Not real random numbers. All the results come from a pseudo-random number generator. What you've proven is that PRNGs don't produce real random numbers.
Yeah. At first I was getting at least 1 or 2 a day then after a while I would go days without finding a single egg. Edit: Then I'd go days more without finding any after I got a couple. And this was across 3 characters I was farming the event on each time, with all 3 not being able to find any eggs for several days at a time. Heck I think I went over a week at least once without any eggs. It's been a common complaint since the first summer event back in 2013.
Yes, I meant the eggs, sorry. Yeah on my unlucky characters I did not find a total of one per day so I had to purchase some eggs to make a parrot each day. However, other characters had 7 or 8 extra eggs. But overall it was 2.5% which I believe is the intent. Maybe the unlucky and lucky characters are also tainted by a 'sticky' RNG.
Captain Jean-Luc Picard: "We think we've come so far. Torture of heretics, burning of witches, it's all ancient history. Then - before you can blink an eye - suddenly it threatens to start all over again."
"With the first link, the chain is forged. The first speech censured, the first thought forbidden, the first freedom denied, chains us all irrevocably."
I just queued up 8 Superior Weapons Tech Upgrades, with a 72% chance of Critical Success.
I crit 1/8 of them.
Using a standard binomial distribution, the odds of this happening are %0.07.
Don't tell me that these things aren't rigged.
The Chance for that happening is non zero, thus possible.
This reminds me of a recent bug report on the forums. Someone claims they have a 0% to fail/disaster on a doff assignment and yet it fails. I explained that even if you have a 0.1% chance it's still a chance and having it that low would make the bar look empty, thus giving the impression you have no failure chance at all.
And yes, when all R&D projects have a 20%+ chance of failure it is very possible they all won't crit.
As pointed out, the odds are individually based, not combined. Just like opening lock boxes you have a chance at a ship from each one and it doesn't mean the more boxes you open the higher your chance is. Each box has the same chance.
I agree with the people who see nothing odd here aside from an improbable event happening.
That said, if I were a systems designer, I'd investigate other methods of handling randomization because the current approach creates very unhappy outliers while generating negligible increased profits for Cryptic.
Odds of getting a grand prize from a lockbox are 1 in 200, roughly. But because each lockbox in a separate chance, some people will never get the grand prize regardless of how many they open. That's probability. There is an issue with limited sampling as well.
Now, does it benefit Cryptic that some people may spend $2000 and never get the grand prize? No. Because figuring out profitability from Cryptic's perspective, they are giving away one ship for every $225 spent. There is no sampling error from their POV. They would make exactly the same money if $225 guaranteed someone the ship. From their POV, the outliers are offset by the people who are lucky and the average acquisition cost is $225. Except they bear the brunt of dissatisfied customers who spend more than $225 which has a net negative on financials in the long run.
Now, let's look at a different model. Instead of a lockbox, imagine the player character is in a cargo bay with 200 boxes. Now imagine one of those boxes is guaranteed to contain the grand prize ship and it's a process of elimination which one has it; boxes reset once a week. Cryptic's profits in this system would be virtually identical to the current system but anyone who spends $225 in one sitting is guaranteed the ship. Cryptic's haul is the same but you avoid the disgruntled player. Because the acquisition cost of the ship on average remains the same which is what matters from Cryptic's POV. And the whale player is happier and less disgruntled which is GOOD because if somebody is willing to spend $225, you don't want a big spender to be a disgruntled statistical outlier. Costs Cryptic NOTHING except initial development costs. Improves longterm prospects.
There's over 100,000 people in this game. If all of them craft 8 beam upgrades at that chance, then about 70 of them will have this happen. I'm sorry that you are one of the 70 but it is not impossible.
Same as any lottery: the odds of you winning the big prize are incredibly small, though still almost every week someone does it. So it is all about the numbers of people playing. And the more are playing, the more will be incredibly lucky - and the more will be incredibly unlucky. To expand: if 100k players tried upgrading 8 times with these odds, the chances of none of them being left completely without success would be abou 2.3%. So most probably there is somebody worse off.
Not real random numbers. All the results come from a pseudo-random number generator. What you've proven is that PRNGs don't produce real random numbers.
We already knew that.
This on the other hand is only true in a very technical sense. There are a lot of RNGs freely available that will produce results that are only getting funny, i. e. start failing mathematical tests, after a few billion iterations. Take that plus the inherent randomness of the game (like who buys what when and - depending on implementation of the RNG's results to the game - who activates a Tachyon Beam in a completely unrelated STF while you're upgrading) and you will have no chance in hell to differentiate between true randomness and a modern RNG like the Mersenne Twister.
