Zero 5* in 29 pulls?

Siv75

ThaRoadWarrior said:

The "best" thing to do (in my opinion) to generally enjoy the game if you aren't a super competitive person is to try and get everyone on roster, and settle in for the long-haul of grinding the feeder system for the upper-tier characters you want. Whether you choose to commit to latest or classic is then up to you, but I would say that dumping all your resources into a single bottomless pit is going to give you the most consistent pull rate experience.

This is good advise. To conclude, I've shared the developers response above. The official stand is there is no seeding and each pull resets ... so is not dependent upon the previous pull. That is quite conclusive a response. 

It doesn't change the feedback I've given. Even if the mechanism is perfect, giving an odds ratio like they do give a wrong impression to someone who has hit 0 pulls in samples larger than the max odds ratio by more than 300%. Yes, it's possible that one can be that unlucky in any aspect of life. However this being a 3rd time with special stores, I've become less interested in CPs. I don't think that's ever going to change. I've not had such issues with LL stores so probably will be more keen to collect those, and cancel my VIP and chill with a lower pace. 

Srheer0

giving an odds ratio
This is something I don't like about the game. Look at the odds of pulling a standard token, elite token, heroic token, support token or one of the special stores with a boosted "rate" on a single 4star.  

In heroic token they list the odds, but don't mention the lowest odd. They should display all the odds and be transparent about everything. 

Bad

Srheer0 said:

giving an odds ratio
This is something I don't like about the game. Look at the odds of pulling a standard token, elite token, heroic token, support token or one of the special stores with a boosted "rate" on a single 4star.  

In heroic token they list the odds, but don't mention the lowest odd. They should display all the odds and be transparent about everything. 

This is because this game has this strange odds system, but it's not because it's not transparent.
Look at the falcon store. It's says there is a 1~14 to a 4*, and 1~4 to a 3* in which falcon has 12x odds. After you finish calculations, the remaining amount will be 2*( possibly a 70% but I expect to be wrong).
Would  it be simpler with a % system? Of course. But the thing is that because of the peculiarities of their rng system they are forced to put it like it is.
As a side note, this response collides with another one recent rumour saying that CS gave a lot of covers for a similar bad luck case, and personally I trust this one a lot more.

entrailbucket

Vhailorx said:

entrailbucket said:

Vhailorx said:

ThaRoadWarrior said:

That search is less about PRNG and more about when it is performed and how often, ie does every player get their RNG seeded one time, and then they are just redeeming along that string of pulls for the rest of their MPQ days, or was each pull a dip into the RNG uniquely. People used to speak on that topic with some authority but I haven’t located the source of that knowledge so I don’t really want to comment on it one way or the other.

It's definitely closer to the former (seeded once and a set stream forever) than the latter (each pull is uniquely generated at the time of pull).  Like most loot box games there were some early controversies over exploiting client-side RNG calculations, but things have been pretty stable for some time now.

RNG is all calculated server side and is seeded once per store and each player just pulls against that stream of predetermined outcomes out to forever. (The per-store seeding is why daredevil points out that the specialty stores are more like gambling.) I seem to recall  that the seeds for long-lssting stores (ie the latest store) have been reset once or twice over the past 4 years, but I can't find any documentation of that on the forums (nor would I expect to as it would be an invisible change to most users).

I do not know if anyone has tested the periodically repeating LT stores (e.g. fan favorites) to see if they reseed every year.

I don't think "seed" is the word you're looking for...that has a specific meaning when you're talking about random number generators.  You're just saying the outcomes are predetermined.

Again we're very close to discussing an exploit here, but if I roll a die 20 times and record each result in order, then read them to you one at a time, those individual rolls were predetermined but they are still random.

My understanding of RNG programs is that they use an arbitrary "seed" value (eg a player iD number) as the starting point for an algorithm that generates randomized numbers.  (And as mentioned earlier in the thread, if the algorithm isn't good enough, the generated numbers will repeat too quickly and are predictable. But I think that is less of a concern nowadays than it was, say, 15 years ago as the programs are much better). So the baaic principle is a (1) fixed, arbitrary value that differentiates player A from player B, and (2) an algorithmic black box that takes (1) as input and produces an infinite series outputs that are both random and also unique for each input  if I am wing about how these RNGs with I would love to be corrected. I am no expert.

