+2 votes
279 views
asked by (231 points)
retagged by
It is said that good behaviour during whole year may be rewarded by nice presents from Santa. Does it work in Tibia? Or that's just absolutely random what we gonna get in present box? I haven't got a skull through last 12 months, hope that Red Guy will notice this thing :)
commented by (958 points)
They aren't random when we are talking to the object of the interest of the question here. What is random are the inputs that are used on the equation, but the resultant gift (what matters here, and what we can quantify and analyze) isn't a random occurrence, it is set to be a low valuable gift (i. e.: red/green/blue bundles) on a huge majority of times. If the data of gifts obtained was in fact random, there would be no pattern to be detectable.

Correct, randomness has nothing to do with equiprobable outcomes, but it can't follow a traceable pattern either.
https://en.wikipedia.org/wiki/Randomness
No contradiction here at all.

Also, slot machines (99% of them at least) aren't random when analyzing the *output* obtained from them. Of course the code will use random/non detectable number (specially because it is the most rational and efficient way to work on a code) on the code line, but the result will eventually be non randomly groups of prizes, with some of then being overwhelming more prevalent than others. A Slot Machine will not give out prizes out on distribution randomness, it will be set to give out a controllable amount of prizes on purpose, in order to result on profit for the Slot Machine's owner. The same with Tibia's Santa Claus, that will give out valuable prizes on a negligible amount. There is no randomness here, the behavior is a expected one.

That's the most direct answer possible to be given to the person who asked, where I also had used Runs Test to explain. The fact that the main char of this user will most likely receive a colored bundle along with snowballs isn't a prophecy by me, but a conclusion that you can have from previously obtained data analyzed as soon you discard the hypothesis of it being a random event.
commented by (47 points)
"Correct, randomness has nothing to do with equiprobable outcomes, but it can't follow a traceable pattern either."

Of course it can follow a traceable pattern. If you try many times (like in the answer below) you will be able to get the frequency of outcomes.

The second statement in the link you sent me just confirms this:
"Individual random events are by definition unpredictable, but in many cases the frequency of different outcomes over a large number of events (or "trials") is predictable"

The santa gift is random, and the output of a single gift is unpredictable. If you try 1000+ times (like in the answer below) you will be able to predict the frequency/pattern outcome.  The event is nevertheless random.

I still don't get your point.
commented by (958 points)
A pattern means a predictable condition, which when established, excludes randomization. The videos that I posted even mentioned other very diverse events analyzed through the Runs Test. The aim of the test was not to tell what will surely happen, but to detect patterns that may indicate that there is a non randomness on the elements presented.

My point was highlighted and it is pretty clear and straight: OP asked if Santa's gifts are random, I said no, negating both that the char behavior can influence it, and also added that there is no randomness on the gift distribution that @Zupakode showed as an example of it being the "completely random" (which made @Ellotris, not me, question on why such values would mean it to be random when they aren't proving it). I sustained my position with a mathematical application that serves exactly for this kind of situation. I wanted to explain that the universe of possibilities will not occur on randomized order: It will be biased to make the player get a low end gift on the majority of times, although there is a small possibility already fixed to be minimal that he can eventually get something good, there is no randomization on the data for the player to receive a gift. Just like Slot Machines, the gift (subject of interest here from the OP) is predetermined as something to be low valued, and that's not aleatory or random. It tends to clear pattern, breaking the basic condition stated on the Wikipedia article that I posted.

By the way, I never questioned the individual experience being unknown for individuals (you and @Yuri seemed to assume that, if yes, I apologize if you both felt that I was trying to confuse people just to win an argument), but I answered that the Santa's behavior is build on such way that not only he doesn't take account of the individual asking for gift, he will obey a gifting order that it is non randomized and gift you with oranges, snowballs, candies and candy canes something around 60% of the time.


Probably the thing missing here is that I need to point out that the mathematical concept is looking from a bigger picture and telling that getting a Dracoyle Statue only once in a lifetime of your Main's Tibia char isn't random at all, but a fact because events were fixed to avoid randomness and to favor a solid distribution of repeated/low value presents. Once I looked at Zupa data, I only treated the question as a mathematical one. Not even once I said that I can predict 100% of the time what people will get from it since I do not consider it being a random event... But I actually should say now that I can predict 60% of the time what people will get from it since I do not consider it being a random event, if that makes things more clear on what I intended with my provided answer.

3 Answers

+1 vote
answered by (195 points)
What is good behavior, that's a topic for a broad discussion as it's an anthropological question.