That said, if I were a systems designer, I'd investigate other methods of handling randomization because the current approach creates very unhappy outliers while generating negligible increased profits for Cryptic. (...)
Certainly a possibility, but it may just open a whole new can of worms. Apart from the implementation aspect (you would have a counter for each player on each lock box or similar on what he has already gotton for numbers, which should be doable), you may get the opposite effect in certain cases and thus diminish profits.
Say I get only those lovely CXP boosters from a lock box for a while, so I will continue (if I know about the mechanics) to buy stuff because I know it will get better. And finally I will have the Annorax or Sheshar or whatever. I am happy, Cryptic is happy because I bought 100 or so - maybe even more - keys.
Now imagine I am one lucky guy, call me Gladstone Gander, and get the ship with the first box. I do know that I will have a really long draught until something decent has the chance to pop again. I won't buy any keys anymore.
In my opinion there is only two options: make it completely random - which will mean some players will be out of luck completely - or make it completely predictable, so everybody knows what he's getting. Both would be okay with me, but I am fairly certain, Cryptic would generate way less revenue when choosing the latter option.
My mother was an epohh and my father smelled of tulaberries
I just queued up 8 Superior Weapons Tech Upgrades, with a 72% chance of Critical Success.
I crit 1/8 of them.
Using a standard binomial distribution, the odds of this happening are %0.07.
Don't tell me that these things aren't rigged.
The Chance for that happening is non zero, thus possible.
This reminds me of a recent bug report on the forums. Someone claims they have a 0% to fail/disaster on a doff assignment and yet it fails. I explained that even if you have a 0.1% chance it's still a chance and having it that low would make the bar look empty, thus giving the impression you have no failure chance at all.
And yes, when all R&D projects have a 20%+ chance of failure it is very possible they all won't crit.
As pointed out, the odds are individually based, not combined. Just like opening lock boxes you have a chance at a ship from each one and it doesn't mean the more boxes you open the higher your chance is. Each box has the same chance.
I agree with the people who see nothing odd here aside from an improbable event happening.
That said, if I were a systems designer, I'd investigate other methods of handling randomization because the current approach creates very unhappy outliers while generating negligible increased profits for Cryptic.
In champs they added lock box scratch cards that give you rewards for opening boxes.
The chance of upgrading is pr-item, not an overall chance.
Like the grab-bags... theres a 0.5% chance you get an item... Not 0.5% overall (wich would mean that you would get the lottery-prize every 200 box), but 0.5% for the box you're opening.
It resets every time.
Don't look silly... Don't call it the "Z-Store/Zen Store"...
This. It's baffling how many people have no idea what chance and probability mean. Even if they claim they do but demonstrate otherwise.
^ Memory Alpha.org is not canon. It's a open wiki with arbitrary rules. Only what can be cited from an episode is. ^
"No. Men do not roar. Women roar. Then they hurl heavy objects... and claw at you." -Worf, son of Mogh
"A filthy, mangy beast, but in its bony breast beat the heart of a warrior" - "faithful" (...) "but ever-ready to follow the call of the wild." - Martok, about a Targ
"That pig smelled horrid. A sweet-sour, extremely pungent odor. I showered and showered, and it took me a week to get rid of it!" - Robert Justman, appreciating Emmy-Lou
That said, if I were a systems designer, I'd investigate other methods of handling randomization because the current approach creates very unhappy outliers while generating negligible increased profits for Cryptic.
even though true randomization is impossible for a computer the methods used to create randomization are so thorough the minuscule discrepancies encountered are hardly anything to gawk over. particularly in a video game. it's only of concern if you are trying to imitate something in nature.
Now, does it benefit Cryptic that some people may spend $2000 and never get the grand prize? No.
Yes. Yes it does. Why?
Except they bear the brunt of dissatisfied customers who spend more than $225 which has a net negative on financials in the long run.
Because there is always another idiot to take that persons place.
F2P is a "thing" not because it makes players happy, it's a thing because it generates more revenue. The reason it generates more money than a static price model is directly dependent on a portion of the player base to overpay. If you remove the opportunity for over-payment you reduce the validity of the F2P model. It's not much more complicated than that. And why should a business give two cares about an outlier customer? Their outliers, edge cases.
