**Mastery Loadouts**
Due to issues related to the release of Mastery Loadouts, the "free swap" period will be extended.
The new end date will be May 1st.
Due to issues related to the release of Mastery Loadouts, the "free swap" period will be extended.
The new end date will be May 1st.
New drop rates display
This discussion has been closed.
Comments
That made no sense.
This is indeed true. I know that when you're after a certain Champ, it can seem like the odds are stacked against you, but I can assure you, the Drop Rates on all of these Champions is the exact same (unless otherwise stated).
To reduce the likelihood of self-reporting bias, I only included data from crystal openings that were originally live streamed. This mostly eliminates the possibility of only including particularly good or particularly bad data.
When I say "on Reddit" I mean I posted it on Reddit, not that I analyzed Reddit posts. Back then I use to post numerical posts there more than here, because honestly the amount of quantitative discussions of the game here was dismal.
I agree statistical bias is very tricky to eliminate from analysis: you have to have experience in doing it or you're likely to run into potential problems. For example, even the recent Blade crystal poll contained a subtle bias: many players chose to spin until they got him. Spin until you win introduces skew, and anyone willing to do the binomial expansion will see that spin until you win can skew a 20% result towards 40%. Because most people didn't do spin until you win, the actual collected data was probably only skewed slightly due to that problem. This kind of thing is notorious to remove completely, and I've had years of practice but still originally missed that one until someone asked about it.
I went 0-10 on blade featured crystals and someone went 1-1.
Just because the theoretical odds are 50% doesn’t mean that will happen in practice. Try flipping a coin. You don’t get head and tails every other time. However, over the course of your experiment you will get back to 50/50 probabilities.
Doesn’t mean I’m still not salty about going 0-10 on featureds though. Arrggfhhhhh. Haha.
If you don't believe their numbers, why would you believe their algorithms? The algorithm for randomized lootboxes is fairly simple. You have a reward table that includes all the entries and numerical weights that determine the odds of each entry being chosen. When you have tiers, for example PHC crystals contain 2* 3* and 4* champions at different relative odds, you use a multi-tier drop table. The game seems to be pulling the rarity table weights and calculating the drop percentage and then displaying it.
Now, if you think Kabam is straight up lying to you, why wouldn't they just tell you what the textbook stochastic loot box algorithm is, whether they use it or not, to make you go away?
That's a bit of a mangle of the word "probability." The only difference between theoretical probability and experimental probability is that your theory might be wrong. I think I understand what you are trying to say, but I wouldn't put it that way because it encourages people to think that probability and statistics sometimes "don't work in the real world." That's not the case at all.
Probability and statistics by definition do not guarantee a particular result. They only describe the likelihood of particular results before they happen. They encapsulate uncertainty. This is a subtle but important distinction. Before I open a featured Blade crystal the odds are (for the sake of discussion) 20% that I will get Blade. *After* I open the crystal the odds are exactly 100% that I will get what I got, because there is now zero uncertainty to that question.
When we say some event has a certain statistical chance to occur, we are not saying that we are guaranteeing that this thing will happen at a specific rate. If we could make that guarantee, then we have absolute certainty about what will happen and statistics no longer applies. The probability of getting one in five is 100% if we are certain this will happen.
When we measure probability, most experiments to measure probability cannot measure probability with absolute certainty, because by definition all probability contains uncertainties. All we can do is increase our confidence that the probability of something happening is within a certain range. But all of Science relies on being confident in measuring probabilities not perfectly, but accurate to within a certain level of confidence. Even the computers everyone is reading this post on relies on integrated circuits that operate in part on confidence in statistical quantum mechanics. There is only a certain probability that electrons will tunnel across transistor gate voltages, and your computer will only function correctly if that probability is high enough to make errors vanishingly unlikely.
Before GMC’s and i got 2 4* GR and another 4* champ. Then you’ll go through 300-400 without a single 4*.
It pretty much means these are the drop rates for each individual crystal you open and not just guesstimated drop rates by someone who claims to be a guru of this game. No “well if you open 10 crystals...”, no “well from what I’ve experienced...”, and no “well you should be due for one soon since...”. Flat drop rates per each individual crystal. All day. Every day.
When we analyze the hundreds of crystals opened, the probability of pulling Blade is 10% (for example), but the statistics show I pulled him 100% of the time (1 for 1).
Yes, I glossed over that, but you are correct that there is a difference between probability and statistics. Also they are not linked. You can use statistical analysis to analyze things with no probability associated with them. For example, you can statistically analyze the words Shakespeare used in all of his works. You end up with the frequency of every word he used, but that data has no random components associated with them.
