Lol I like how these team instinct members have become. Including [iN] in their own thread titles now. Luckily there is one in my arena which I drop on a regular basis.
The sample size needed depends on the probability itself. For something with a probability in that range, 100 samples is actually enough to establish the rate to a pretty high level of certainty.
Thanks for the data, real data is rare around here and always appreciated.
It's a relevant sample size because 99.9% of the player base isn't going to refresh their energy 9 times in a row. So even this extreme example demonstrates how bogus drop rates are
The sample size needed depends on the probability itself. For something with a probability in that range, 100 samples is actually enough to establish the rate to a pretty high level of certainty.
Thanks for the data, real data is rare around here and always appreciated.
It's a relevant sample size because 99.9% of the player base isn't going to refresh their energy 9 times in a row. So even this extreme example demonstrates how bogus drop rates are
How any times you actually run it doesn't mean squat related to a relevant sample size. By this logic I should always get 1/3 every day I aim 3 hard nodes. Just because we don't or can't run an accurate sample size doesn't make the true sample size any less relevant.
As the definition of probability goes, you would have to run it an infinite amount of times to actually prove the drop rate is what you claim it is. Being that is impossible to do, a much higher number of samples is usually sufficient. 116 is in fact, a very small sample set. It would need to be on the order of hundreds of thousands of samples. If you ran the above experiment for 100 days and came up with the same results, I would be inclined to agree that your claim is correct. Other than that, I'd have to say you had extremely terrible luck today.
I don't think you know what probability means. There are statistical tests for this. His sample size is more than enough for an adequate estimate on the drop rate. Only very low % values require such big groups. He's not "proving" anything, he's giving an estimate.
Which is good, considering all the crybabies saying they ran it 50 times and got 1-2 pieces.
As the definition of probability goes, you would have to run it an infinite amount of times to actually prove the drop rate is what you claim it is. Being that is impossible to do, a much higher number of samples is usually sufficient. 116 is in fact, a very small sample set. It would need to be on the order of hundreds of thousands of samples. If you ran the above experiment for 100 days and came up with the same results, I would be inclined to agree that your claim is correct. Other than that, I'd have to say you had extremely terrible luck today.
83% of us would get the EXACT same results. The rest would get a few drops +/-.
Using your linked calculator assumes you know the drop rate. We have a polling problem here. There is an infinite population. How many samples do you need to take to have 99% confidence that your sampled value is +/- 2.5%
100% correct. In fact, if you use the calculator, and assume that 0.33 is the actual probability and use 20 as the number of successes, you can see that there's a .999999 probability that the average person will get 20 . So if one knows how to use that calculator, this just further proves that the actual drop rate is correct at 0.33 probability of you were to only obtain 20 on any given 100 sample run.
The sample size needed depends on the probability itself. For something with a probability in that range, 100 samples is actually enough to establish the rate to a pretty high level of certainty.
Thanks for the data, real data is rare around here and always appreciated.
It's a relevant sample size because 99.9% of the player base isn't going to refresh their energy 9 times in a row. So even this extreme example demonstrates how bogus drop rates are
How any times you actually run it doesn't mean squat related to a relevant sample size. By this logic I should always get 1/3 every day I aim 3 hard nodes. Just because we don't or can't run an accurate sample size doesn't make the true sample size any less relevant.
What relevance does using statistics have in regards to this game, though? It's not TRULY random. It's all designed to give you a certain payout over a certain period of time. You will NEVER go past that limit. They will never give you true random chance. Gear is the only way they can keep whales from outgrowing content faster than the devs can produce it. You will ALWAYS be behind on gear
The sample size needed depends on the probability itself. For something with a probability in that range, 100 samples is actually enough to establish the rate to a pretty high level of certainty.
Thanks for the data, real data is rare around here and always appreciated.
It's a relevant sample size because 99.9% of the player base isn't going to refresh their energy 9 times in a row. So even this extreme example demonstrates how bogus drop rates are
How any times you actually run it doesn't mean squat related to a relevant sample size. By this logic I should always get 1/3 every day I aim 3 hard nodes. Just because we don't or can't run an accurate sample size doesn't make the true sample size any less relevant.