This isn't Gambler's Fallacy. The OP isn't saying the 8th attempt has a higher chance of success because the previous 7 attempts failed, he's saying the chance that 7 out of 8 attempts fail, when each attempt has a 72% chance of success, is incredibly small, and it is.
My guess is that 72% chance is a lie, like some others have said more or less.
This isn't Gambler's Fallacy. The OP isn't saying the 8th attempt has a higher chance of success because the previous 7 attempts failed, he's saying the chance that 7 out of 8 attempts fail, when each attempt has a 72% chance of success, is incredibly small, and it is.
Trust me I get it, I did this at school and it's not difficult. Which is why I don't understand why the OP is quoting that as his proof.
The OP is only correct in that his odds of getting the crit were better than not critting. To then say his odds of getting the result he did, while small, is to some extent a logical fallacy as it is ignoring the fact that it is "possible" for them all to have critted. There is always going to be an element of luck involved with where the die will land even if it is loaded in your favour.
Ten soldiers wisely led will beat a hundred without a head. - Euripides
I no longer do any Bug Hunting work for Cryptic. I may resume if a serious attempt to fix the game is made.
Not real random numbers. All the results come from a pseudo-random number generator. What you've proven is that PRNGs don't produce real random numbers.
We already knew that.
This has nothing to do with pseudo-random numbers. It could happen with real random numbers. A low probability decidedly does not mean the same as impossible. Otherwise no one would have ever won in the lottery.
Star Trek Online Advancement: You start with lowbie gear, you end with Lobi gear.
Not real random numbers. All the results come from a pseudo-random number generator. What you've proven is that PRNGs don't produce real random numbers.
We already knew that.
This has nothing to do with pseudo-random numbers. It could happen with real random numbers. A low probability decidedly does not mean the same as impossible. Otherwise no one would have ever won in the lottery.
I'll just go ahead and disagree with you there. Unless the engine uses Mersenne Twister (and I admit that it might) or some equally sophisticated PRNG, there will be issues with the PRNG showing a bias toward certain outcomes. I'd already noted some bias in the PRNG inside City of Heroes, and this engine is at least based on that. Even with a very sophisticated PRNG, a modulus ranging step at the end can force a bias in the outcome.
Obviously, 25% + 25% + 25% + 25% = 100%. Or does it? You are forgetting one major thing. Statistics don't mean anything, at all. Having multiple stuff with % chances do not add together. The number of projects you are running are not accumulative.
The first project is 72%.
The second project is 72%.
The tenth project you run is still 72%.
So, just because you have multiple projects running, that does not increase your chances. EACH project, INDIVIDUALLY is 72%. And this means that you very easily can, and will end up without a crit on any of them.
And this is also the reason why you can, and WILL dump more than $100 into lockboxes, and ONLY get duty officers and mining claims. If each lockbox is 1%, then obviously, the 40th lockbox should be 40%, and the 90th box should be 90%. And the 100th box is a guaranteed Temporal Science Vessel. Correct? Wrong. Your first lockbox is 1%. Your 20th lockbox is 1%. Your 100th lockbox is still 1%.
So, no matter how many crafting missions you have running, EACH project has its own individual % chance, and that is what will determine your result.
Basically, it is a slot machine. And your results with a slot machine all depends on who made that slot machine. Whether mechanical, or pure digital, slot machines can easily be rigged to never land on the jackpot.
That said, if I were a systems designer, I'd investigate other methods of handling randomization because the current approach creates very unhappy outliers while generating negligible increased profits for Cryptic.
even though true randomization is impossible for a computer the methods used to create randomization are so thorough the minuscule discrepancies encountered are hardly anything to gawk over. particularly in a video game. it's only of concern if you are trying to imitate something in nature.
Now, does it benefit Cryptic that some people may spend $2000 and never get the grand prize? No.
Yes. Yes it does. Why?
Except they bear the brunt of dissatisfied customers who spend more than $225 which has a net negative on financials in the long run.
Because there is always another idiot to take that persons place.
F2P is a "thing" not because it makes players happy, it's a thing because it generates more revenue. The reason it generates more money than a static price model is directly dependent on a portion of the player base to overpay. If you remove the opportunity for over-payment you reduce the validity of the F2P model. It's not much more complicated than that. And why should a business give two cares about an outlier customer? Their outliers, edge cases.
Death and taxes.
Right but in a room with 200 boxes, 112.5 ZEN average cost per attempt and only one grand prize in the room, how much does Cryptic make per ship? $225.
In the current scenario where you have keys that cost 112.5 ZEN in a bulk purchase with a 0.5% droprate, how much does Cryptic make per ship? $225.
