I love this thread. There are two sides attacking the OP, but they are also unwittingly attacking each other.
One side says that it's normal to kill the warden 30 times at MP5 and have only 1 key drop; the other side says that the OP is lying, because they think that killing the warden 30 times at MP5 and only getting 1 key is almost impossible.

Maybe side #1 should start attacking side #2.

Given that only one of these arguments has empirical validity. It doesn't really matter.

I love this thread. There are two sides attacking the OP, but they are also unwittingly attacking each other.
One side says that it's normal to kill the warden 30 times at MP5 and have only 1 key drop; the other side says that the OP is lying, because they think that killing the warden 30 times at MP5 and only getting 1 key is almost impossible.

Maybe side #1 should start attacking side #2.

That's not true, both sides are correct! #1 says, it's normal, though very rare, #2 says it's very rare therefore OP is lying. We have the same conclusion, #2 just takes it further and applies it to human behaviour. The only side that can not win either way is the OP...

Probability mathematics and randomness is very interesting topic. This topic, however, is silly.

The probablity of an event with a 0.5 chance of happening (not getting a key) to happen only once out of 30 times is (0.5^{30}) x 30. That is 2,794 x 10^{-8}, which in turn is 1 in 35,791,394. Just so you know.

The probablity of an event with a 0.5 chance of happening (not getting a key) to happen only once out of 30 times is (0.5^{30}) x 30. That is 2,794 x 10^{-8}, which in turn is 1 in 35,791,394. Just so you know.

Counted this myself too and then thought that which is more probable: Op is lying and/or just trolling OR he is just that unlucky.

I have read these whinies in diablofans and at official forums and none of them has actually understood the probabilities/statistics and call things broken. Yet I havent seen any plausible data/videos of those insane "30 runs with 1 key".

As an avid enthusiast of mathematics, I'd like to point out you can't throw the book of stats at the OP because Diablo 3 is NOT using an accurate model for producing the kind of randomness you're talking about here. Many games that utilize statistical variance ALSO employ a checks and balance system that variably changes the likelihood of a given event occurring to increase/decrease its resulting outcome based upon previous results.

If you flipped a coin and it lands heads, depending on how many times you flipped the coin before that and what the results were, the game would increase/decrease the chance of it landing tails on the next flip to bring the average result in line with the specified chance. Another alternative is the use of number tables. In either case, true randomness cannot occur as the system is being rigidly controlled.

The idea that "random is random" is a fallacy when it comes to video games. The OP might be wearing a tinfoil hat, but you can't defend Diablo 3 with traditional mathematics. This is programming, a very different beast.

As an avid enthusiast of mathematics, I'd like to point out you can't throw the book of stats at the OP because Diablo 3 is NOT using an accurate model for producing the kind of randomness you're talking about here. Many games that utilize statistical variance ALSO employ a checks and balance system that variably changes the likelihood of a given event occurring to increase/decrease its resulting outcome based upon previous results.

There is absolutely no evidence that Blizzard has employed such an algorithm. It would mean that after a bad streak the system would definitely give you a good streak, or after a long run of no legendary items you will find a godly Echoing Fury. In turn, this would mean that after you find a 2 billion item, you can stop playing because the chance for a good item will now be significantly lower.

I highly doubt that Blizzard uses such a system, based on drops for real lucky players I've seen I would even say it's definitely not the case, but my sample is obviously not big enough. If you don't have evidence for your claim (i.e., you have seen the code itself) I would be careful, because it would not make sense for Blizzard to implement a drop pattern that takes previous events into account.

About RNG/PRNG... this discussion is happening right now in the MF thread as well. Feel free to join if you think my logic is flawed ;-)

You should read my post more carefully. I am not saying that Blizzard is using such a system. I pointed out two very real systems that are being used today in many games and provided a link to further explain what I briefly outlined.

The fact is, and it is a fact, video games CANNOT create true randomness at this time. This means that arguing "random is random" is pointless as it is precisely the opposite. Programming places very rigid controls on statistical outcomes. They may "appear" random, but they are, in fact, not.