What relevance does using statistics have in regards to this game, though? It's not TRULY random. It's all designed to give you a certain payout over a certain period of time. You will NEVER go past that limit. They will never give you true random chance. Gear is the only way they can keep whales from outgrowing content faster than the devs can produce it. You will ALWAYS be behind on gear
It's completely relevant to use statistics because those complaining that the drop rates aren't what they say they are can be proved using stochastic calculations. And when it is proved, or demonstrated with high confidence, the argument then changes to "well it's not truly a probability thing, they programmed it to not be that way", which is insane because the only way they can program it so that some people have seen a high drop rate and some have witnessed a low drop rate is by using random numbers.
That being said I don't truly believe that all shard nodes or gear nodes are the exact same probability, by design. Lower gear is higher probability in all likelihood. But you can't say that statistics have no relevance here. It's the only sane argument to a topic fueled by observation bias and emotion.
As the definition of probability goes, you would have to run it an infinite amount of times to actually prove the drop rate is what you claim it is. Being that is impossible to do, a much higher number of samples is usually sufficient. 116 is in fact, a very small sample set. It would need to be on the order of hundreds of thousands of samples. If you ran the above experiment for 100 days and came up with the same results, I would be inclined to agree that your claim is correct. Other than that, I'd have to say you had extremely terrible luck today.
I don't think you know what probability means. There are statistical tests for this. His sample size is more than enough for an adequate estimate on the drop rate. Only very low % values require such big groups. He's not "proving" anything, he's giving an estimate.
Which is good, considering all the crybabies saying they ran it 50 times and got 1-2 pieces.
Apparently you didn't get his point. He's arguing the terminology used. Also, I wouldn't say that 100 is a good sample size for even an "adequate" estimate. I'm at exactly 50 handcuff pieces in 92 attempts. Can I now go ahead and create a thread stating "Proof of handcuffs having 50% drop rate"? No, of course no. That was the drop rate of my alt (which apparently always is luckier than my main) and based on history I know that such luck will probably not happen to me again within the next 6 months or perhaps ever again.
A rough estimate could be made at 1000 attempts. A good claim, like Valent_Antona stated would require 100*100 attempts. The video shows nothing more but what drop rate he had based on the attempts. It gives you an idea not to expect a 100% drop rate, and not to expect a drop rate lower than 5%.
I don't think you know what probability means. There are statistical tests for this. His sample size is more than enough for an adequate estimate on the drop rate. Only very low % values require such big groups. He's not "proving" anything, he's giving an estimate.
Which is good, considering all the crybabies saying they ran it 50 times and got 1-2 pieces.
No, 116 is an insufficient sample size to determine the probability of this purple drop. See quoted section below. The drop rate identified by that test is 21 +/-9 19 times out of 20.
Also when you say only very low values require large groups is sort of misleading and the value of error doesn't change with the probibility. The signicance of the error does so you want larger sample sizes
But you can't say that statistics have no relevance here. It's the only sane argument to a topic fueled by observation bias and emotion.
I don't believe statistics is relevant here because I don't believe gear drops are true random chance. I believe they're controlled by an algorithm that won't dispense over a certain amount of mats during a certain period of time, or something similar to that
Streak breakers to end cold or limit hot streaks make sense but even with those the overall drop rates would represent a binomial distribution outside of the outliers.
So statistical analysis would still be relevant even with noise generated at the boundaries.
I'm dont understand the drivers for a non random system.
Streak breakers to end cold or limit hot streaks make sense but even with those the overall drop rates would represent a binomial distribution outside of the outliers.
So statistical analysis would still be relevant even with noise generated at the boundaries.
I'm dont understand the drivers for a non random system.
"to those that do not understand electricity, a light bulb seems like magic".
Streak breakers to end cold or limit hot streaks make sense but even with those the overall drop rates would represent a binomial distribution outside of the outliers.
So statistical analysis would still be relevant even with noise generated at the boundaries.
I'm dont understand the drivers for a non random system.
A mathematician(statistician): If I halve the distance between each step, it will take an infinite number of steps to reach that wall.
An engineer(person designing an actual house to be built): You will very quickly be close enough to that wall so the remaining distance doesn't matter.
Mathematician: Technically, I'm still right.
Engineer: And realistically, it doesn't **** matter.
The amount of **** measuring on this forum is amazing. Thanks for the data. It measures up pretty close to what I have (79/401 according to excel)
I'm dont understand the drivers for a non random system.