The proposed systems can be tweaked around the edges if necessary (adjusting number of boxes, adjusting Lobi averages, etc). But it should generate Cryptic the same return because they'd generate the same money per ship. You're dealing with variances being behavioral here but from what I know of gambling/gaming research, the knowledge that one of the boxes in a physical room has the prize should drive UP participation if anything.
I love these discussions, because almost everyone always falls into the trap of assuming that a pseudo-number generator (especially in an MMO) is actually random, lol. Debating the nature and calculation of probability is futile, when your base assumption is incorrect: that the event(s) is truly random in the first place.
I love these discussions, because almost everyone always falls into the trap of assuming that a pseudo-number generator (especially in an MMO) is actually random, lol. Debating the nature and calculation of probability is futile, when your base assumption is incorrect: that the event(s) is truly random in the first place.
Of course it's not random. The numbers are fixed to TRIBBLE us over so we get tempted to spend more monies.
Who cares anyway? It's been established that it's not impossible and there really isn't any more to add to the discussion, other then to argue about math and stuff.
For the uninitiated into Pseudo Random Number Generators, a PRNG uses a seed number then puts it through various calculations to get the random numbers, for a single player game, providing you don't need hundreds of numbers generated per second, your processors clock speed is usually sufficiently random enough to generate these numbers, too many numbers required though and you get slowdown as the computer is constantly checking the computers clock speed.
So what does something of the scale of STO use, where the servers are nearly constantly bombarded with PRNG requests for crafting, upgrading, lockboxes, drops, crit chances, ect. ect.?
Usually the current time value, it's why in some other large games some people report that doing things at a certain time gives higher chances than others, because the value used has a higher chance of generating the numbers needed (not guaranteed, please do not read that as guaranteed) to get the desired results.
Here though, it means that if you fail a lot, you're more than likely to keep failing so just stop, take a deep breath, and try again later, it doesn't however mean that the odds are guaranteed to be zero or that the probabilities shown are incorrect, it just means that the die used has a slight nick that's causing it to roll slightly unevenly at the moment.
TLDR: If you're failing a lot, stop and try later.
Take the whole world population, roughly 7,300,000,000 people, and have everyone toss a coin.
Given a 50% chance, 3,650,000,000 people just tossed heads. Let them toss again.
Given a 50% chance, 1,825,000,000 people just tossed heads. Let them toss again.
Rinse and repeat...
After 20 tosses you will have roughly 7000 people that just tossed 20 heads in a row.
The probability of tossing 20 heads in a row is: 0.5 ^ 20 = 0.000095%
This is highly improbably, however 7000 people just did it
I love these discussions, because almost everyone always falls into the trap of assuming that a pseudo-number generator (especially in an MMO) is actually random, lol. Debating the nature and calculation of probability is futile, when your base assumption is incorrect: that the event(s) is truly random in the first place.
Random enough is effectively random.
If you can control the latency and the number of requests per period, why are you wasting your time playing a video game? Use that power to optimize the data flow of the internet for the good of all humanity!
This is my Risian Corvette. There are many like it, but this one is mine.
obviously the author of the thread is a victim of an evil scientist conspiracy planned and executed by PWE in China. They not only want your jobs and your wifes, but also your random numbers. He better start loving delta rising too, or these "coincidents" will happen to him more frequently.
This doesn't really prove your point. I don't think you really understand what 1 - ( ( 1 - x ) ^ y ) is telling you.
What it is saying is that at 72% with 8 tries you have a 99.99% chance to crit ONCE. Not all 8 times. For each roll the chance is still 72% that doesn't change which means every roll 28% chance of fail. You could look at this math the other way... and at 28% chance to fail in 8 rolls you have a 92.77% chance to fail at least one of them.
As others have mentioned go do some reading about the Gambler's fallacy Each roll has the same 28% chance to fail as the one before it. Your thinking is how Casinos take your money.
Comments
We already knew that.
"With the first link, the chain is forged. The first speech censured, the first thought forbidden, the first freedom denied, chains us all irrevocably."
I agree with the people who see nothing odd here aside from an improbable event happening.
That said, if I were a systems designer, I'd investigate other methods of handling randomization because the current approach creates very unhappy outliers while generating negligible increased profits for Cryptic.
Odds of getting a grand prize from a lockbox are 1 in 200, roughly. But because each lockbox in a separate chance, some people will never get the grand prize regardless of how many they open. That's probability. There is an issue with limited sampling as well.
