Extremely low chance on drawing 5* event heroes

It is poor understanding of science. Occam’s razor isn’t a rule. Falsifiability is more important for this case. And as we can’t check any hypothesises we can’t use any of them.


I have to say that I like the random factor. There are games where you don’t have the random factor and the top teams are all the same, there’s no variation. Having the random factor ensures not everyone will have the same teams.

1 Like

yet we see Wu Kong in all Titan teams regardless of titan color…

1 Like

Wu Kong and dead heroes goes together on high level titans if you don’t throw time stops at titan’s face

1 Like

Sorry dude. You are simply incorrect. And I’m surprised that you seem so adamant on this. Nobody is claiming that the findings of the efforts are 100% accurate. My standards are indeed fairly low on this as it is a free mobile app and not a definition of auto safety regulations or some such.

However you can’t actually believe that there is zero value in the data collected. If I even look at just 100 epic pulls and receive say 5 5* heroes, I can be 100% certain that the drop rate of 5* heroes is under 100%. There. I’ve learned something. I can be reasonably sure that the drop rate is under 50%. It is true degree of certainty lowers the tighter I try to draw the conclusion, but this fact does not invalidate the learning.

As more and more data is collected, we can start to zero in on an acceptably accurate finding of the true drop rate. While it is certainly possible that the drop rate changes or has other factors at play, what the data describe is not these changes nor the other factors. We can only make a broad prediction based on the factors in play during the time the data were collected.

If you choose to believe this is based on faulty assumptions, that’s your choice and is certainly defensible. However it is really only defensible in the same way that you could say that our understanding of physics is based on assumptions like ‘The laws of physics are universal and unchanging’

As to your statement of ‘guess are…guesses’ C’mon man. That’s crazy. Guesses include both wild stabs at truth AND informed probability based assertions. We guess in our lives all the time. Literally every thing you do in your life rests on evaluating probability and going with your best guess. From stepping up on a curb, to choosing your next job. It’s a LOT better to look at the available data in making these guesses and not just throw up your hands and saw ‘might as well just throw a dart, I can’t know for sure anyway.’

And dude, you CAN reliably put numbers to the processes as long as you are OK with a degree of uncertainty in those numbers. While it is definitely wrong to say we know the drop rate to a specific percent, I think you’re really stretching to say we don’t know within, say, 10%. Honestly I have not done the math in a long time, but it would surprise me that 1000 pulls is not a significant enough sample size to determine within 1-2% what the rate is, but would be happy to be shown to be wrong in terms of how reliable that sample size is. What I find absurd is that you would argue that you can have no confidence in any assertion as to the probability.

Finally, good job stacking your super high horse on top of other high horses. You look um…really cool up there.


See this thread:

Sample size was over 1000, which allows a reasonable calculation of confidence intervals under various assumptions about the underlying statistical processes.

The data were collected across many player. Were some lying? Perhaps, but there’s no incentive to do so.

I believe SG to be under-resourced because so many things are released with bugs, and we know approximately the staff size, and we also know they’re working on at least one new title. This isn’t to say that someone hasn’t written a super-duper revenue maximization algorithm—I’ve seen the code that lets one major airplane do that, though, and it’s not simple stuff.

The answer to insufficient data is not to throw up one’s hands and walk away. It’s to collect more data, preferably in a way that allows refutation of a null hypothesis. A natural null hypothesis is that the odds of pulling the HotM are the same, small value for each pull. Let’s see if we can group-collect enough data to test that hypothesis.


Sure thing man, but also add:

And this is the best data available so we would be fools to ignore it, despite the inherent uncertainty.

1 Like

I went through some of the topics and you are THE village… fanboy hands down. Completely insulated against any reasonable argument. Passive aggressive too. I am sorry I didn’t know who I was interacting with. :wink:

What reverse phrasing? Building times are dragged out for no good reason. Its a lazy design and nothing else. Instead of content to keep us occupied we get timers. I know it has a certain OCD appeal to but not to me. Your rationalizations attempts are very sweet… and quite typical for a person who tries to justify and defends their choices. Its silly but it is your right to do as you please.

