Or a doppel with DPE
The main problem at this time is knows DPK from valuable itens like monsters gems unless they are pure RNG.
it feels like monster gems are pure luck
I got two hannaming gems one right away and another 3 kills later. I would doubt that there were so many people on that same channel with me
You got 3 Hanaming kills in 5 kills without changing channels?
2 hannamings gems without changing channels in 5 kills
That coincides with my other theory about DPK.
So far everyone who i’ve spoken to who has gotten 2 gems in a row(nearly) without swapping channels were killing mobs that have 2 spawns on the map.
edit: Assuming both spawns were killed about equally it would make sense for 2 gems to drop soon after each other with a DPK system.
oO it is double spawned… interesting
yeah just a theory though but makes sense. I’d like to hear more people comment on it but getting gems in general is a rather low chance, and not many people get 2 in a row as well to see if they got it from a multiple spawned mob.
You’ve simulated a great number of chances, around 10.000 tries (kills) to reach a sucess (drop) but it seems like you only tried it with a few screenshots (at least this is what i understood haha)
Also, this is the same effect as few people farming many monsters each one…
For statistical purposes Wouldn’t it be better simulating around 10.000~15.000 people?
Hell, why don’t we simulate 100,000 times or maybe 103,355,924,999,729,472 times…
I don’t think simulating a gem dropping 10,000 is more accurate. There aren’t 10,000 of any one gem that have been dropped in game. You guys are crazy.
What is the reason for all this hate? @ _ @
But okey, let’s begin
If I’m interested in knowing HOW MANY KILLS until get AT LEAST ONE (because this is a rare item, i think nobody here want to farm 15 gems of a same monster), it does not make much difference simulate 100.000times or 103,355,924,999,729,472 times … for a 0:01% chance this will increase the accuracy of 99.9999997% to 99.9999998% or at least something of that range, I agree with you too, it is useless
I don’t think simulating a gem dropping 10,000 is more accurate
In the first example you were right (considering i’ve understood what you’ve said =P)
, but know it’s a bit different
if I want a AVERAGE of how many kills until the gem, yes, simulating 10,000 people farming it will be more accurate than simulating a few people, because in a average the more you have the better it is
There aren’t 10,000 of any one gem that have been dropped in game.
But dude, this isn’t supposed to be a simulation of what is happening right now in Tree of Savior @_@
This is a simulation of the 0.01% chance.
So, in this sense, it doens’t matter how much gems were dropped until now by players at ToS because this simulation work with others things that uses 0.01% too
Only thing is, having 2 spawn points or not, getting 2 gems in less than 400 kills soon as server started(note that OP said after maint all the kill count gets flushed and start a new which makes in that channel only 400 were killed) while I was definitely alone makes the DPK theory invalid.
Which can only end up with game revolves around RNG and might or might not have DPK system on top of it to boost rare drops
It may also just be a mix of DPK on semi rare things while the rarest things are just RNG.
I’m leaning towards whichever theory that does not complicate things too much because there is also no reason the developers would spend extra time just to implement a more complicated system. Having 1 or 2 generic ways to generate drops would make more sense to me.
I did some testing on hanamings last night in a channel with 0 others and I was getting aras recipe from every 200 to 240 kills (testing was killing nonstop to get 5 recipes). So IF there’s a DPK it may be a range thing as well.
#Here:
http://puu.sh/pLBt0/e3bce55e92.png
Ten million drops. You won’t see 10 million drops in the entire life of this game from server open to server close.
It took my code 24 minutes to finish processing this due to the loop needing to continuously run until 10 million drops happened.
Perhaps a more realistic sample is 100 drops:
http://puu.sh/pLBAB/322aee6ac4.png
#50 drops:
http://puu.sh/pLBDI/c66395e924.png
#25 drops:
http://puu.sh/pLBED/30852bc98c.png
#10 drops:
http://puu.sh/pLBFq/568104dd14.png
#5 drops:
http://puu.sh/pLBGA/a0b15170a4.png
It’s worth noting that, aside from the 5 drops (it’s not a large enough sample size), the average remains between 9-10k roughly. That is the important part of this math. Despite a few unlucky or lucky players, people are OVERALL rewarded for their efforts in around 9-10k kills.
That makes this a much better system than DPK.
4 Likes
There can also be a progressive drop rate system where when reaching certain threshold, the drop rate of an item would increase from every kill that does not drop it and then reset once it’s dropped. (This is to balance out the drop rate of very rare stuff)
I’m just throwing out ideas here now. Not saying that’s what’s being used but that kind of system is pretty common in some games.
Here is some data to back up DPK theory
- I choosed Mark Tunic as sample. Mark Tunic is rare drop from Black Maize with “?%” drop chance.
- Addons were used for the research: “Item Drops”, “Monster Tracker”, “cwFarmed”, “LKChannel”
- Three experiments were made: Experiment #1 - empty channel in during one session. Experiment #2 - channel with people. Experiment #3 - prediction of drop Mark Tunic on empty channel
Experiment #1
Experiment #1 was made in 5 sessions:
Session 1
Session 2
Session 3
Session 4
SS3. Kill #1854. Silver mob (you can see amount of silver droped). He droped tunic, sound was not produced tho. I can't say if drop from silver/golden mob do reset channel's kill count or do not. Need more tests with silver/golden mobs
Session 5
As you can see there is definite range of kills before tunic droped. Except silver mob
Regarding my data for Mark Tunic: Minimum = 96. Maximum = 109
Experiment #2
You can see 4 people are on channel. One of them is Masaya. I asked Masaya to kill Black Maizes as much as she/he can. Masaya made 138 kills and found 1 Mark Tunic
As you can see kill count is not individual but global (i guess per channel)
Experiment #3
I picked empty channel than reseted channel’s kill count (by finding mark tunic). After that I made 95 kills and call Masaya to make 1-13 kills more
Video #1
Experiment was repeated second time
Video #2
I want to thank Masaya for assistance!
Updated info in post below
12 Likes
people were saying they found their Vubbe Blood after around 100/130 kills so it might be around there too.
Though good luck with that horrendous spawn time.
Im Guessing Mithril would be in the thousands
@jjjsong @atari DPK is no longer a theory. We have video evidence of a thaum getting a DPK item, killing until the kill count was correct, using swell body on the next mob, and then getting a double DPK item drop on the next kill. There is NO proof that gems are DPK items, but for known DPK items the proof is clear cut. Under any other system, it would be impossible to know when a rare was going to drop, but DPK makes it possible.
We can maybe say they use this DPK system on regular drops including semi rare drops. However it’s very possible on the really rare items (like monster gems) it’s just RNG. So far the testings were on white / blue items or recipes.
It’s also possible that the rate to drop an item increases substantially after a certain amount of kills with no drops. This would be almost just like DPK except there’s a possibility of getting the item “within” the “set kill counts” still.
I.E. for something that drops at 1% chance, after 100 kills of no drops, the rate increases to 50% (or 100% for DPK system). However, there’s still a chance to get the item within the next 100 kills because each one of them still hold 0.5% drop chance.
Basically these testing and observations can spawn many theories and DPK is just one of them.