And if there is an actual mathematician reading this, strictly speaking computer PRNGs are not absolutely random: knowing the algorithm and the seed makes them absolutely deterministic. But since we don't know what the initializing seed is and we do not know how quickly the generated sequence is being consumed by the game we cannot predict what the visible output sequence will be and statistical analysis then measures our uncertainty about the generated sequences represented as a probability.
These distinctions are not normally important in colloquial conversation, but they become important when people don't fully understand or misapply these tools and draw incorrect conclusions.
All that is gunna happen is instead of
“I cant get a 4* ur crystals are rigged” “phc should be just called a 2*”
All you are gunna get now is
“**** it is not 2% chance stop lying” and “2% chance for your whales, not for the f2p”
Nothing will change at all.... salty people will always be salty... stupid people will always be stupid and complainers will always complain....
Can't save everyone. Some people live in an evidence-based world, and the rest you have to let live in their own encapsulated phantasms. They shouldn't be the determiner for the information everyone else benefits from.
-Niko
Point proven..... lol.... right on queue..... lol....
Correct, a computer does not have the capability to generate random, it’s usually some equation, with ticks tied into it, but if you knew the formula and had impossibly perfect timing, you could theoretically get only the absolute best pulls.
The benefit of living in an evidence-based world is that mass delusions require constant approbation from willing co-conspirators. There's no such requirement when all you care about is the facts. There is no threat to the facts that laughter can ultimately pose. I've seen it come, and more importantly on the time scales I care about I've always seen it go.
Even if you knew the formula used to generate the result..... and you could reverse that calculation to determine the input seed required to get your desired result how would you ensure that your crystal opening entered that exact seed....
Without being able to hack and enter the seed you want....
Its not like time of day would determine thenseed used as that would not be random.... if there was a determining influence surely we would be able to see a pattern without knowing the formula
Knowing the formula would just allow us to understand it properly.... would still need the way to actually affect he pull
Hypothetically speaking if we knew the algorithm and precisely how it was used in crystal drops, if it was the right algorithm like say Mersenne Twister it is mathematically possible to reverse the seed after seeing enough drops. You could then choose the date and time of a crystal opening to coincide with the right generated number to get what you wanted, if you also knew the precise nature of the reward table construction and all of the generated numbers between then and now. None of these things is practical except under very contrived circumstances, but it is not mathematically impossible. It would probably be easier to hack the game and just add the champion to your roster directly.
I doubt the PRNG is reseeded often. Most developers don't do that because under most circumstances it doesn't have any real benefit, because attacks against the seed in a game environment are so unlikely and impractical in the first place. If I had to guess, my guess would be that the server side PRNG(s) are reset when the servers are rebooted.
Reseeding itself can be a potential weakness if you don't do it right so it is not automatically correct to reseed constantly. Some generators, for example, require you to throw away the first few generated numbers after initialization. If you forget to do that, reseeding can weaken your random number generation. Probably still not enough for game players to notice, but still.
Meaning mathematically possible but far from practical to do so
I'm both a coffee & a tea person, although I drink more coffee each day
But in this case, I would rather join you for tea
In the end, it is nice to have the drop rates. But whether it is accurate, bugged or working as intended, we will never know.
Right, I highly doubt they have an uncrackable setup containing a very precise barometric pressure device, and this value tying into the algorithm as this would be absolute overkill for this situation.
There's no reason to use barometers in this situation, when collecting entropy with network packet timing would always work perfectly fine when you're dealing with massively multiplayer game servers.
When the Premium Hero Crystal was introduced, the landscape of crystals was different. So you had the free daily crystal (that wasn't just champions), and the free 4 hour crystal (that has no champions). And that was pretty much it. You had special featured crystals like romance, or avengers crystals, and 3 and 4 star crystals were only available through offers (no shards in the game). So it was kind of a premium crystal. They should probably update the name.
Kabam says spinning or not spinning doesn't affect crystal odds. This is backed up by the fact that you get the reward before the crystal even starts visually spinning, which makes it very unlikely that spinning alters the odds for that crystal.
People say this a lot but it’s not true. Champ is determined once the crystal stutters. Don’t believe me? Buy a crystal and spin it. Then close game. It will still be in your inventory. But wait until you see the hesitation and then close game, and your champ will be in your inventory.
Mmm... once you place the crystal and release your finger from it, that should be it.
Once you place the crystal in the spinner, your entire alliance sees what you got in alliance chat. That is the moment when the servers determine your reward, and I've tested this myself when I first heard about it.