Rogue Master MJ
Well-Known Member
Same game you are... But the return on selling GB's isn't so great....You do so using the sell button in your city. What game are you playing?
Same game you are... But the return on selling GB's isn't so great....You do so using the sell button in your city. What game are you playing?
Still doesn't mean you don't sell them when it's time for them to go. The OP used the exact correct terminology, which you clearly didn't understand. Now you know. You sell buildings whether it's to the AD, or not. You must not predate the AD. I suspect it's the noobs who equate selling only with the AD.Same game you are... But the return on selling GB's isn't so great....
Actually, your overall relic generation rate is outside of the 99% confidence interval. I believe another member commented earlier about regularly experiencing fewer total relics than expected. Your data statistically supports that there may be an issue with the relic generator calculation. Certainly a red flag, and probably warrants additional investigation.I looked at the ToR data (expected - actual)
L57E 30% total / 56.18% common - 29.26% uncommon (gold) - 14.56% rare (jade)
L57A 27.0% total / 52% common - 29.3% uncommon (gold) - 18.7% rare (jade)
498 relics/1842 encounters (553 relics needed for 30%, 55 more)
L62E 30.75% total / 55.1% common - 29.54% uncommon (gold) -15.36% rare (jade)
L62A 27.6% total/ 52.1% common - 31.6% uncommon (gold) - 16.3% rare (jade)
503 relics/1824 encounters (561 relics needed for 30.75%, 58 more)
So, nerfed? no, I don't think so. I did notice the overall total relics were 3% less for both players (in bold). Under each, I listed what was needed, how many more relics, to have the correct % total relics and I would say it is possible to make up the shortage over time.
? you gonna make a conclusion like that based on a sample size of 2000? increditable.Actually, your overall relic generation rate is outside of the 99% confidence interval. I believe another member commented earlier about regularly experiencing fewer total relics than expected. Your data statistically supports that there may be an issue with the relic generator calculation. Certainly a red flag, and probably warrants additional investigation.
Not really a conclusion, that’s just what the statistics say about the data. Otherwise, yes, the sample size is plenty sufficient and actually creates a reasonably narrow width of confidence.? you gonna make a conclusion like that based on a sample size of 2000? increditable.
so I guess in foe standard deviation doesn't exist. good to knowNot really a conclusion, that’s just what the statistics say about the data. Otherwise, yes, the sample size is plenty sufficient and actually creates a reasonably narrow width of confidence.
One player's stats are within 3% of expected and you consider it a red flag and warranting additional investigation? Seriously? You do realize that there are literally thousands of players, right? Your "insight" is like taking a thousand people and having them take turns rolling the same die 2000 times, then picking one person at random and saying the die is faulty based on their stats. Actually, that scenario would be more valid than your conclusion here. You fancy yourself an expert on stats, but you make the most ridiculous statements about them.Actually, your overall relic generation rate is outside of the 99% confidence interval. I believe another member commented earlier about regularly experiencing fewer total relics than expected. Your data statistically supports that there may be an issue with the relic generator calculation. Certainly a red flag, and probably warrants additional investigation.
I would recommend googling “how many standard deviations in a 99% confidence interval” to see what is wrong about your statement.so I guess in foe standard deviation doesn't exist. good to know
Where to start with this gem?One player's stats are within 3% of expected and you consider it a red flag and warranting additional investigation? Seriously? You do realize that there are literally thousands of players, right? Your "insight" is like taking a thousand people and having them take turns rolling the same die 2000 times, then picking one person at random and saying the die is faulty based on their stats. Actually, that scenario would be more valid than your conclusion here. You fancy yourself an expert on stats, but you make the most ridiculous statements about them.
That 3% you refer to is approximately a 10% departure from the expected value. I consider it a red flag when somebody presents data that generates a statistical variance beyond accepted norms. Further, that’s how data analysis works. Once an anomaly has been detected, further evaluations are done to confirm. The nice person that posted it, even went so far as to provide two examples of an anomalous distribution rate.One player's stats are within 3% of expected and you consider it a red flag and warranting additional investigation? Seriously?