Its a success of a lazy and in part abusive (gambling bit) design. The game is repetitive as f…, lacks new content, its the design that forces interaction rather than genuine player interest, its parades as completely random but I doubt its purely random. Its success is hooking people in by pretending to be sth else for the first weeks of playing. Then comes the progression paywall and lack of end game. Some people enjoy being part of such profit oriented system because they were lucky or have the cash and will try to justify their choice. Some see the big picture for what it is. Some buy DLCs, lootboxes and so on and support abusive business practices and some don’t.

:joy::joy: No, we didn’t. You have conjured sth and pass it as a fact (in plural). I like grind, I don’t like gambling and I don’t like abusive mechanics. In this game grind is designed to give players enough time to make the “right decision” and spend money. Been gaming long enough to and see this game for what it is. I actually find this forum more fun than the game itself. Might stick around.

No buddy. You have a degree in behavioral psychology so you should know that introducing gambling mechanics to games can be very damaging to the young cause it normalizes such behavior. There is a known relationship between access and prevalence of associated disorders. Gambling fair and ethical - you must be having a laugh? There is no such thing as ethical gambling.

Its an interesting way of looking at games but it lacks the fun factor imho. I like how you are concerned for me and my time. Thanks its very sweet <3.

Again, argument from ignorance. I am very happy. Are you? This forum seems to be important part of your life.


No, Occam’s Razor isn’t a ‘rule’. It is, however, a very useful principle that is on point in this case. Perhaps more importantly, you are thoroughly wrong about our hypotheses not being falsifiable. I’ve already outlined what we would need to reject our primary hypothesis. If players choose not to collect that data, perhaps it’s because they don’t really doubt the hypothesis.

1 Like

You don’t know what you are talking about. :joy::joy: Consider this a professional opinion.

C’mon guys, my intent wasn’t to make this thread closed, too easy this way.

I really hope someone of the staff come here an answer to us, even a STFU.

I’m a simple guy, no need big stuff.

I am not but we will get to this later. I am also correct I will try to explain it better why. Finally I have never said that people claim this. This is an example of straw-man fallacy and reduction ad absurdum in one.

Sure you learn some useful information from this exercise but this information cannot reasonably quantified. People here assume that because they have what they think is a large enough sample it must offer some significance. Its fair logic but sadly its flawed.

Representative sampling:

  • sample should be random (and it isn’t cause most people don’t bother with forums they just want to play the game, there is an over representation of seasoned gamers and their habits differ from general gamer population)
  • sample should be sufficient in size to represent the population it studies within acceptable error (for basic system like an opinion poll we have 3% confidence on a 50/50 opinion for sample size of 1000 with a 100% return - but drop rate is not an binary opinion so it more complex system).

So we should have representative sample size for EACH state or drop value to estimate it reliably from player outcomes which is the end state. We don’t have this and we are likely not to have it in the future. Drop rates are adjusted and will be adjusted cause this is how the in game economy/revenue is regulated. There will be optimum drop range for the economy/revenue but we don’t know how narrow it is. We don’t know if selected player populations are not experiencing increased or reduced drop rates. We will likely never find out cause it would cause uproar if confirmed. Maybe on special occasions they increase/reduce drop rates? Maybe drop rates are tied to spending? These are valid questions we cannot be answered by the players through end outcome data grind.

Sure its better to know that drops are likely to be below certain percentage that not to know it but because this state can be changed without us knowing its a very crude estimate with a lot of uncertainty attached to it. For some it will be good enough and they will take it at the face value and its fine but what I take issue with is people say we KNOW. We don’t. There is a high risk associated with taking highly uncertain results as a accurate. The error could be easily multiple of the value ie drop of 5* from the gate is 3 +/- 5%. We don’t even know with certainty the population size to calculate a reasonable sample size… I mean for a million strong player base, confidence level of 95%, and error margin of 1% (as our estimate is as low as 3-5%) the desired sample size is ca 10k for a simple system.