My worst series in act 2 was 14 runs without a key on mp5. Had 3-in-a-row keyless act 2 runs on mp8 too, and many other unlucky runs. Act 2 key farming is really frustrating to me. While I don't think Blizzard screwed it up for real, I do think sometimes that I should be logging my runs, just to have hard data. It helps staying sane

The fact is, and it is a fact, video games CANNOT create true randomness at this time. This means that arguing "random is random" is pointless as it is precisely the opposite. Programming places very rigid controls on statistical outcomes. They may "appear" random, but they are, in fact, not.

There are many hardware RNG-s available for PCs/servers, and since all RNG happens server side in Diablo, you can't know. For example Intel's Ivy Bridge processors provide a true RNG as part of the AVX instruction set. No special hardware needed there.

The probablity of an event with a 0.5 chance of happening (not getting a key) to happen only once out of 30 times is (0.5^{30}) x 30. That is 2,794 x 10^{-8}, which in turn is 1 in 35,791,394. Just so you know.

When calculating such odds, I believe it's customary to include the probability of the event of not finding a key out of the 30 runs as well. This is because the converse of only getting 1 key out of 30 should be the probability of getting 2 or more keys, not 2 or more and 0 keys.
So the odds should be 31*0.5^{30} which changes everything!

Statistically, if you want to check if the droprate is 50% at mp5 you need atleast 10 drops and 10 not drops in order to allow CLT deal with the non normal distribution of the samples.
But lets run the data for lulz...
Minitab:

Test and CI for One Proportion

Test of p = 0,5 vs p not = 0,5

Sample X N P-Value
1 1 30 0,000

That doesn't tell most peeps anything,,, So lets go through the data. X represents getting a key, N represents the number of runs, the P-Value is the chance of getting one or zero keys in 30 tries assuming the hypothesis that the probability to get a key is 50%. Generally if the P-Value is below 0.05 you discard the hypothesis.

But as we don't don't have enough data, we can't really use the test that's been used there... So we can't really attack the problem with statistics based on your small sample. Give me one where you have 10 drops and 10 non-drops and we can start talking statistics. Untill then, keep it out of it!

As an avid enthusiast of mathematics, I'd like to point out you can't throw the book of stats at the OP because Diablo 3 is NOT using an accurate model for producing the kind of randomness you're talking about here. Many games that utilize statistical variance ALSO employ a checks and balance system that variably changes the likelihood of a given event occurring to increase/decrease its resulting outcome based upon previous results.

If you flipped a coin and it lands heads, depending on how many times you flipped the coin before that and what the results were, the game would increase/decrease the chance of it landing tails on the next flip to bring the average result in line with the specified chance. Another alternative is the use of number tables. In either case, true randomness cannot occur as the system is being rigidly controlled.

The idea that "random is random" is a fallacy when it comes to video games. The OP might be wearing a tinfoil hat, but you can't defend Diablo 3 with traditional mathematics. This is programming, a very different beast.

Yes, of course. My calculations were based on the assumption that the events were independent, which wouldn't be the case in any of the "controlling randomness" systems with decribed. I do believe Blizzard would have told us if they were using any, and in any case the chance of the OP only getting 1 key out of 30 would be even less if any of these systems was being used.

Ok, MP5 is similar to flipping coins, 50% chance for head, 50% chance for tails. The chance to get 29 head and 1 tails is really really small, about one in 35 million.

Now, this is very unlucky. However one has to keep in mind that a lot of people play D3, and many more "coin flips" are being done throughout the game. This increases the chance that it happens to someone. The chance is comparable to winning the lottery, which, while unlikely for an individual, overall happens quite regularly.

edit: Just saw page two, math has been done before my post.

Given that only one of these arguments has empirical validity. It doesn't really matter.

That's not true, both sides are correct! #1 says, it's normal, though very rare, #2 says it's very rare therefore OP is lying. We have the same conclusion, #2 just takes it further and applies it to human behaviour. The only side that can not win either way is the OP...

