I opened 13 bonds (across 3 account) and got 1 Grade V.

If the drop rate was 50% (double the stated 25%) then the odds of that happening are 0.16%, or 625 to 1 against. 

There are loads of Bi-nominal (events with only two possible outcomes) calculators online so it's not difficult to find the probability for whatever experience you have had.

With a statistical confidence level of over 95% you can say the drop rate isn't as stated. Btw I am not saying it isn't, that is just the statistics / maths of it.

Who knows maybe I am just that really really special that 625th lucky person. In the last event one of the other captains got 1 from 8. So it's not like these extremely unlikely events aren't happening frequently. It's far beyond the simple explanation of bad luck.

@Mehorm the chances of getting 0 from 3 if the drop rate is 50% is 12.5% or 1 in 8. So every eight people that open 3 bonds would get nothing.

I opened 13 bonds (across 3 account) and got 1 Grade V.

If the drop rate was 50% (double the stated 25%) then the odds of that happening are 0.16%, or 625 to 1 against. 

There are loads of Bi-nominal (events with only two possible outcomes) calculators online so it's not difficult to find the probability for whatever experience you have had.

With a statistical confidence level of over 95% you can say the drop rate isn't as stated. Btw I am not saying it isn't, that is just the statistics / maths of it.

Who knows maybe I am just that really really special that 625th lucky person. In the last event one of the other captains got 1 from 8. So it's not like these extremely unlikely events aren't happening frequently. It's far beyond the simple explanation of bad luck.

@Mehorm the chances of getting 0 from 3 if the drop rate is 50% is 12.5% or 1 in 8. So every eight people that open 3 bonds would get nothing.

When you use the % symbol it means chances out of 100 Period.  Breaking in down into fractions that don't reflect the 1% or 1 out of a hundred or fifty out of one hundred  50%  and have that be your reasoning will always fall short of a person's expectations for the most part. 

And averages across a number of accounts are completely invalid ! 

your supposing a list of factors that that game doesn't imply are relevant to the equation that your putting forward!

With me  it's sometime I'm the windshield sometimes I'm the Bug  so I shrug and hope it's better next time. My last draw was 3 golds and 1 purple  the time before that I had 4 purples  and 3 purples the time before that

  it is what it is 

Wesley Pringle said:

When you use the % symbol it means chances out of 100 Period.

It doesn't. The sign actually represents a fraction. 

Maybe you never considered why the symbol looks the way it is but its one number being divided by another (the numbers being the two o's) with the slash being the divide by sign.

If we keep it simple. percentages between 0% to 100% represent fractions between between 0 and 1. 

For instance, if you want to calculate 90% of something you calculate it as the number times 0.9 (strictly speaking you are multiplying it by 0.9/1 but the /1 drops out of the calculation as it has no impact)

And there is actually no reason why percentages have to stop at 100%. 200% would be like multiplying by 2 (stickily speaking 2/1).

Wesley Pringle said:

And averages across a number of accounts are completely invalid ! 

This is also not correct.

The larger a sample the more statistically meaningful it would become. So if you were trying to calculate drop rates, collecting the experience of as many players as possible would be exactly the way to improve the accuracy of your result. 

You would do the maths as a weighted average, which would simply equate to totaling all the positive results and dividing by all the positive and negative results added together. It wouldn't matter how many ppl collected the data itself.

Hope this is helpful.

We will just need to disagree Because I find your numbers to be failing and misrepresenting to fit your narrative in my opinion

djmoody said:

Wesley Pringle said:

And averages across a number of accounts are completely invalid ! 

This is also not correct.

The larger a sample the more statistically meaningful it would become. So if you were trying to calculate drop rates, collecting the experience of as many players as possible would be exactly the way to improve the accuracy of your result. 

You would do the maths as a weighted average, which would simply equate to totaling all the positive results and dividing by all the positive and negative results added together. It wouldn't matter how many ppl collected the data itself.

Hope this is helpful.

The only way you could even hope to have a valid averages when dealing with  percentage across multi accounts would be to 100 consecutive Actions entered for each account you are averaging not 5 -7 actions per account! 

djmoody said:

Wesley Pringle said:

And averages across a number of accounts are completely invalid ! 

This is also not correct.

The larger a sample the more statistically meaningful it would become. So if you were trying to calculate drop rates, collecting the experience of as many players as possible would be exactly the way to improve the accuracy of your result. 

You would do the maths as a weighted average, which would simply equate to totaling all the positive results and dividing by all the positive and negative results added together. It wouldn't matter how many ppl collected the data itself.

Hope this is helpful.

The only way you could even hope to have a valid averages when dealing with  percentage across multi accounts would be to 100 consecutive Actions entered for each account you are averaging not 5 -7 actions per account! 