Now, does it benefit Cryptic that some people may spend $2000 and never get the grand prize? No. Because figuring out profitability from Cryptic's perspective, they are giving away one ship for every $225 spent. There is no sampling error from their POV. They would make exactly the same money if $225 guaranteed someone the ship. From their POV, the outliers are offset by the people who are lucky and the average acquisition cost is $225. Except they bear the brunt of dissatisfied customers who spend more than $225 which has a net negative on financials in the long run.
Now, let's look at a different model. Instead of a lockbox, imagine the player character is in a cargo bay with 200 boxes. Now imagine one of those boxes is guaranteed to contain the grand prize ship and it's a process of elimination which one has it; boxes reset once a week. Cryptic's profits in this system would be virtually identical to the current system but anyone who spends $225 in one sitting is guaranteed the ship. Cryptic's haul is the same but you avoid the disgruntled player. Because the acquisition cost of the ship on average remains the same which is what matters from Cryptic's POV. And the whale player is happier and less disgruntled which is GOOD because if somebody is willing to spend $225, you don't want a big spender to be a disgruntled statistical outlier. Costs Cryptic NOTHING except initial development costs. Improves longterm prospects.
Same as any lottery: the odds of you winning the big prize are incredibly small, though still almost every week someone does it. So it is all about the numbers of people playing. And the more are playing, the more will be incredibly lucky - and the more will be incredibly unlucky. To expand: if 100k players tried upgrading 8 times with these odds, the chances of none of them being left completely without success would be abou 2.3%. So most probably there is somebody worse off.
This on the other hand is only true in a very technical sense. There are a lot of RNGs freely available that will produce results that are only getting funny, i. e. start failing mathematical tests, after a few billion iterations. Take that plus the inherent randomness of the game (like who buys what when and - depending on implementation of the RNG's results to the game - who activates a Tachyon Beam in a completely unrelated STF while you're upgrading) and you will have no chance in hell to differentiate between true randomness and a modern RNG like the Mersenne Twister.
Certainly a possibility, but it may just open a whole new can of worms. Apart from the implementation aspect (you would have a counter for each player on each lock box or similar on what he has already gotton for numbers, which should be doable), you may get the opposite effect in certain cases and thus diminish profits.
Say I get only those lovely CXP boosters from a lock box for a while, so I will continue (if I know about the mechanics) to buy stuff because I know it will get better. And finally I will have the Annorax or Sheshar or whatever. I am happy, Cryptic is happy because I bought 100 or so - maybe even more - keys.
Now imagine I am one lucky guy, call me Gladstone Gander, and get the ship with the first box. I do know that I will have a really long draught until something decent has the chance to pop again. I won't buy any keys anymore.
In my opinion there is only two options: make it completely random - which will mean some players will be out of luck completely - or make it completely predictable, so everybody knows what he's getting. Both would be okay with me, but I am fairly certain, Cryptic would generate way less revenue when choosing the latter option.
My character Tsin'xing
The chance of upgrading is pr-item, not an overall chance.
Like the grab-bags... theres a 0.5% chance you get an item... Not 0.5% overall (wich would mean that you would get the lottery-prize every 200 box), but 0.5% for the box you're opening.
It resets every time.
This. It's baffling how many people have no idea what chance and probability mean. Even if they claim they do but demonstrate otherwise.
Get the Forums Enhancement Extension!
even though true randomization is impossible for a computer the methods used to create randomization are so thorough the minuscule discrepancies encountered are hardly anything to gawk over. particularly in a video game. it's only of concern if you are trying to imitate something in nature.
Yes. Yes it does. Why?
Because there is always another idiot to take that persons place.
F2P is a "thing" not because it makes players happy, it's a thing because it generates more revenue. The reason it generates more money than a static price model is directly dependent on a portion of the player base to overpay. If you remove the opportunity for over-payment you reduce the validity of the F2P model. It's not much more complicated than that. And why should a business give two cares about an outlier customer? Their outliers, edge cases.
Death and taxes.
My guess is that 72% chance is a lie, like some others have said more or less.
Trust me I get it, I did this at school and it's not difficult. Which is why I don't understand why the OP is quoting that as his proof.
The OP is only correct in that his odds of getting the crit were better than not critting. To then say his odds of getting the result he did, while small, is to some extent a logical fallacy as it is ignoring the fact that it is "possible" for them all to have critted. There is always going to be an element of luck involved with where the die will land even if it is loaded in your favour.
