Something needs to be done about Alex Normal looting system.

[Zova]

.....

I don't have the time to do the math right now but I can at least point out the flaw in your methodology that is leading to your abnormally low numbers. The problem is that you are only considering getting 1 drop per run as a single combination.

For instance, if I were to receive a single item as a drop, we would need to consider all the possible combinations that this can occur in. Looking at the instance in which I win the first drop but lose all the other drops, then we need to consider all possibilities for the last 3 drops in which I do not win.

The number of possible combinations for this outcome of Win|Lose|Lose|Lose is:

N = 1*7*7*7 = 343

It is 7 per roll for the last three since there are 7 possible winners for each of the three remaining drops, since I have stipulated that I only win one roll, I don't count myself. Then the odds I win one drop, and that drop is the first drop is:

Odds = 343/4096 = .08 or 8%.

Remember that there are actually 4 possibilities of one win and three losses, so my real odds of winning 1 item is closer to 30%. Adding in the possibility of winning more than one increases these odds even further.

[Kaurie]

Each loot roll is independent of all others. Combining the probabilities as you did was your mistake for determining a roll outcome.
A player has a 1 / (8 - passers) chance of winning a loot role.

Now, if you want to determine the probabilities of winning specific combination of rolls (e.g. wyz) your math is fine.

I appreciate the feedback. I understand that each roll is independent of all others, but that shouldn't negate that there are 4092 different possible outcomes, right? I feel the part that is incorrect must be the possible outcomes in which you receive loot (I counted 14 total, but maybe I was wrong)?

[Kaurie]

snip

Great, thanks! I figured it was something along those lines.

[Pibz]

Boo hoo why didn't RNGesus smile upon me today

Wow, really?!
So first, you reduce my proposal and arguments to QQing, when it's your post that absolutely no substance. Basically you are non-contributing to the discussion, why even post?

Second , yeah i'm being so unreasonable, such fun to be had in running something an obscene amount of times for 1 piece of gear, so fun right?
Why don't we just roll back the measure to reduce this absurdity in other encounters and make it so primals dont drop tokens or so that you don't get tokens on your Alex Savage clear anymore either? I mean it was so much better before right? They were so silly to change it... ( /s)

(edit: sniped out lots of bad math, updated OP to reflect correct info, kept the above bits as the sentiment still stands as dismissing something as "It's just RNG, lols" is pretty irritating)

[Archaell]

]...

Expected number of runs for the chest piece would be still 32, as they drop only in A4 and why would you farm the harder thing, if you don't want the chest piece? Also if you could get only one item, players would be rolling for the shafts first.

Btw. There are also double drops, which improves the odds.

[Dotsusama]

I have nothing to contribute except that one time I had someone win three items in a run, and the fourth (accessory token) went to someone that only rolled a 50. So yes, please one item per run seems more than fair. I doubt very many tokens will be wasted. I always hit greed on the ones I don't need right away, because collecting for alts or potential GC Seal turn-ins later.

[Zova]

From what I understand of the loot system, there are four drops, and all 8 players can roll on any item. This means that for a single run you have the following odds of obtaining at least one item, provided that you, and every other person roll on every item.

P(Get at least one item) = 1 - P(Get No Items)
P(Get at least one item) = 1 - (7/8)^4
P(Get at least one item) = 1 - .5861
P(Get at least one item) = .4139

So by changing the system you have only increased your odds of obtaining any item in a run by approximately 9%.

Since, the math shown by both of us only applies to receiving any item and I don't really know the specifics of the system, I can't comment on how much it would speed up you obtaining that one drop you need for whatever piece it is you want.

[Pibz]

So by changing the system you have only increased your odds of obtaining any item in a run by approximately 9%.

You are indeed correct, as i changed the OP to reflect, it was just bad math (or none at all rather, i had it stuck in my head it was 1/8 for some reason).

But yeah, it's not nearly as big an issue as it would've been if the chances of any drop would be 1/8 (again /facepalm). Though in my opinion it does make it more frustrating to see someone get multiple rolls when you get nothing and there's no real reason not to up the chances on outdated content. But yeah you are absolutely right, thanks for dropping by and correcting it, i probably wouldn't have given it a second though and actually have done the math without it.

Either way, something should be done to smooth out the number of runs required to get say a full set, but im not gonna push as i already lost some credibility with all the bad numbers.

For what it's worth it doesn't make much difference to me i don't really have that many alts i care to gear and my mains don't really need the tomes anymore, it was more about other people who actually need it (ie new players for example) and those people not getting frustrated by it, but oh well.

[Kaurie]

With the help of a few of the responses (mainly by Zova) to my stats (thank you, was having a mind blank from my old stats courses), I've put together an odds table.

Odds of not getting an item 2401 in 4096 58.62%
7 7 7 7 =2401

Odds of getting 1 item 1372 in 4096 33.50%
7 7 7 1 =343
7 7 1 7 =343
7 1 7 7 =343
1 7 7 7 =343

Odds of getting 2 items 294 in 4096 7.18%
7 7 1 1 =49
7 1 7 1 =49
1 7 7 1 =49
7 1 1 7 =49
1 7 1 7 =49
1 1 7 7 =49

Odds of getting 3 items 28 in 4096 0.68%
1 1 1 7 =7
1 1 7 1 =7
1 7 1 1 =7
7 1 1 1 =7

Odds of getting 4 items 1 in 4096 0.02%
1 1 1 1 =1

Odds of getting a specific item 343 in 4096 8.37%
Odds of getting any 1 or more items 1695 in 4096 41.38%

I'd like to apologize for challenging the thought of 50%, it's actually quite accurate. I was definitely in the mindset of odds of getting a specific item after having spent 15 hours trying to get bolts and shafts :\

[Koujiro]

I find it amusing that people rather correct math issues than to actually talk about the loot system and try to give better solutions (on the hope that maybe a dev. will see this topic and promote some changes in a near future?).

*ahen*

Giving my own opinion now. The "one item per run" is interesting (at least half of the people in the run will get >something<), but I thought on something that follows the Savage example to eliminate the RNG completely. Give a "Gordian Prototype's Page", or however you want to name it, for each clear we have on each individual floor, and allow us to exchange some of those pages for bolts, chains, pedals, you name it. That will eliminate a bit of the RNG of those places while securing that people will have to run them a sufficient amount of times for them to not be simply "given" to the players. And, of course, if RNG is on your favour, you will be able to reduce the time you spent there.

The reason for the different pages instead of a single token is to make sure that everybody runs all the floors, instead of spamming the same floor over and over again for easy tokens.

I would go for something like, but things could always be rearranged:
A1 pages: Bolts and Chains
A2 pages: Pedals and Lens
A3 pages: Cranks and Springs
A4 pages: Springs and Shafts

The only issue I see with this method is a inventory one. But inventory is ALWAYS a problem, so, what else can we do? ¯\_(ツ)_/¯