability advertised.

This has major implications.

If a coin head count can get ahead by two, it most certainly can get behind by two.

I claim it must or the probability advertised is not random nor accurate.

Quote: DoubleGold

Quote: gordonm888

As rsactuary has been eloquently explaining, the concept of "regressing to the mean" is based on the assumption that all the subsequent trials will be random. For coin flips, that translates into an assumption that heads and tails will occur randomly in the future with a statistical expectation of a 50/50 probability, and that such an outcome over the long-term will bring you closer to the mean of 50% in the long-term.

Because "regression to the mean" is based on the concept that all subsequent flips will be random, it is impossible to use the principle of "regression to the mean" to predict some skew away from randomness in the future.
link to original post

Coin flip probability is 50%.

That implies balance of heads and tails.

If a series of flips gets out of whack by two, an offset the other direction by the same magnitude has to occur to get to 50% average.

But I am not using a time limit nor count limit.

If there was either, then that could prevent the balancing.

Anything less than balancing will not reach the 50% probability advertised.

This has major implications.

If a coin head count can get ahead by two, it most certainly can get behind by two.

I claim it must or the probability advertised is not random nor accurate.
link to original post

Other than possible trolling, what's the point/goal here? Is it just a conversation/debate about something that will never benefit you in any actual monetary way? Sure, there's more to life than making money, but I assumed the goal of gambling was to win money.

____________


the dreams of gamblers
dreaming that they can find a system against a -EV game that will provide them with steady winnings
something that the greatest math minds in the world cannot find and have stated over and over again is not possible

let them dream
if the greatest minds in the world can't convince them they're wrong - surely I can't


.

AxelWolf, thank-you for your patience.



I think I am trying to determine if apophenia is real and if I have it.

Because I certainly see patterns in supposedly random data.

For example, I've noticed more than anything, I will get a hot streak (slot machines), and the credit balance will eventually temporarily peak.

Then it will find a temporary floor.

The credit balance will bounce in a figurative box (rectangle), not exceeding the resistance (previous high balance) and not breaking through the support (previous low balance).

Then after several bounces it'll breakout of the box.



The example I previously supplied of a historical loser machine (someone called it a betting system) would not be something I'd be interested in even if it worked.

I was just showing how to apply a concept for more reader insight.



I think what I am learning in this thread will help me in a monetary way.

I'm always learning as I'm that type.

But I am more a lifetime student that likes to learn over the greedy type.

lilredrooster, I'd be interested in maximizing what's available or minimizing if down.

I'd like to be efficient as I can be.

Quote: DoubleGold

lilredrooster, I'd be interested in maximizing what's available or minimizing if down.


link to original post

so would I

and every other gambler - millions and millions of us

but we can't in a -EV game

we can reduce the size of our bets when we think we're likely to lose - but no guarantee that we won't start winning exactly then - meaning our choice actually hurt us

we can increase the size of our bets when we think we're going to win - but no guarantee that we won't start losing exactly then - meaning our choice actually hurt us

but keep looking, keep dreaming - until you're satisfied

until you realize that your time would be much better spent by looking for +EV opportunities



.