Probability mathematics and randomness is very interesting topic. This topic, however, is silly.

^{30}) x 30. That is 2,794 x 10^{-8}, which in turn is 1 in 35,791,394. Just so you know.Counted this myself too and then thought that which is more probable: Op is lying and/or just trolling OR he is just that unlucky.

I have read these whinies in diablofans and at official forums and none of them has actually understood the probabilities/statistics and call things broken. Yet I havent seen any plausible data/videos of those insane "30 runs with 1 key".

If you flipped a coin and it lands heads, depending on how many times you flipped the coin before that and what the results were, the game would increase/decrease the chance of it landing tails on the next flip to bring the average result in line with the specified chance. Another alternative is the use of number tables. In either case, true randomness cannot occur as the system is being rigidly controlled.

http://en.wikipedia....mber_generation

The idea that "random is random" is a fallacy when it comes to video games. The OP might be wearing a tinfoil hat, but you can't defend Diablo 3 with traditional mathematics. This is programming, a very different beast.

There is absolutely no evidence that Blizzard has employed such an algorithm. It would mean that after a bad streak the system would definitely give you a good streak, or after a long run of no legendary items you will find a godly Echoing Fury. In turn, this would mean that after you find a 2 billion item, you can stop playing because the chance for a good item will now be significantly lower.

I highly doubt that Blizzard uses such a system, based on drops for real lucky players I've seen I would even say it's definitely not the case, but my sample is obviously not big enough. If you don't have evidence for your claim (i.e., you have seen the code itself) I would be careful, because it would not make sense for Blizzard to implement a drop pattern that takes previous events into account.

About RNG/PRNG... this discussion is happening right now in the MF thread as well. Feel free to join if you think my logic is flawed ;-)

The fact is, and it is a fact, video games CANNOT create true randomness at this time. This means that arguing "random is random" is pointless as it is precisely the opposite. Programming places very rigid controls on statistical outcomes. They may "appear" random, but they are, in fact, not.

There are many hardware RNG-s available for PCs/servers, and since all RNG happens server side in Diablo, you can't know. For example Intel's Ivy Bridge processors provide a true RNG as part of the AVX instruction set. No special hardware needed there.

http://en.wikipedia.org/wiki/Hardware_random_number_generator

Taking your OP into consideration, I would bet my money on it being the latter.

When calculating such odds, I believe it's customary to include the probability of the event of not finding a key out of the 30 runs as well. This is because the converse of only getting 1 key out of 30 should be the probability of getting 2 or more keys, not 2 or more and 0 keys.

So the odds should be 31*0.5

^{30}which changes everything!But lets run the data for lulz...

Minitab:

Test and CI for One Proportion

Test of p = 0,5 vs p not = 0,5

Sample X N P-Value

1 1 30 0,000

That doesn't tell most peeps anything,,, So lets go through the data. X represents getting a key, N represents the number of runs, the P-Value is the chance of getting one or zero keys in 30 tries assuming the hypothesis that the probability to get a key is 50%. Generally if the P-Value is below 0.05 you discard the hypothesis.

But as we don't don't have enough data, we can't really use the test that's been used there... So we can't really attack the problem with statistics based on your small sample. Give me one where you have 10 drops and 10 non-drops and we can start talking statistics. Untill then, keep it out of it!

Yes, of course. My calculations were based on the assumption that the events were independent, which wouldn't be the case in any of the "controlling randomness" systems with decribed. I do believe Blizzard would have told us if they were using any, and in any case the chance of the OP only getting 1 key out of 30 would be

even lessif any of these systems was being used.Now, this is very unlucky. However one has to keep in mind that a lot of people play D3, and many more "coin flips" are being done throughout the game. This increases the chance that it happens to someone. The chance is comparable to winning the lottery, which, while unlikely for an individual, overall happens quite regularly.

edit: Just saw page two, math has been done before my post.

And a Liar is a Liar.

Currently played toon:https://eu.battle.net/d3/en/profile/Rage-2973/hero/97362116