I hope this is helpful  for your future comparisons 

Wesley Pringle said:

The only way you could even hope to have a valid averages when dealing with  percentage across multi accounts would be to 100 consecutive Actions entered for each account you are averaging not 5 -7 actions per account! 

I hope this is helpful  for your future comparisons 

There would be absolutely no need to take consecutive samples at all. Nor would it matter if 100 people sampled 1 event each or 1 person sampled 100 times. The accuracy of your sample would be completely unchanged. The only driving factor to improving accuracy would be increasing the total sample size.

Like tossing a coin, the result of each roll is completely separate and independent of the rolls that have come before. There is no linkage between the rolls at all.

Our pattern seeking brains would tend to think that after 10 heads in a row it would be less likely that the next roll would be a head. Actually the chances for the next roll are still what they always are 50/50. Same for thinking, my lucks been down but it must soon change.... the next random event is not contingent in any way on what has gone in the past.

Taking consecutive samples would therefore have no impact. All the rolls are separate and independent. Each one has the same 50/50 chance of being a head. Whether a roll was 1st in a list or 50th or 100th, it would still be 50/50 and so each roll would be representative of the underlying drop chance. None of them would be superior to sample compared to the others.

We will need to agree to disagree then 

Opened 10 holy bonds(same account) previous event-that one before this one.. got 4 golds and 6 purples.I tried again my luck(cause IT IS a matter of LUCK) this time and i opened 4..got 3 golds and 1 purple-was very happy because those 3 i could enlight 3 new ones and make them lvl 40.I disagree with any1 who says "useless" champ... ANY gold champ matters-at least to me..as my target is to collect all 20 of them and to make them lvl 40 at least for Tower event to help my league to reach floor 100.One thing is sure.. the odds are better with X2..but as Limad said you can get 0 in 8 or a best 6 of 8-a friend got once 5 of 6..once)

djmoody said:

Wesley Pringle said:

And averages across a number of accounts are completely invalid ! 

This is also not correct.

The larger a sample the more statistically meaningful it would become. So if you were trying to calculate drop rates, collecting the experience of as many players as possible would be exactly the way to improve the accuracy of your result. 

You would do the maths as a weighted average, which would simply equate to totaling all the positive results and dividing by all the positive and negative results added together. It wouldn't matter how many ppl collected the data itself.

Hope this is helpful.

The only way you could even hope to have a valid averages when dealing with  percentage across multi accounts would be to 100 consecutive Actions entered for each account you are averaging not 5 -7 actions per account! 

I hope this is helpful  for your future comparisons 

A 50% chance is exactly the same as tossing a coin - Half the time you would expect to get 'heads' and half the time you would expect to get 'tails'
'Half' is an easily understood fraction

You do not need to toss the coin exactly 100 times, nor would it help to do so

If you do not get a result within 'reasonable' limits, then you can assume that the coin is weighted

I have the same problem with champion crit rates - after dozens and dozens of hits on different accounts and servers, I am utterly convinced the crit rate is a fake

http://prntscr.com/s3i5ds

Inigo Montoya said:

A 50% chance is exactly the same as tossing a coin - Half the time you would expect to get 'heads' and half the time you would expect to get 'tails'
'Half' is an easily understood fraction

You do not need to toss the coin exactly 100 times, nor would it help to do so

If you do not get a result within 'reasonable' limits, then you can assume that the coin is weighted

I have the same problem with champion crit rates - after dozens and dozens of hits on different accounts and servers, I am utterly convinced the crit rate is a fake

http://prntscr.com/s3i5ds

When dealing with  percent your dealing only with results Out of 100 (full stop)

If it was a 50/50 chance that would be how it would be listed ! 

Just because people want to change what facts are doesn't change the facts!

i really liked tossing a coin analogy..iv been using this in chat with friends asking me to explain stuff

i agree with both wesley and inigo both

but few things here

1. u cant posibily have enough draws in holy bonds each time to see if the coin is weighed heavier in one side

u toss a coin as many times as u can.. i mean i just did,right now.i had a coin in my pocket and after seeing i swear :D i tossed it like 20 times... funny thing.. first 5 times were all tails, so yes a person can be unlucky 5 times in a holy bond

2. now i tried even applyin other factors,like using more or less force while tossin it ,angling it .well in the end to me it seems like it all comes down to luck

3.a team mate of mine Loken RP ,during this last 1 in 3 event opened 3,all turned grade 5, well a league mate even got the event champion in first..