So "seed" as I am using it means putting a fixed value into the algorithm. When a new mpq account is created, the latest store is seeded for that account and all LT pile from n=1 out to infinity are fixed no matter when they are used (though the character pool in LTs changes over time). And if the LT store is ever "reseeded" for any reason, then an entirely new set of pulls, starting from the next pull, would be generated.

Not trying to discuss any exploits.

Sorta.  The RNG function takes a seed value.  If you pass it the same value forever, it'll spit out the same sequence forever.  That's not very random and it can cause problems, because if someone can guess the seed value then they know the entire sequence.   

The way these work now is that they're fed a random, different seed value every time they're called.  That seed value is some very large random-ish number, like a hash of electronic noise collected from a hard drive or some other source of mathematical entropy.  These things aren't just available now, they're the default way of doing things.

Someone might choose to do things the way you describe for some reason, but they'd have to choose that way.

HoundofShadow

It's 4x odds for 3* Falcon HP store. My understanding of 4x odd is simply 46 + 4 = 50

Instead of 1/5*1/47 odds or 0.425% chance of drawing 3* falcon, it's 1/5*4/50 or = 1.6% chance.

Blackstone

My record for pulling with no 5* is 62.
But I kept pulling and drop rate got much better. Like a 5* every 3-4 that just kept coming.

It's how I learned to only pull when I had a a lot of pulls available.

entrailbucket

I've always found it to be a little weird that people choose a game with RNG-based rewards and then do everything they can to avoid the vagaries of RNG. 

Like, OP here makes a great point!  The game is extremely unfair.  You can work really hard, earn a bunch of rewards, and then cash in your tokens and get nothing.  You can spend real money and get a big pile of nothing.  It's all random.

If you're a gambler, it's really fun to bet this way, and losing makes winning more exciting.  Some of us like that feeling.  But if you're not a gambler, or if you're really concerned with fairness...why choose a game that's made this way, out of a universe with tens of thousands of other available, fun games? 

They're not lying to you or misrepresenting what this is -- some of the very first rewards you get are tokens that give out random stuff.  I guess I don't understand why you'd stick with this game given so many other alternatives that are so much more fair.

fight4thedream

**Mod note: Removed post of Customer Service ticket conversation. Please be aware that it is a violation of Rule 8 to post such content on the forum. 

bbigler

I’m a software developer as well, and any good random number generator function will seed off of another constantly changing number, like microseconds. Thus, reaching true randomness.  It can’t be predicted. 

But when you do a quick pull of 50 tokens, (for the sake of processing speed), it’s not calculating the next random cover in a fraction of a second, but instead just loading the graphic for the next predetermined cover.  So, at some point in time before pulling, the random sequence was generated. 

As another poster stated, this is likely done on the server (perhaps during account creation).  But it makes me wonder how far out the sequence is calculated.  It must be finite, which means it must re-seed and regenerate periodically.  After 1000 pulls? 10,000 pulls? 

Perhaps it generates a random sequence of 100,000 pulls and then checks if it’s close to 15%.  If it’s not, then it recalculates a new sequence.  I would have programmed it that way and it would be fair to everyone.  

entrailbucket

bbigler said:

I’m a software developer as well, and any good random number generator function will seed off of another constantly changing number, like microseconds. Thus, reaching true randomness.  It can’t be predicted. 

But when you do a quick pull of 50 tokens, (for the sake of processing speed), it’s not calculating the next random cover in a fraction of a second, but instead just loading the graphic for the next predetermined cover.  So, at some point in time before pulling, the random sequence was generated. 

As another poster stated, this is likely done on the server (perhaps during account creation).  But it makes me wonder how far out the sequence is calculated.  It must be finite, which means it must re-seed and regenerate periodically.  After 100 pulls? 1000 pulls? 10,000 pulls? 

Perhaps it generates a random sequence of 100,000 pulls and then checks if it’s close to 15%.  If it’s not, then it recalculates a new sequence.  I would have programmed it that way and it would be fair to everyone.  

We're all pretty much saying the same thing here, except for the people who are worried that random isn't actually random and that it's skewed somehow.

bbigler

entrailbucket said:

bbigler said:

I’m a software developer as well, and any good random number generator function will seed off of another constantly changing number, like microseconds. Thus, reaching true randomness.  It can’t be predicted. 

