I have just taken my GCSE in statistics so the infomation is fresh in my mind but if I make any mistakes please correct me. Some of this may be a bit confusing (hey, I don't fully understand it).

 

 

 

I think that we can assume that the distribution for rare drops is normal as it would be too much effort for Jagex to make it any other way (unless they are cruel and made it negatively skewed just to spite us). This means that the mean will be at the center of the bell-shaped curve. We could get a pretty accurate estimation of the mean by asking everyone who got a rare drop how many kills it took them. Whilst it is true that many people wouldn't know or would give wrong infomation an experiment could be conducted (if everyone had way too much time on their hands).

 

 

 

The definition of a normal distribution is that 67% of results lie within one standard deviation of the mean, 95% within two standard deviations and 99.9% withing three. This means that if we had a big table of results for the number of kills it took to get a drop, we could work out the mean, find out the range of values between which 67% of the results lie and devide this figure by two to give us the standard deviation. I hate this part of statistics but if we had this data, could we not work out the drop rate?

 

 

 

If we factor in the ring of wealth, I believe that this would give the results a slightly positive skew so the experiment would have to be conducted without rings of wealth so as to make the results accurate.

 

 

 

There is also the issue of clustering. For example, if you had 10 numbers between 1 and 100 you may get these results:

 

 

 

4

 

23

 

26

 

28

 

75

 

91

 

75

 

81

 

1

 

46

 

 

 

As you can see, we get a cluster of three numbers (23, 26 and 28) even though the numbers were random. This can explain why some people are luckier than others. The luck people may cluster in the 90s (for this example) whereas the unlucky ones may cluster in the 20s.

 

 

 

These are just my thoughts but can someone please give me feedback on them?

I have played rs since 2001, lvled an 88 plate pure on classic, a 109 rsc-rs2 main a 113 combat main, and a pure with 92 range..

 

I have never got a dragon drop, i went to kbd with my mate arch we killed him about 12 times of which i got the kill 9 times, arch got d med and left half, i got arrows and yew logs.

 

This game [bleep]ing hates me.

 

 

 

It'd probably (I say probably because I know jack about runescript) take more work to program every spawn for every monster to keep track of how many have spawned so it can "release" the item than to just give the monster itself (meaning that monster "code" for every monster, regardless of spawn point) a 1/3000 chance of dropping it.

 

 

 

 

i think it can't be that way because: how could the Ring of Wealth work?

Nickelliston I cant see why you would say the distribution is normal.

 

 

 

For any given rare drop, from any single monster you will, on a specific kill, either get it or not get it.

 

 

 

This is a Bernouli distribution, so if you are repeatadly killing this same monster over and over again what you have is a sum of independant Bernouli distrubtions, which is the binomial distribution, where x is the total number of the drop you are looking for, n is the number of the monster you have killed and p is the probability.

 

 

 

Now based on the two wheel apprach to rare drops N64jive mentioned, if you say that P(A) (A = 0 , no rare drop. 1 , rare drop. this is simplistic, but completly enumerates all possible outcomes and covers our purposes) is the probaility of getting any rare drop on the first wheel, and the second wheel giving the rare drop in question is P(B) (B = 0 some other rare drop. 1 the rare drop in question)

 

 

 

So the probability of getting the drop in question is P(A)P(B) (normally this would require independance, which we dont have, but the definitions in the model give this result to use anyway)

 

 

 

Now choosing P(A)P(B) to be something tiny what you find is that the odds of not getting the rare drop in question after killing 1/P(A)P(B) of the monsters (n = 1/P(A)P(B)) which is where you get the mean of binomial distribution to be 1 (mean = np) the probability of not getting it tends to about 0.368.

 

 

 

There is a binomial distribution probability calculator here if anyone is interested http://www.stat.tamu.edu/~west/applets/ ... ldemo.html

Ahh it's Qeltar! I trust your figures, as I have seen alot of good from you (even though you do argue alot :P).

 

 

 

Thanks for posting this, most people are totaly obvivous to the luck drop rates and frankly it gets quite annoying! I think the idea of beeing a lucky player or an unlucky player is fully based in opinion....... but it cant be ignored some players are redicuosally lucky :P

 

 

 

I know people who have got 3 or 4 barrows items a day, and dont often go a week without a dragon drop. Based in statistics I would exist in beeing an unlucky player.... I have killed 1000's of metal dragons, and done 100's of barrows trips to no reward :( (note I am mentioning metal dragons interms of their precived drop rate of 1:200 for dragon legs/skirt and not the visage)

 

 

 

 

 

I think the rate you have said though seems low, and I guess I would say a 3rd age kite should have a higher rate. Paul once stated on the offical froum that the 3rd item (any) kite from a level 3 clue is the same as getting a dragon half sheild drop from a skelaton, I belive this to be in the region of 1:10000 due to some reports I hear, which would make me thing the drop rate may be around 1:12000!

 

 

 

Obviousally this is totaly guess work, and I am sure you can work it out better than me (waves my nuramacy dyslexia card).

 

I can't understand why people can't accept that it's random.

 

 

Because it's human nature. People overestimate their chances of being both very lucky AND very unlucky. On the former, why do you think lotteries and casinos are so wildly successful? On the latter, you have folks who worry about planes crashing or being hit by lightning, when they are far more likely to die from something mundane like a car crash or heart attack.

