Paid Spin Stats

Ohm

So, I have been spinning A LOT

Of late I decided tracking spins to see what I was getting back on them. I'll add more as I spin more. After spending more than $600 - I have decided to track what I'm winning/ losing on spins.

Observations:

1) 250/500/750/1000 credit drops are observed to be much rarer than the first 7-10 days (my past 50-75 spins also haven't yielded any of these)

2) Without the four drops above its always going to be difficult to break even

3) You lose a lot of value on trade-ins (33%)

4) I'm addicted

5) I'm spending more money on this than I would like to

Spin 1,Chain (2),Chain (2)

Spin 2,Mirror (2),Familiar (2)

Spin 3,Rope (2),Troggle (30)

Spin 4,Assembly (15),Phylactery (15)

Spin 5,Rope (2),Assembly (15)

Spin 6,Shackles (2),Rope (2)

Spin 7,Mirror (2),Phylactery (15)

Spin 8,75 Credits (75),Vanity_pet (2)

Spin 9,Familiar (2),Chain (2)

Spin 10,Chain (2),Artifact upgrade token (50)

Spin 11,Token (20),Map fragment (20)

Spin 12,Vanity_pet (2),Phylactery (15)

Spin 13,Chain (2) ,75 credits (75)

Spin 14,Assembly (15),Vanity_pet (2)

Spin 15,Phylactery (15),25 credits (25)

Spin 16,25 credits (25),Artifact upgrade token (50)

Spin 17,Rope (2),Vanity_pet (2)

Spin 18,Artifact upgrade token (50),Mirror (2)

Spin 19,Mirror (2),Phylactery (15)

Spin 20,troggle (30),Shackles (2)

Spin 21,Clover (2),Chain (2)

Spin 22,Token (20),Familiar (2)

Spin 23,Vanity_pet (2),25 credits (25)

Spin 24,75 credits (75),Clover (2)

Spin 25,Familiar (2),Shackles (2)

Spin 26,25 credits (25),Shackles (2)

Spin 27,Troggle (30),vanity_pet(2)

Spin 28,Assembly (15),Shackles (2)

Spin 29,Familiar (2),Familiar (2)

Spin 30,Mirror (2),Chain (2)

Column Totals,445,380

Spent,1050,

Gained,835,

Iniar

This is your problem:

4) I'm addicted

Ohm

See the thing is each spin costs you 35 credits - but returns 27.65 credits on average (n > 30 is a sufficient sample size?)

Salik

n > 30 is, presuming a bell curve, enough. The less regular the curve, the more data will be needed. Generally you can assume 30 is enough.

Ailish

Ohm said:

See the thing is each spin costs you 35 credits - but returns 27.65 credits on average (n > 30 is a sufficient sample size?)

I believe you just discovered the genesis of the expression "the house always wins".

Iniar

And in this case, the house is @Jeremy. No winning to be found here. 

Theophilus

I started with 150 credits. I didn't spend anything on the wheel.

I now have a sandstorm horn, death choker, lightning crown, lightning cloak, two Moradeim anklets, forest anklet, two treasure maps, a snow globe every familiar/vanity pet, monocle, artifact pet token, house token, token fishing stuff, propaganda pamphlet, two vials of perpetuity, three arti pipes and...

A resplendent halo currently costs 6050 credits.

It radiates with the powers of Aesir's constitution, diamond curative, diamond energy, invigorating Baar, magick's bane and steelweave veneer.

I'm not complaining.

I probably won 10,000 credits, but lost 4,000 spinning. I had 3,000 credits in tokens alone.

Ohm

I swear this game is conspiring against me

*hears @Jeremy's evil laughter in the background*

Lartus

Muahahhhaha. These bonus criticals are evil.

Skye

How concerned about @Ohm should we actually be right now? I'm not sure if this is all fun and games or if we should be genuinely worried and getting him some help...  :-?

Ohm

Seriously I need help  

I don't get it - how I have such terrible luck on the wheel. Is it a bug or a feature?

People spinning before me - RAWR - 1000 credits/ 750 credits

I spin - 900 credits converts 10 mounts/ 10 Tokens in 10 mins

Kabaal

If you spent all of that on Mega Millions tickets, you'd likely have lower returns.

IRE: Better than state-run lottos!

Dec

I haven't run the odds of winning a large credit amount from a spin, but the nature of gambling is that some people win and some people lose.  This is supposed to be a fun diversion with prizes that are typically cute and not always available.  If you are treating it as a means to gamble on credits and don't care about those prizes, remember that no system is a sure thing.

That being said, gambling addiction is a real thing.  Even when gambling fake money, if its causing problems with your well being it's time to talk to a behavioral health professional.  

Ohm

Dec said:

I haven't run the odds of winning a large credit amount from a spin, but the nature of gambling is that some people win and some people lose.  This is supposed to be a fun diversion with prizes that are typically cute and not always available.  If you are treating it as a means to gamble on credits and don't care about those prizes, remember that no system is a sure thing.