With that being said, I have some points to make.

You mention that you haven't gotten a skull on the last 12 months, but everyone on non-pvp servers haven't either.

I can't tell you with certainty that it is or that it isn't, on it's current version, a random chance.
What I can tell you is, on the old C++ version that leaked and led to the outbreak of open tibia servers and all its variants is that, it was, in fact a random chance as it was literally written on the code.

My conclusion is, I believe it's random, but I also believe in the secrets that Tibia has, I still find Tibia an incredibly misterious game, despite knowing that it's most just 'fake', 'unfinished' and 'misleading' things, I like to believe there's something I don't know or don't understand, like a riddle; So... yeah, it's random BUT is it really random ?
commented by (958 points)
edited by
I guess your concept is an extremely heterodoxy approach in order to conclude that it is a random event. I guess you have an idea more related to philosophical definitions (which by the way I personally disagree, since it is a personal take related to philosophy point of views which can be affected by cultural/sociological differences, like Relativism or Utilitarianism). I appreciate for the explanation about the C++. Indeed, the values on the equation can be random, but the outcome shows a pattern, so it doesn't follows the lack of pattern on randomness (broad definition here, you can only affirm it mathematically after using a test like Runs Test).

 One thing should be added in order to any conclusion that you can obtain from it: The probability of any loot drop on Tibia is an outcome which we can observe and quantify, it is not a theoretical/non-apparent one. The use of Runs Test serves in a pretty solid way, since the number of categories of rewards can be converted to numbers (a random sequence would show values that do not merge themselves on groups that we can also correlate further with the market/rarity value if we plan to do a bigger study on it, for example, bunnyslipers are very expensive and rare, but candy canes are somewhat cheap and abundant usually). So it is not only a non random event, but also a very good example of an adjusted code line, that regardless of the elements used for output generated on an equation, it will give in the end a much minor number of AoLs compared to colored bundles, and even a less significant amount of Santa Teddy. Modern MMORPGs are 100% related to altered sequences and adjusted events that should be meticulously set. I object to any objection to the use of the Runs Test, since it was designed exactly for situations like that one. I studied other applications and principles available (like Newcomb-Benfords Law) and the Runs Test suits adequately. As you can see on the class video used on references, it was even applied when comparing the amount of sand on soil. There is no overthinking by my part, but I respect your stance.

You (and anyone else wondering about the subject, apparently) affirms the fact that at the moment you ask Santa, you can't predict what you will receive (correct), but the people responsible for Santa's code OR anyone with the data from 1002 chars gifted certainly will most likely know what you will probably get more than half of the times that you goes to Santa with a char (and it fits perfectly with Wiki data my personal experience with Santa by the way). The lack of knowledge of the gift outcome by the receiver does not prove that such thing was made with a randomness of output by the code creator. The pattern stills there, and even the most broad definition of Randomness states that Randomness should present "lack of pattern". Until any mathematical counter proof shows up, I guess my answer will still persist on the approach that there is no randomness to the fact that you got on the last ten years only snowballs and colored bundles with your char: https://en.wikipedia.org/wiki/Randomness

I used "Times to get one", but even the column with the gifts obtained out of 1002 would show the apparent pattern.

Anyway, I will do a big study about it during the Santa Event (got some spare time on the next weeks). Hopefully I will be able to get some reliable data if you think that Wiki's data can be inaccurate.
commented by (958 points)
Rest assured that I respect what you said. I only want to say that for this case, we can detect a clear pattern operating, although you can't say that you will certainly get a good or bad gift on your individual experience. But it was designed against a high value gift.
commented by (195 points)
edited by
The first paragraph of the wikipedia mentions it's relation with entropy.
The wikipedia link itself has science related articles, none of which point to the other study you mentioned.

By reading the entropy entry on wikipedia, you'll be able to see how it directly relates to the possibility of the statistical distribution of the data.

The data you're taking into account, those 1002 reports are innacurate.
You should not take those into account, and of course, you'll need a much larger pool to start evaluating what the chances really are.
So no, we cannot detect a clear pattern, as the data cannot be taken into account.
Yes, it was designed against a high value gift, of course.

Perhaps, its how you're approaching the randomness of the gift.

Santa gives 6 prizes, 1, 2,3,4,5,6 gold, each related to the side of a six sided dice.
That means that you have the same chance to get them, all at random.