 

 

 

Thanks for the nice discussion so far.

 

 

 

~q

wow, nice math work :D

 

 

 

I agree how luck is more of an algorithm when it comes to numerical chances. Thanks for the statistics.

Utopianflame

 

 

 

You obviously know more about this subject than I do but this is my reasoning for saying that the distribution is normal.

 

 

 

If the chance of getting a certain rare drop ® for each kill is 1 in 2000, the probability is 0.0005. This means that the probability of not getting the drop is 0.9995. If player x kill a monster 50 times, the probability of him not getting the drop is (0.9995)^50 which comes to about 0.975 which means that player x has a 1 in 40 chance of getting the drop. If x kills another 50 monsters, the chance of not getting r decreases to 0.951. To get a chance of getting the drop that is over 0.5, x would have to kill about 1400 monsters (as (0.9995)^1400 comes to 0.496 which is the chance of not getting the drop).

 

 

 

If 1000 players were doing this, some would get it before 1400 kills, some would get it on their 1400th kill and some would still not have it after 1400 kills. As fewer people (on average) would get it before 100 kills (for example) than before 200 kills, If you plotted a graph of number of kills before drop (on the x axis), against number of people getting that drop on that number of kills (on the y axis), the line would increase as it does on a normal distribution graph as the chance of you getting it (overall) increases as the number of kills increases.

 

 

 

As about 50% of players have got the drop before or on their 1400th kill, the numbers from their on in would decrease as more and more players get r. This would mean that the graph would go like the other side of a normal distribution graph as the number of people getting the drop from the 1000 players (for the first time at least) would decrease as most people would have got the drop before 10000 kill (as an example).

 

 

 

Also, as there is no certainty of getting the drop even after 100,000 kills, the line would be an asymptote, just like a normal distribution graph.

 

 

 

Like I said I have just taken my GCSE in stats (I don't even know what my result is yet) I cannot claim to be an expert in any shape or form but that is the logic I was working from.

Ah I can where your coming from now, the main flaw in the logic is simply that both ways you can model this (number of killls until the drop, or number of drops in a given number of kills) are discrete, while the normal distribution is continous.

 

 

 

Of course by the central limit you can consider the binomial distribution to be approximatly normal , however the very small value of p means that you would need a very large value of n (about 5 times 1/p or more) for it to be a good approximation.

Nickelliston I cant see why you would say the distribution is normal.

 

 

 

For any given rare drop, from any single monster you will, on a specific kill, either get it or not get it.

 

 

 

This is a Bernouli distribution, so if you are repeatadly killing this same monster over and over again what you have is a sum of independant Bernouli distrubtions, which is the binomial distribution, where x is the total number of frop you are looking for, n is the number of the monster you have killed and p is the probability.

 

 

 

Now based on the two wheel apprach to rare drops N64jive mentioned, if you say that P(A) (A = 0 , no rare drop. 1 , rare drop. this is simplistic, but completly enumerates all possible outcomes and covers our purposes) is the probaility of getting any rare drop on the first wheel, and the second wheel giving the rare drop in question is P(B) (B = 0 some other rare drop. 1 the rare drop in question)

 

 

 

So the probability of getting the drop in question is P(A)P(B) (normally this would require independance, which we dont have, but the definitions in the model give this result to use anyway)

 

 

 

Now choosing P(A)P(B) to be something tiny what you find is that the odds of not getting the rare drop in question after killing 1/P(A)P(B) of the monsters (n = 1/P(A)P(B)) which is where you get the mean of binomial distribution to be 1 (mean = np) the probability of not getting it tends to about 0.368.

 

 

 

There is a binomial distribution probability calculator here if anyone is interested http://www.stat.tamu.edu/~west/applets/ ... ldemo.html

 

 

 

I love tipit. actually intelligent people. or at least they took STAT.

 

I already posted my opinion a few pages back

Drop rate is 0% + luck *lol*

 

 

 

I seriously dont think that RS has an algorithm with a fixed range of numbers of kills after which you simply HAVE TO get the drop.

 

I think the algorithm for randomization does offer the possibility of NEVER giving the player the drop, no matter how hard you try

*Ahem*

 

 

 

You all are wrong. Runescape, much like any other computer based system, runs off of a series of algorithms. If the game deems you as 'lucky' you will forever be lucky; Conversely, if it deems you as 'unlucky' you will forever be unlucky. There's nothing random about it.

 

 

 

Drop rate is 0% + luck *lol*

 

 

 

I seriously dont think that RS has an algorithm with a fixed range of numbers of kills after which you simply HAVE TO get the drop.

 

I think the algorithm for randomization does offer the possibility of NEVER giving the player the drop, no matter how hard you try

 

 

 

That's what I said on page 4 (And just above).

Sly I cant help but feel that the only accurate thing in your post is when you said there was nothing random about it, of course there is not, this is a computer, it's only psuedo-random, computers are incaple of generating truly random results, but under the right conditions and useing the right generators, they can give a very good impression of it.

You all are wrong. Runescape, much like any other computer based system, runs off of a series of algorithms. If the game deems you as 'lucky' you will forever be lucky; Conversely, if it deems you as 'unlucky' you will forever be unlucky. There's nothing random about it.

 

Lol I hope that's not true