I already said why. To keep whales from consuming content faster than they can produce it. Drop $5,000 and you pretty much have all the characters at 7* or close to it. That's a random system. What's next? Gear them. And that's where EA sets the cap. Players will NEVER push past this cap. It's set at an artificial level to keep them chasing. Then slowly they introduce things like the latest Aurodium pack that contain some purple items and the whales gobble these up like starving refugees.
The problem is you're thinking like a mathematician. This game is run by businessmen. They want money. They're absolute geniuses when it comes to manipulating people for it. And going by responses in this topic they've been doing an amazing job
Here's the true percentages. 99.9% of players hate the rare purple drop rates, .01% aka developers love it.
Everyone agrees EA is **** greedy with their drop rates, what the **** is new? They're not changing **** because they're overpaid as ****.
If that 1% top earner of this game gave away his money and even gave a ****, this game would be so much better, **** drop rates
A mathematician(statistician): If I halve the distance between each step, it will take an infinite number of steps to reach that wall.
An engineer(person designing an actual house to be built): You will very quickly be close enough to that wall so the remaining distance doesn't matter.
Mathematician: Technically, I'm still right.
Engineer: And realistically, it doesn't **** matter.
The amount of **** measuring on this forum is amazing. Thanks for the data. It measures up pretty close to what I have (79/401 according to excel)
And software is about 60% math, 40% engineering. Mess up either side and you have a broken system.
Your history is getting closer to a good sample set, which is promising. The only thing it proves is that at the specified drop rate probability, the actual standard dev is larger than expected. It doesn't mean the actual probability is wrong.
Over the course of a very long time, we can expect what they indicate the rate is. On average some may find more or less.
I'm dont understand the drivers for a non random system.
I already said why. To keep whales from consuming content faster than they can produce it. Drop $5,000 and you pretty much have all the characters at 7* or close to it. That's a random system. What's next? Gear them. And that's where EA sets the cap. Players will NEVER push past this cap. It's set at an artificial level to keep them chasing. Then slowly they introduce things like the latest Aurodium pack that contain some purple items and the whales gobble these up like starving refugees.
The problem is you're thinking like a mathematician. This game is run by businessmen. They want money. They're absolute geniuses when it comes to manipulating people for it. And going by responses in this topic they've been doing an amazing job
Your missing the psychology. This entire game is a slot machine with no jackpot. Small incremental wins keep people feeding quarters into the machine. There would also be outcry from the megawhales if their drop rate disappeared. Also if caught after saying the drop rate doesn't change they would kill their credibility. The escalating cost of energy is what prevents whales from burning through content. You fix the amount of energy available and don't allow banking and you restrict pace. And if they did take time to code a complex drop system they would have also put in streak breaking logic at the bottom end.
The fact that their is consensus forming at somewhere between 15-25% suggests that it is random.
Replies
Great job on the sound!
Heres a simple online calculator for confidence. Saves doing the math. Just put in a large number for population.
So in the above example if you use a 95% confidence level and 116 samples your confidence interval is 9.
So the above test correctly stated is the drop rate for stun cuffs is 21% +/- 9 19 times out of 20.
To get better results doesn't actually take that many more samples.
To get 99% confidence +/- 2.5 we would need 2600 samples. At a 95% confidence just 1500 samples.
So 6%?? That's crazy low man, were you doing the same node as OP?
It's a relevant sample size because 99.9% of the player base isn't going to refresh their energy 9 times in a row. So even this extreme example demonstrates how bogus drop rates are
How any times you actually run it doesn't mean squat related to a relevant sample size. By this logic I should always get 1/3 every day I aim 3 hard nodes. Just because we don't or can't run an accurate sample size doesn't make the true sample size any less relevant.
I don't think you know what probability means. There are statistical tests for this. His sample size is more than enough for an adequate estimate on the drop rate. Only very low % values require such big groups. He's not "proving" anything, he's giving an estimate.
Which is good, considering all the crybabies saying they ran it 50 times and got 1-2 pieces.
100% correct. In fact, if you use the calculator, and assume that 0.33 is the actual probability and use 20 as the number of successes, you can see that there's a .999999 probability that the average person will get 20 . So if one knows how to use that calculator, this just further proves that the actual drop rate is correct at 0.33 probability of you were to only obtain 20 on any given 100 sample run.