But when you do a quick pull of 50 tokens, (for the sake of processing speed), it’s not calculating the next random cover in a fraction of a second, but instead just loading the graphic for the next predetermined cover.  So, at some point in time before pulling, the random sequence was generated. 

As another poster stated, this is likely done on the server (perhaps during account creation).  But it makes me wonder how far out the sequence is calculated.  It must be finite, which means it must re-seed and regenerate periodically.  After 100 pulls? 1000 pulls? 10,000 pulls? 

Perhaps it generates a random sequence of 100,000 pulls and then checks if it’s close to 15%.  If it’s not, then it recalculates a new sequence.  I would have programmed it that way and it would be fair to everyone.  

We're all pretty much saying the same thing here, except for the people who are worried that random isn't actually random and that it's skewed somehow.

Totally agree.  It is random.  It is fair.  You aren’t entitled to a 15% rate even after 100 pulls. 

There’s nothing wrong with the game, but there may be something wrong with one’s perspective of the game.  

ThaRoadWarrior

If everyone’s pull destiny was randomly generated out to infinity all at once, and the sequence could be known by the dev team, one could perhaps not unreasonably lobby to have any atypical gaps artificially smoothed out. But if each pull is generated randomly at the time of the pull that is not as realistic a request.

bbigler

ThaRoadWarrior said:

If everyone’s pull destiny was randomly generated out to infinity all at once, and the sequence could be known by the dev team, one could perhaps not unreasonably lobby to have any atypical gaps artificially smoothed out. But if each pull is generated randomly at the time of the pull that is not as realistic a request.

That would create a customer service nightmare with everyone contacting them because they didn’t get a 5* after 20 pulls.  What if someone gets 3 x 5*s after 10 pulls? Should D3 take some back to even out their odds?  They logistically can’t (and shouldn’t) fix everyone’s pull rate for such small amounts.  If you pulled 500 and got <10% then you could make a case.  But I’ve never heard of that happening to anyone.  

I used to work in customer service and you don’t want to create more work for yourself.  When someone angry calls, the best way to handle it is by acknowledging the reason they’re angry and then investigating what happened.  Usually, the solution is to educate the caller on how the system works instead of giving them something for free. If the company made a mistake, then you fix it, but otherwise you educate your customer.  

ThaRoadWarrior

My suggestion was more like “if all these number lines exist, just run a smoothing algorithm on them to avoid any wide gaps or wild clumps” before people have to complain

bbigler

ThaRoadWarrior said:

My suggestion was more like “if all these number lines exist, just run a smoothing algorithm on them to avoid any wide gaps or wild clumps”

That’s not a bad idea.  It may cut down the complaints…….unless people start complaining because they didn’t get a 5* after 10 pulls. 

I will say that I understand the frustration of the OP because it takes a lot of work just to get 1 pull.  So if your accumulation of hard work nets very little, you get angry.  Thus, my earlier comment about proper expectations.  

ThaRoadWarrior

It could even be something they’d be comfortable publishing at that point, like they did when they added the bonus shards counter that gets more and more likely to hit up to I think 40 0
pulls where you are guaranteed to get some bonus shards? That feels like the right number, I’d have to search that thread to confirm but I do recall we were told.

ThaRoadWarrior

https://forums.d3go.com/discussion/82009/r199-new-feature-bonus-shard-forced-timer

bbigler

I haven’t read up on this, but I think the gambler addiction comes from getting a string of wins (making you very happy), so you continue to gamble hoping for another string of wins, pushing you through a string of losses.  As long as those random clumps of wins keep happening, you keep gambling even if your sum total is negative.  So perhaps the devs don’t want to smooth out the pull rates.  

I’ve noticed in my small sample of 224 Latest pulls (since restarting) that I get clumps and droughts instead of regular periodic 5* covers.  I’ve had a series of 20+ pull droughts as well as tight clumps of 5*s, such as 4 in a row.  

ThaRoadWarrior

It’s generally pretty clumpy in my experience as well. It’s not hyper common to get droughts above the teens but it’s not unheard of either.

entrailbucket

I have heard stories of people getting compensation for very low rates after they made a very large number of pulls, but I've never seen any proof that it actually happened.