That being said, gambling addiction is a real thing.  Even when gambling fake money, if its causing problems with your well being it's time to talk to a behavioral health professional.  

I don't disagree with what you're saying. However, the expected value of spins over a very large number of spins for any 2 people should converge. If that's not the case, the system may be biased.

Gurn

Ohm said:

Dec said:

I haven't run the odds of winning a large credit amount from a spin, but the nature of gambling is that some people win and some people lose.  This is supposed to be a fun diversion with prizes that are typically cute and not always available.  If you are treating it as a means to gamble on credits and don't care about those prizes, remember that no system is a sure thing.

That being said, gambling addiction is a real thing.  Even when gambling fake money, if its causing problems with your well being it's time to talk to a behavioral health professional.  

I don't disagree with what you're saying. However, the expected value of spins over a very large number of spins for any 2 people should converge. If that's not the case, the system may be biased.

This is correct IN THEORY. However, you don't know what astronomically large sample size you actually need in order to calculate expected value. Remember that expected value is an average of ALL spins. There is no reason to assume the system is biased or cheating you in any way, because you simply have not enough information. It is also absolutely completely reasonable for the system as a whole to be having an EV of 0, making it completely fair, but completely screwing you with all the negative values while everyone else takes the positive values.

That's just how the cruel mistress of statistics and gambling works.

Ohm

Gurn said:

Ohm said:

Dec said:

I haven't run the odds of winning a large credit amount from a spin, but the nature of gambling is that some people win and some people lose.  This is supposed to be a fun diversion with prizes that are typically cute and not always available.  If you are treating it as a means to gamble on credits and don't care about those prizes, remember that no system is a sure thing.

That being said, gambling addiction is a real thing.  Even when gambling fake money, if its causing problems with your well being it's time to talk to a behavioral health professional.  

I don't disagree with what you're saying. However, the expected value of spins over a very large number of spins for any 2 people should converge. If that's not the case, the system may be biased.

This is correct IN THEORY. However, you don't know what astronomically large sample size you actually need in order to calculate expected value. Remember that expected value is an average of ALL spins. There is no reason to assume the system is biased or cheating you in any way, because you simply have not enough information. It is also absolutely completely reasonable for the system as a whole to be having an EV of 0, making it completely fair, but completely screwing you with all the negative values while everyone else takes the positive values.

That's just how the cruel mistress of statistics and gambling works.

Oh, I'm not saying it is cheating or targeting me in any way.

You don't need astronomically large sample sizes. Because each incremental spin after a certain number of spins barely moves the expected value at all. All I'm saying is there should be convergence between EVs of two players. And if there is a marked difference, then something is biased systematically (not intentionally). 

Gurn

Ohm said:

Gurn said:

Ohm said:

Dec said:

I haven't run the odds of winning a large credit amount from a spin, but the nature of gambling is that some people win and some people lose.  This is supposed to be a fun diversion with prizes that are typically cute and not always available.  If you are treating it as a means to gamble on credits and don't care about those prizes, remember that no system is a sure thing.

That being said, gambling addiction is a real thing.  Even when gambling fake money, if its causing problems with your well being it's time to talk to a behavioral health professional.  

I don't disagree with what you're saying. However, the expected value of spins over a very large number of spins for any 2 people should converge. If that's not the case, the system may be biased.

This is correct IN THEORY. However, you don't know what astronomically large sample size you actually need in order to calculate expected value. Remember that expected value is an average of ALL spins. There is no reason to assume the system is biased or cheating you in any way, because you simply have not enough information. It is also absolutely completely reasonable for the system as a whole to be having an EV of 0, making it completely fair, but completely screwing you with all the negative values while everyone else takes the positive values.

That's just how the cruel mistress of statistics and gambling works.

Because each incremental spin after a certain number of spins barely moves the expected value at all. <---- This statement only holds if you know how many samples you need to have a decently accurate measurement. Because you don't know what you need, you need astronomically high numbers to prove anything, BECAUSE the variance is so high. If I spin 1000 times and I get 2 credit things each time, then get a 2000 credit thing, then my 1001 spins can't really be considered representative of the mean value.

You're still correct in theory, but in all practicality, it's impossible.

Khizan

Ohm said:

You don't need astronomically large sample sizes. Because each incremental spin after a certain number of spins barely moves the expected value at all. All I'm saying is there should be convergence between EVs of two players. And if there is a marked difference, then something is biased systematically (not intentionally). 

You need a fairly large number, though. Certainly larger than you think.

Let's say you spin 30 times, since you used that as an example. You get a 1500 credit spin mixed in with 29 rope/rope spins. This gives you an estimated value of roughly 50cr/spin. I'm going to assume ropes have no tradein value here just to make things easier on me, because their 1cr/per wouldn't really change things substantially.