What if santa gives the the same prizes, but now, the rules are:
if you roll, 1,2 and 3 on the dice you then have to flip a coin and you'll get 1 gold if heads and 2 gold if tails.
if you roll, 4 and 5 on the dice you then have to flip a coin and you'll get 3 gold if heads and 4 gold if tails.
If you roll 6 on the dice the you have to roll another dice, if you roll 6 again, you then flip a coin and you'll get 2 gold if heads and if tails you roll another dice a 20 sided dice, if you get 20 on the dice you get 6 gold, all else, you get 1 gold.

This means you have a very, very different chance of getting the money but still, both are random, the same kind of random.
To be random and to have the same probability chance are two very different things.

And 'I guess your concept is an extremely heterodoxy approach in order to conclude that it is a random event.'
It's not my concept, I have no orthodox or heterodox way of approaching as this is a statistical discussion that can't have my views of the world, none of which I posed, ofc.
I don't have a philosophical idea about the subject nor did I mention any other approach besides the Entropy, which has no philosophical intent or view.

I merely pointed out a few things.
'Random is hard'
'C++ source code used srand(TimeInMillis);'
'Study does not apply'
'Data is not reliable'


-------
This is a great video somewhat related to the topic.
https://www.youtube.com/watch?v=sMb00lz-IfE
and
https://www.youtube.com/watch?v=9rIy0xY99a0
0 votes
answered by (1.2k points)

The present from Santa Claus Npc is completely random.
I can paste you some statistics from santa.
 

Obtaining gifts from Santa Claus, 1002 trials.
 

ItemAmountPercentageTries to get 1
10 Candy Canes + Green Christmas Bundle22122.06%5
10 Candy Canes + Red Christmas Bundle20220.16%5
10 Candy Canes + Blue Christmas Bundle19119.06%6
10 Candy Canes + 10 Snowballs15815.77%7
10 Candy Canes + Piggy Bank737.29%14
Dwarven Ring + Ring of Healing + Energy Ring454.49%23
Red Gem323.19%32
Bullseye Potion + Mastermind Potion + Berserk Potion181.8%56
Shiny Stone161.6%63
Violet Gem131.3%78
Gold Ingot111.1%92
Amulet of Loss60.6%167
Dracoyle Statue40.4%251
Panda Teddy30.3%334
Stuffed Dragon30.3%334
Snowman Doll20.2%501
Santa Doll20.2%501
Dragon Claw10.1%1002
Santa Teddy10.1%1002
commented by (706 points)
The table does not prove they are completely random. Is there any official confirmation of this answer?
commented by (958 points)
Ellotris is right. The mathematical test proves that there is no randomness attached to the gifts generated by Santa Claus NPC.

By the way, the data was collected from here https://tibia.fandom.com/wiki/Santa_Claus/Presents


I myself will do a "Santa Claus Test" this year. I guess that further data obtained can be pretty useful for the discussion here.
commented by (1.2k points)
Well by random I mean there is no thing that affects on the static % of each item, when I say random I say that all people have the same % for each item to get it.  Obviously each item has a % of drop like in a normal monster or boss .
commented by (958 points)
edited by
Hm, sorry if I look kind of picky, but your answer itself shows a little confusion that even the user asking the question probably committed as well. This is a similar concept that people may have about Slot Machines. Most of people play it thinking (well, maybe not your case, but just to illustrate on how statistics can be confused) that getting a nice reward on these casino machines is a random event with a lot of difficult due to the several possible combinations that you can get, BUT it isn't random at all. A biased code is inserted in order to prevent randomness (on the case of Slot Machines, usually they are set to give out big prizes only after a certain number of money is inserted). Neither are monsters drop a random event at all, because they follow a biased distribution. You won't see Dragon Scale Mails at the same rate of Dragon Ham on Dragon Lords loot, because these events weren't made to be random, they were made to follow a distribution.


Nevertheless, I understand your point. You can be lucky enough to get the Dragon Lord that is carrying a Dragon Slayer (Extremely rare) or the Frost Dragon Hatchling carrying a Spellbook of Enlightment (very rare). I only killed  7 Frost Dragons Hatchlings (According to bestiary) on the Kingsday event at the Arena on my Tibia main char life, and after killing 7 Frost Dragons Hatchlings, he got a very rare item. Believe or not, I got a Dragon Slayer on  DL spawned on that arena at the same day.