i agree with wesley more

i myself believe the chances are for each draw or opening .. u have a 50% chance to get a gold ,open and might or might not get one ,so for next draw it dsnt mean it increases the chance of getting another.. its still 50% each time it wil be just like this

also in case of opening 15 holy bonds and receiving no gold,its extremly rare but i believe it still could happen,u just have to have my bad luck :p 

cheers guys

i really liked tossing a coin analogy..iv been using this in chat with friends asking me to explain stuff

i agree with both wesley and inigo both

but few things here

1. u cant posibily have enough draws in holy bonds each time to see if the coin is weighed heavier in one side

u toss a coin as many times as u can.. i mean i just did,right now.i had a coin in my pocket and after seeing i swear :D i tossed it like 20 times... funny thing.. first 5 times were all tails, so yes a person can be unlucky 5 times in a holy bond

2. now i tried even applyin other factors,like using more or less force while tossin it ,angling it .well in the end to me it seems like it all comes down to luck

3.a team mate of mine Loken RP ,during this last 1 in 3 event opened 3,all turned grade 5, well a league mate even got the event champion in first..

i agree with wesley more

i myself believe the chances are for each draw or opening .. u have a 50% chance to get a gold ,open and might or might not get one ,so for next draw it dsnt mean it increases the chance of getting another.. its still 50% each time it wil be just like this

also in case of opening 15 holy bonds and receiving no gold,its extremly rare but i believe it still could happen,u just have to have my bad luck :p 

cheers guys

fURY said:

1. u cant posibily have enough draws in holy bonds each time to see if the coin is weighed heavier in one side

u toss a coin as many times as u can.. i mean i just did,right now.i had a coin in my pocket and after seeing i swear :D i tossed it like 20 times... funny thing.. first 5 times were all tails, so yes a person can be unlucky 5 times in a holy bond

I strongly suggest you look up bi-nominal calculators on the web as I previously suggested in the thread.

Rather than guessing how it works you can actually learn a lot about how the drops work, the likelihood of different results. You can exactly determine how lucky or unlucky you have been in a given situation and if you want determine the required sample size to confidently determine whether the drop rate during the double up event really is 50% or not you can do that also.

Here is one if you don't want to do your own google search.

Link to a Bi-Nominal Tool

Here are the actual probabilities of to the two situations you suggested.

Only getting 5 tails or less in 20 coin tosses. Probability of that happening is 14.8% or approx 1 in 7 times, or if you prefer around 15 people in 100 would experience that level of "bad luck".

The chance of someone getting no holy out of 5 drops during double chance event should be 3.1%, so around 1 in 32 people would experience that level of "bad luck".

And with regards to determining whether the drop rate is really 50% or not.

The above calculator gives you the z-test value. You can use that to determine the confidence level that you sample size is sufficient to draw a reasonable conclusion about the drop rate:

Simple table to convert your z-test value to a confidence level.

https://prnt.sc/s4ou98

For reference 95% confidence interval is used for most polling e.g. polling to predict general election results. You can pick your poison though, the higher the degree of confidence, the bigger the sample size required.

If enough people wanted to share their drop experience on the forum it would be possible to check whether the drop rate really was 50% or not.

YDoom said:

fURY said:

1. u cant posibily have enough draws in holy bonds each time to see if the coin is weighed heavier in one side

u toss a coin as many times as u can.. i mean i just did,right now.i had a coin in my pocket and after seeing i swear :D i tossed it like 20 times... funny thing.. first 5 times were all tails, so yes a person can be unlucky 5 times in a holy bond

I strongly suggest you look up bi-nominal calculators on the web as I previously suggested in the thread.

Rather than guessing how it works you can actually learn a lot about how the drops work, the likelihood of different results. You can exactly determine how lucky or unlucky you have been in a given situation and if you want determine the required sample size to confidently determine whether the drop rate during the double up event really is 50% or not you can do that also.

Here is one if you don't want to do your own google search.

Link to a Bi-Nominal Tool

Here are the actual probabilities of to the two situations you suggested.

Only getting 5 tails or less in 20 coin tosses. Probability of that happening is 14.8% or approx 1 in 7 times, or if you prefer around 15 people in 100 would experience that level of "bad luck".

The chance of someone getting no holy out of 5 drops during double chance event should be 3.1%, so around 1 in 32 people would experience that level of "bad luck".

And with regards to determining whether the drop rate is really 50% or not.

The above calculator gives you the z-test value. You can use that to determine the confidence level that you sample size is sufficient to draw a reasonable conclusion about the drop rate:

Simple table to convert your z-test value to a confidence level.

https://prnt.sc/s4ou98

For reference 95% confidence interval is used for most polling e.g. polling to predict general election results. You can pick your poison though, the higher the degree of confidence, the bigger the sample size required.

If enough people wanted to share their drop experience on the forum it would be possible to check whether the drop rate really was 50% or not.

And we all know how accurate election polls are :P  

Good info overall though, thanks.