Thanks. The view is magnificent.

1 Like

I would say the answer is to eliminate the source of error and reduce uncertainty. But we can’t do that without devs help as they have the answers to some fundamental questions. What about the HOTM drop rate being adjusted, different during the event, different afterwards? How to account for this?

1 Like

The super frustrating thing about the RNG and not knowing what the rates are is that there is no way at all to hurry progress in AMs or heroes at all.

Everything that leads to gems or tokens or AMs has a frequency absolutely controlled by the game ( flags, titan spawns, quests, chest etc). That is why bad luck hurts so much , there is nothing you can do about it except wait until the next time , or spend more.

I think fewer would complain about not knowing the percentages if they could grind them out on some activity that was not limited , but took time and resulted in gems.People would play huge number of hours to get the result ( and the complaints on here would be about not being able to compete with get-a -life gamers , not big spenders :slight_smile: ).

Anyway it would be nice if SG introduced some small elements like that into the game, here is some example off the top of my head

  • Every million ore ( or whatever number) mined you get a bonus roll ’ Your mine produced 50-100 gemstones and something else 'mine related ’ this month.

  • Your watchtower caught an enemy patrol ( get some troops).

  • Your Tc found and exceptional hero, very small chance independent of tc level hero token ( kind of like ‘great people’ production in Civ).

  • Introduce some difficult ( probably need battle items) ZERO flag provinces with chances of 3/4 * AMs and\or tokens at an extremely low rate ( and no bloody rope or potent herbs please !!!)

For anyone that thinks SG are small ( ok they are not large …) and don’t have resource to do some of the things we as customers ask the following,from February, is an interesting read

I’ll quote just one thing.

‘Despite such a large funding round, Soininen said the he remains a “big believer in small teams and talented individuals, as our company name suggests”.’

And for anyone too full of suggestions and dev skills -


I’m not sure about this. Let’s test it: try to come up with a reasonable argument. I’ll wait. :grin:

So you don’t actually mean that building times have been “dragged out”, which implies some nefarious tweaking by the devs; you just mean that building times are longer than you would like. That’s perfectly fine, of course - de gustibus non est disputandum - but it’s also impatient and entitled. Not an ‘instant gratification’ type, huh?

That bitter little rant certainly makes it sound like you dislike the game. It might be nice to have the forum graced by the voice of someone who flat out dislikes E&P, just for the contrast. I know that I spend time playing and discussing it because I’m very fond of it; I get the impression that’s also the case for other contributors. Seems a bit masochistic to invest your time in something you hate, but hey - whatever floats your boat. Some people like clothes pegs, don’t they?

There’s no paywall. And no end game, because the game isn’t designed to end.

Just wrong. In this game, grinding allows you to assemble a team that can be competitive with players that spend. That’s unusual and it’s a good thing.

When you say foolish things like that then you signal that you’re not rational. It’s good sense to recognise the dangers of gambling, the high susceptibility of some to its charms, and the possibility of an addictive relationship with the pastime developing. But to claim there’s no such thing as ethical gambling is puritanical rubbish. Gambling is harmless fun for millions.

I am as happy as a box of fluffy ducks. :hatching_chick::hatching_chick::hatching_chick: And while this forum isn’t exactly an important part of my life, I do like discussing the game with others who enjoy it. And if a hater or two also want to participate, then what the heck: all are welcome.


I agree with most of what you have presented.

Yes if we were conducting a study worthy of publication, we would need both a randomized data sample and a larger sample. However given the limited funding for this effort, some concessions had to be made. As a result the findings must be viewed with some healthy skepticism. Happy to concede that ‘we know that’ or ‘we have found that’ are inappropriate ways to present the outcomes. That said, I’m glad to accept that you concede that the current analysis of the limited data set has some utility.