I'll just go ahead and disagree with you there. Unless the engine uses Mersenne Twister (and I admit that it might) or some equally sophisticated PRNG, there will be issues with the PRNG showing a bias toward certain outcomes. I'd already noted some bias in the PRNG inside City of Heroes, and this engine is at least based on that. Even with a very sophisticated PRNG, a modulus ranging step at the end can force a bias in the outcome.
The first project is 72%.
The second project is 72%.
The tenth project you run is still 72%.
So, just because you have multiple projects running, that does not increase your chances. EACH project, INDIVIDUALLY is 72%. And this means that you very easily can, and will end up without a crit on any of them.
And this is also the reason why you can, and WILL dump more than $100 into lockboxes, and ONLY get duty officers and mining claims. If each lockbox is 1%, then obviously, the 40th lockbox should be 40%, and the 90th box should be 90%. And the 100th box is a guaranteed Temporal Science Vessel. Correct? Wrong. Your first lockbox is 1%. Your 20th lockbox is 1%. Your 100th lockbox is still 1%.
So, no matter how many crafting missions you have running, EACH project has its own individual % chance, and that is what will determine your result.
Basically, it is a slot machine. And your results with a slot machine all depends on who made that slot machine. Whether mechanical, or pure digital, slot machines can easily be rigged to never land on the jackpot.
My Ship Builds: USS Conqueror, HMS Victorious, HMS Concord, ISS Queen Elizabeth, Black Widow III
Click here to view my DeviantArt.
Right but in a room with 200 boxes, 112.5 ZEN average cost per attempt and only one grand prize in the room, how much does Cryptic make per ship? $225.
In the current scenario where you have keys that cost 112.5 ZEN in a bulk purchase with a 0.5% droprate, how much does Cryptic make per ship? $225.
The proposed systems can be tweaked around the edges if necessary (adjusting number of boxes, adjusting Lobi averages, etc). But it should generate Cryptic the same return because they'd generate the same money per ship. You're dealing with variances being behavioral here but from what I know of gambling/gaming research, the knowledge that one of the boxes in a physical room has the prize should drive UP participation if anything.
Of course it's not random. The numbers are fixed to TRIBBLE us over so we get tempted to spend more monies.
My Ship Builds: USS Conqueror, HMS Victorious, HMS Concord, ISS Queen Elizabeth, Black Widow III
Click here to view my DeviantArt.
So what does something of the scale of STO use, where the servers are nearly constantly bombarded with PRNG requests for crafting, upgrading, lockboxes, drops, crit chances, ect. ect.?
Usually the current time value, it's why in some other large games some people report that doing things at a certain time gives higher chances than others, because the value used has a higher chance of generating the numbers needed (not guaranteed, please do not read that as guaranteed) to get the desired results.
Here though, it means that if you fail a lot, you're more than likely to keep failing so just stop, take a deep breath, and try again later, it doesn't however mean that the odds are guaranteed to be zero or that the probabilities shown are incorrect, it just means that the die used has a slight nick that's causing it to roll slightly unevenly at the moment.
TLDR: If you're failing a lot, stop and try later.
Rayzee
excellentawesome#4589
torgaddon101
raeat
Given a 50% chance, 3,650,000,000 people just tossed heads. Let them toss again.
Given a 50% chance, 1,825,000,000 people just tossed heads. Let them toss again.
Rinse and repeat...
After 20 tosses you will have roughly 7000 people that just tossed 20 heads in a row.
The probability of tossing 20 heads in a row is: 0.5 ^ 20 = 0.000095%
This is highly improbably, however 7000 people just did it
Random enough is effectively random.
If you can control the latency and the number of requests per period, why are you wasting your time playing a video game? Use that power to optimize the data flow of the internet for the good of all humanity!
Yes yes the chances are small in and just a bit under 1% mark really.
Still the chances of pulling a lockbox ship is around the same... and the market is flooded with them. So ya do the math my friend.
This doesn't really prove your point. I don't think you really understand what 1 - ( ( 1 - x ) ^ y ) is telling you.
What it is saying is that at 72% with 8 tries you have a 99.99% chance to crit ONCE. Not all 8 times. For each roll the chance is still 72% that doesn't change which means every roll 28% chance of fail. You could look at this math the other way... and at 28% chance to fail in 8 rolls you have a 92.77% chance to fail at least one of them.
As others have mentioned go do some reading about the Gambler's fallacy Each roll has the same 28% chance to fail as the one before it. Your thinking is how Casinos take your money.