That's a 97% chance of ropes, roughly. Spin 30 times and you have roughly a 40% chance of nothing but ropes. Spin 42 times and you have a 27% chance of getting nothing but ropes. Once you pass 42 spins you will have spent more than 1500cr on spins, and your cr/spins ratio will be negative without multiple wins.

Now increase the chances. You get two 1500 cr prizes in your 30 spins. That's roughly a 94% chance of ropes. Now 30 spins has a 15% chance of nothing but ropes and 42 spins has an all-rope chance of roughly 7%. Just adding 1 more win to your spins has altered your probability assumptions dramatically. Tell somebody they have a roughly  1 in 4 chance of being screwed and fewer people will take the chance than if you tell them they have a roughly 1 in 10 chance of being screwed.

And what if you were actually lucky and the odds are actually 1 in 40? Now you have a 54% chance of all ropes over 30 spins. If you're using the 1 in 30 data set, your odds here are about twice as bad as you think they are. 

Small data sets = bad for probability

Iniar

The hardest thing to accept is that the system owes you nothing. It will do what it wants to do, that is the nature of gambling. The sample size to demonstrate convergence (from the individual point of view) may well be n = 5, or n = 5000, and the system will never be obliged to explain to you why IT HASN'T LANDED ON BLACK (for you (this time))!!

Ohm

Gurn said:

Ohm said:

Gurn said:

Ohm said:

Dec said:

I haven't run the odds of winning a large credit amount from a spin, but the nature of gambling is that some people win and some people lose.  This is supposed to be a fun diversion with prizes that are typically cute and not always available.  If you are treating it as a means to gamble on credits and don't care about those prizes, remember that no system is a sure thing.

That being said, gambling addiction is a real thing.  Even when gambling fake money, if its causing problems with your well being it's time to talk to a behavioral health professional.  

I don't disagree with what you're saying. However, the expected value of spins over a very large number of spins for any 2 people should converge. If that's not the case, the system may be biased.

This is correct IN THEORY. However, you don't know what astronomically large sample size you actually need in order to calculate expected value. Remember that expected value is an average of ALL spins. There is no reason to assume the system is biased or cheating you in any way, because you simply have not enough information. It is also absolutely completely reasonable for the system as a whole to be having an EV of 0, making it completely fair, but completely screwing you with all the negative values while everyone else takes the positive values.

That's just how the cruel mistress of statistics and gambling works.

Because each incremental spin after a certain number of spins barely moves the expected value at all. <---- This statement only holds if you know how many samples you need to have a decently accurate measurement. Because you don't know what you need, you need astronomically high numbers to prove anything, BECAUSE the variance is so high. If I spin 1000 times and I get 2 credit things each time, then get a 2000 credit thing, then my 1001 spins can't really be considered representative of the mean value.

You're still correct in theory, but in all practicality, it's impossible.

Iniar said:

The hardest thing to accept is that the system owes you nothing. It will do what it wants to do, that is the nature of gambling. The sample size to demonstrate convergence (from the individual point of view) may well be n = 5, or n = 5000, and the system will never be obliged to explain to 

you why IT HASN'T LANDED ON BLACK (for you (this time))!!

@Khizan

30 is small sample size for sure, somewhere between 500 and 1000 is probably not.

@Iniar

Oh it doesn't owe me anything. I understand that. But - at the same time, it shouldn't be a system like Excel's earliest RAND function was - where the random outcomes were in a predetermined order every time - since the seed was the same. If that's the case, it will invariably result in biased outcomes.

Ohm

Honestly, this thread was not created to suggest the system owes me anything. It was just a way to compare and contrast spins with other power spinners. But it's taken a turn in an unintended direction. 

Iniar

I spun four times today and won 250 credits on one.

Cardi

Ohm said:

I work as an investment risk manager

From another thread, but I would hate to have you as my investment risk manager,

Me: Does it sound like a good idea to try my luck on 100 spins

Ohm: Better make that 200 spins sonny

Ailish

umm, an investment risk manager who seemingly doesn't understand how the speed of convergence to the mean would be affected by the variance of the distribution? Yikes.

Anette

I think the problem many people can have is that a lot of the "winnings" are probably useless to them.  Getting a few lessons?  If you're omnitrans?  Though they did seem to finangle those out as I haven't gotten them in some time.  (Or maybe it's everyone else getting them now, it's always hard to tell)

Vanity pets and familiars are nice the first time you get them, but when you're on your 10th, not so much.

Personally, I wouldn't remove them, but I would probably add a thing that would skew a "vanity pet" result towards giving you one you don't have, at the very least.  If it does this, I certainly couldn't tell.

As for the ethereal stuff, well, there's always the "consolation prize" in such gambling systems, lockboxes or crates or in this case spins, but I've been tempted to trigger getting one on a sound clip of that Charlie Brown Halloween special line "I got a rock :c"