What relevance does using statistics have in regards to this game, though? It's not TRULY random. It's all designed to give you a certain payout over a certain period of time. You will NEVER go past that limit. They will never give you true random chance. Gear is the only way they can keep whales from outgrowing content faster than the devs can produce it. You will ALWAYS be behind on gear
It's completely relevant to use statistics because those complaining that the drop rates aren't what they say they are can be proved using stochastic calculations. And when it is proved, or demonstrated with high confidence, the argument then changes to "well it's not truly a probability thing, they programmed it to not be that way", which is insane because the only way they can program it so that some people have seen a high drop rate and some have witnessed a low drop rate is by using random numbers.
That being said I don't truly believe that all shard nodes or gear nodes are the exact same probability, by design. Lower gear is higher probability in all likelihood. But you can't say that statistics have no relevance here. It's the only sane argument to a topic fueled by observation bias and emotion.
Apparently you didn't get his point. He's arguing the terminology used. Also, I wouldn't say that 100 is a good sample size for even an "adequate" estimate. I'm at exactly 50 handcuff pieces in 92 attempts. Can I now go ahead and create a thread stating "Proof of handcuffs having 50% drop rate"? No, of course no. That was the drop rate of my alt (which apparently always is luckier than my main) and based on history I know that such luck will probably not happen to me again within the next 6 months or perhaps ever again.
A rough estimate could be made at 1000 attempts. A good claim, like Valent_Antona stated would require 100*100 attempts. The video shows nothing more but what drop rate he had based on the attempts. It gives you an idea not to expect a 100% drop rate, and not to expect a drop rate lower than 5%.
No, 116 is an insufficient sample size to determine the probability of this purple drop. See quoted section below. The drop rate identified by that test is 21 +/-9 19 times out of 20.
Also when you say only very low values require large groups is sort of misleading and the value of error doesn't change with the probibility. The signicance of the error does so you want larger sample sizes
I don't believe statistics is relevant here because I don't believe gear drops are true random chance. I believe they're controlled by an algorithm that won't dispense over a certain amount of mats during a certain period of time, or something similar to that
Streak breakers to end cold or limit hot streaks make sense but even with those the overall drop rates would represent a binomial distribution outside of the outliers.
So statistical analysis would still be relevant even with noise generated at the boundaries.
I'm dont understand the drivers for a non random system.
"to those that do not understand electricity, a light bulb seems like magic".
Profit? Just sayin.
Amazing to me.... In stuck @6%....I spent 8k energy for 50 cuffs.
An engineer(person designing an actual house to be built): You will very quickly be close enough to that wall so the remaining distance doesn't matter.
Mathematician: Technically, I'm still right.
Engineer: And realistically, it doesn't **** matter.
The amount of **** measuring on this forum is amazing. Thanks for the data. It measures up pretty close to what I have (79/401 according to excel)
Mk 3 Czerka Stun Cuffs (S)
Attempts 391
Drops 99
Drop % 25%
I already said why. To keep whales from consuming content faster than they can produce it. Drop $5,000 and you pretty much have all the characters at 7* or close to it. That's a random system. What's next? Gear them. And that's where EA sets the cap. Players will NEVER push past this cap. It's set at an artificial level to keep them chasing. Then slowly they introduce things like the latest Aurodium pack that contain some purple items and the whales gobble these up like starving refugees.
The problem is you're thinking like a mathematician. This game is run by businessmen. They want money. They're absolute geniuses when it comes to manipulating people for it. And going by responses in this topic they've been doing an amazing job
Everyone agrees EA is **** greedy with their drop rates, what the **** is new? They're not changing **** because they're overpaid as ****.
If that 1% top earner of this game gave away his money and even gave a ****, this game would be so much better, **** drop rates
Your history is getting closer to a good sample set, which is promising. The only thing it proves is that at the specified drop rate probability, the actual standard dev is larger than expected. It doesn't mean the actual probability is wrong.
Over the course of a very long time, we can expect what they indicate the rate is. On average some may find more or less.
Your missing the psychology. This entire game is a slot machine with no jackpot. Small incremental wins keep people feeding quarters into the machine. There would also be outcry from the megawhales if their drop rate disappeared. Also if caught after saying the drop rate doesn't change they would kill their credibility. The escalating cost of energy is what prevents whales from burning through content. You fix the amount of energy available and don't allow banking and you restrict pace. And if they did take time to code a complex drop system they would have also put in streak breaking logic at the bottom end.
The fact that their is consensus forming at somewhere between 15-25% suggests that it is random.