 But this is a "lucky event", also showed by the statistic. Just because it was very non probable event, it doesn't mean that it won't happen. But again, not random at all.
0 votes
answered by (958 points)
edited by

Actually, I guess that the chances of getting certain presents were proved as being very low according to the table posted by @Zupakode, so it is not a random event. Most of the time you will get a colored Christmas Bundle along with something, as your data already shows.

Mathematics can provide solid evidence for it on the field of statistical analysis. One example is the Runs Test. This test can reveal a random pattern if present, through mathematical calculation. We can use an Online Tool as a resource. We can see a very easy website to operate with this statistical test right here

You can do it with any group of number. A P-value inferior to 0.05 will indicate that the group of numbers tested are actually following some kind of pattern or modification. Going for the analyses of the column called "Tries to get 1" resulted on a P-value of 0.00003. This is a very low number, which is accompanied by the website tool signalling as the conclusion the following statement: "Very strong evidence against randomness (trend or seasonality)".

For more information about using the Runs Test on Excel, you can access this link and do it by yourself: https://www.youtube.com/watch?v=YWlod6Jdu-k

This friendly class can tell more in depth about details regarding the theory behind Runs Test, along with other examples: https://www.youtube.com/watch?v=9YLgBwMQoMI

You will usually receive valuable items much less often compared to Bundles. On my personal experience, after several years of Tibia, the best that I got from it was an AoL on a character that doesn't even exist anymore. It is basically a "lottery" that every character can join once a year each December with 0 cost per try.

enlightenedAlso, I should highlight that there is zero evidence about player's behavior being something that can influence the present that you will receive from Santa Claus NPC (unfortunately).enlightened

As my mathematical approach revealed, the group of numbers obtained from the huge amount of trials (chars gifted) contained on the answer made by @Zupakode showed that there is no randomness at all when analyzed. The programmers set Santa Claus to give certain presents more frequently, so there is no random number generation on the code for Santa Claus NPC currently. If you are looking for the biggest prizes, you have a mathematical disadvantage set against you (just like on Slot Machines in real life, although many people think that they are random as well).

commented by (47 points)
I think the problem is that you used the wrong data in the test. You can't use the column 'tries to get one' as the input to the test.

Actually you can't run that test with only the table posted by @Zupakode. You would need the whole sequence that the presents were given. For example: 1st trial: Candy, 2nd trial: Candy, 3rd trial: Red Gem, ... 1000th trial: Candy.

The test is based on the 'runs' that the output is given. If you consider the 'Tries to get one', it simply does not make sense.

Suppose there are two presents only: present 1 and present 2. If you test 20 times and receive the presents in this order: 1, 2, 1, 2, 1, 2, 1, 2, ...,  then the Runs test will say that this is not random, because there is a clear pattern there.

If you receive in this order: 2,2,2,1,1,2,1,2,2,1,1,1,2,1,1,1,1,1,1,1 the test will say it could be random, because it couldn't detect any order in it. In fact I used a virtual random coin flipper to generate this sequence.

If you simply add the 'tries to get one' as input, it is simply wrong.
commented by (958 points)
I visualized the column that I used as the number of "effort" requested to get such presents, so since I coudn't find a detailed sequence, I had the number of "tries spent" that could be run as the "sequence of effort" (statistically, you would see only certain items after a given number of hundreds of tries, sometimes earlier, sometimes sooner, but not on the big picture it would be something like this), but I admit that the utilization of such is probably not the best, but it was the most fitting thing to utilize.

Nevertheless, I'm aiming to do exactly what you said, and I will run a test using a perfect sequence in details. I will edit the post with new conclusions as soon as I can get it and organize the data.
commented by (47 points)
In your answer you are basically showing that the sequence: 5 5 6 7 14 23 32 56 63 78 92 167 251 334 334 501 501 1002 1002 (in this order) is not random. Of course it isn't, the sequence is sorted. There is a clear pattern in it.
If you could get, somehow, the presents received in the order that they were received, then you could try this test to check if it is random or not.

You don't need to do that because it is random.

I read some of your comments and you seem to fail to understand that random events can have outcomes with different probabilities (like slot machines, lottery, santa present.. ).

Imagine you are santa and you use a random dice to decide the present you will give to someone. Suppose you decide to gift a pair of golden boots if the dice shows a 6 and a pair of boots of haste if it shows any other number. It should be clear that the dice is completely random and so is the gift. You have a 1 in 6 chance to get a golden boots, and a 5/6 chance to get a boots of haste. That does not mean that it is not random, just because the presents will be unevenly distributed. The output is completely random, because the output is based solely on the dice, that is random.
...