Thanks for your cursory walk-through of determining the necessary sample size. Although your larger point seems to be that even if we had the entirety of the data set, we still wouldn’t know enough to ask the right questions since SG can change the rules at their whim without warning or announcement. But appreciated nonetheless since as I mentioned my math on this front is rusty and I’m lazy as well. Due to my laziness I will assume you have done the math correctly. Due to my confirmation bias, I’ll likely quickly change my tune if someone throws up conflicting math that leans towards my current beliefs :wink:

Just a quick final note on the fallacies you invoke. This deliberate attempt to embarrass ends up being a misapplication of some basic book learnin’ without acknowledging the clearly intended meaning behind my statement. My ‘100%’ remark was at worst hyperbole and more fairly simply a colloquialism. Careful when you use this stuff. You try to sound smart but look deliberately pedantic. In the end it dilutes the effectiveness when these fallacies can be used accurately. Of course if you did it just to get my goat, touche :wink:

1 Like

hahaha in other words you agree with @Kerridoc that the appropriate response is to throw up your hands?

1 Like

I agree …it’s legitime that developers should earn money but this is definitely greedy .I have spend around 300 euro and I am after nearly 1 year stacked with 3x5* heroes! Development of sg should rethink their sell strategy or people will end purchasing anything as they realise beeing ripped off. IT CAN’T be ALWAYS EXPLAINED it’s random then within 1 year getting always the same low rated heroes isn’t random anymore and not funny just frustrating and a step a way to deinstall .!! A request to SG development keep the game fair and giving more chance to purchase 5*heroes on recruitment or summon ;TY

1 Like

There was already more than one presented on the usability of end state data. You were not interested. Besides this comment was referring to your rebuttals in other threads…

Yes. Too long for me and its a pretty common opinion since the collective moaning resulted with VIP passes being added. I am quite patient when I see a point. Here there is no point other than a lazy design. That’s my opinion. I am perfectly fine with the though that there are people who wake up, fire up the game and think “yes! only 6 days left! efing delighted. I achieved so much already.”. To me its the simplest and laziest solution there is. But just so its not just criticism, some alternatives:

  • common enough strategy is to include other resources in the mix
  • bind upgrades to quest
  • bind upgrades to story line progress
    all more interesting and entertaining (imho) than the excitement of… waiting.

I might be ranting a bit but at the same time its my diagnosis. Instead of focusing so much on sounding eloquent maybe address the criticism for once?

  1. Is it repetitive? Yes it is. Grind of provinces fo AM. Raids with more or less the same meta teams. Titans are repetitive as well. Rare ones are some interesting direction but because they made slot machine out of them as well, they kind have a ambivalent feel to the atm. Is it a good thing? Its not in a mobile app imho.
  2. Lack of new content? Surely. Generic quests, recycled events and one story.
  3. Randomness thing is just my feeling. I don’t believe it is random. I think they constantly play with the rates, odds and so on to shape the outcomes to their desire.

There is one. Its a sort of a “soft paywall” and low odds are the gatekeepers. The game is designed to incentivise spending hence the daily offers, splash screens, sales and so on. You have alredy admitted this above so I am not sure why are you backtracking now? Obviously lack of end game in your mind is a good thing right? Good for whom? In principle grind is ok but too much grind to squeeze cash out is not.

Fair enough, this was a bit too broad. I was referring to certain types of gambling ie targeting minors not all games of chance. Obviously you know this but choose the easy score instead.

I am delighted to learn this, that makes at least two of us. Lets hope it stays this way. Btw did you know that ducks despite their fluffiness are quite violent birds? Just a digression.

As my final remark I would like to propose we tone down the pleasantness. Lets keep it civil.

I am just ana*l this way. I can’t help it. I get a bit social awkward label because of this. And honestly I swear to G I don’t mean to embarrass anyone. Its more of rhetoric contest at best. If you or anyone else felt embarrassed my apologies. Peace.


Our? Primary? Could you clarify this?

Do you collect data? Can I see it?

In the era of big data is really strange to deny complex algorithms by referring to Occam’s razor.

1 Like

Cookie Settings