This is literally irrelevant. The trials are independent and the results of one GE encounter have no effect on any other GE encounter for anyone in the game. We aren’t trying to represent the player demographics in some type of population study. The entirety of statistical sampling is taking a varying amount of data and being able to narrow the confidence intervals as the sample grows. Smaller samples lead to wider, less specific intervals, but intervals nonetheless. If the outcomes still don’t fit in these wide intervals, it is still anomalous.You do realize that there are literally thousands of players, right?
Not even close. All you are doing here is increasing the sample size to 2 million. The process wouldn’t change. The only thing that would differ is the confidence interval would shrink even further in width. If the result fell outside of that interval, it is still anomalous. Also, this isn’t insight. It’s basic statistics that you can go plug into a calculator right now.Your "insight" is like taking a thousand people and having them take turns rolling the same die 2000 times, then picking one person at random and saying the die is faulty based on their stats. Actually, that scenario would be more valid than your conclusion here.
Have never called myself an expert, but sure I’ll confess to understanding high school statistics. It is Interesting that you don’t seem to find 1800 trials (x2) to be sufficient for saying something looks off, but it is sufficient to say it is working just fine.You fancy yourself an expert on stats, but you make the most ridiculous statements about them.
Please show me where I drew any conclusion based on this data. My point is that the data is insufficient to indicate anything. One player out of tens of thousands. 3600+ occurrences out of tens of millions. Two ToRs out of over 255,000. Yeah, go ahead and call for an investigation.Have never called myself an expert, but sure I’ll confess to understanding high school statistics. It is Interesting that you don’t seem to find 1800 trials (x2) to be sufficient for saying something looks off, but it is sufficient to say it is working just fine.
190,766*1%*2.5=4770. so what you saying again?I would recommend googling “how many standard deviations in a 99% confidence interval” to see what is wrong about your statement.
getting lots of junk can be good at times selling them in the auction and Can get good stuff by bidding on itemsI understand that the drop rate changed to be more accurate...
But what's with the rewards? I have gotten so many Faces of the Ancient and Ritual Flames that it makes me want to demolish my Temple of Relics just to free up more space in my city.
I would rather get a random number of forge points from 2 to 5 in number than have to get any more of those
crappy buildings.
Anyone else feel this way?
The probabilities experienced by any single player do not impact any other player. It is irrelevant how many ToRs there are or how many players there are. A 30% distribution model looks the same. We could be pulling balls from a bucket of 100 green and red balls. If after about 1500 draws, we are getting 27% red, it becomes unlikely that there are 30 or more red balls in the bucket (or less than 24). If we had only performed 1000 draws at 27%, the confidence interval would include 30% and we could say it is operating within expectations. This is precisely what statistics does - small sample creates data that describes the overall probabilities. The larger the sample, the smaller of an interval that describes the data.My point is that the data is insufficient to indicate anything. One player out of tens of thousands. 3600+ occurrences out of tens of millions. Two ToRs out of over 255,000.
I don’t know what you are showing here. Number of players, exception rate, standard deviations? If what you mean to say is that 99% isn’t 100%, then yes that is true. It is possible that the one person that has tracked and posted their data just so happened to be outside of the 99% confidence level. That alone would probably warrant a closer look. This person then posted a second dataset that would also fall out of the 99% confidence level. The chances of each are actually less than 1%, but we’ll round up and say it was exactly 1%. That leaves the probability of this person‘s data being exceptional at .0001. That again sure would seem to warrant a closer look.190,766*1%*2.5=4770. so what you saying again?
Yeah, thanks for that nugget, Sherlock.The probabilities experienced by any single player do not impact any other player.
Sure it's relevant. You're saying that you think there's a flaw in the all the ToRs based on these two. If they were the only two, you might be right. But the more there are, the higher the probability that some will perform outside your narrow expectations. So the fact that there are over 255,000 ToRs is relevant. Two ToRs is one thousandth of one percent (.00001) of all ToRs. You can't draw any conclusions based on that.It is irrelevant how many ToRs there are or how many players there are.
No. However, when people make misleading or incorrect statements, some of us feel obliged to call them out so others aren't negatively impacted.Does every single thread have to turn into an argument?!