I'm in the top 1%!

Pongie

mpqr7 wrote:

It's not just "having a bad day". It's pretty frustrating how hard it is to truly progress in this game. That's why I'm spending more and more time on other games instead. I haven't pulled a new oml token in over 6 months, and it's not from a lack of trying. And my other 5*s aren't very useful either, as they either have too few covers or are not yet with all three colors.

I'm doing okay, not great. Hopefully in another few years, I'll get somewhere...

5*s Covers
OML 5..1..4
JG 0..2..5
SS 3..1..1
BSS 1..3..0
IM 2..0..1
Cap 1..1..1
GG 0..0..1
BB 0..0..0

mine is very similar

OML 4/1/4
PHX 4/1/3
SS 2/1/2
BSS 0/0/0
IM46 4/1/2
Cap 1/0/0
GG 0/1/0
BB 0/0/0

Let's meet up in a few years to compare again

HaywireII

Updated the original post with numbers to add a little more value to the discussion. Thank you fight4thedream for unlocking the thread and thank you Malcrof for feeding me cupcakes in Slice 4.

BlackSheep101

I arbitrarily picked 40 pulls as my "unreasonably high" cutoff and added up how many cases were higher out of 100k. There are way too many thousands for comfort...

HaywireII

I decided to run the numbers another way. I simulated pulling 200 tokens 100,000 times and checked how many 5 star covers would show up. Amazingly it was never zero. I would hate to be one of the 11 people who only got one though. The distribution here pretty much looks like the standard bell curve ranging from 1 to 56 covers with the peak at 30 as expected for a 15% draw rate.

11	1
12	3
13	14
14	27
15	39
16	90
17	185
18	368
19	632
20	1068
21	1613
22	2236
23	3134
24	4065
25	5106
26	6155
27	7106
28	7593
29	7755
30	7934
31	7574
32	7026
33	6416
34	5451
35	4514
36	3714
37	2880
38	2106
39	1612
40	1184
41	863
42	558
43	359
44	228
45	156
46	97
47	62
48	36
49	20
50	11
51	5
52	3
56	1

HaywireII

Well it's over on pull 30. I just championed Beast and the Legendary token gave me Bruce Banner purple. Back to record keeping.

HaywireII

aesthetocyst wrote:

I believe that's showing that 1 run yielded "only" 11 covers and that was the minimum. Not that 11 runs had 1 cover each. If it was the latter, then your table implies that 30 runs coughed up 7,934 covers on only 200 pulls

Now that would be some positively broke RNG, harhar.

No it was just the way I modified my existing sim to repurpose it. 11 of the runs only yielded 1 cover and 7,934 of them yielded 30 covers. I guess it would read better if I had swapped the columns.

If you go back to the original post you can see where I ran 100,000 dry spells. The longest one was 70 of those 100,000 runs. Not sure what you'd consider to be average longest but you can cut and paste the numbers out of there if you want to play with them.

Hadronic

aesthetocyst wrote:

If you can, set it up to look at the long runs. What's the average longest dry spell in each run? That's where the rub is.

I opened 200 tokens at a time, and looked for the longest dry spell. Did this 50,000 times and got an average of 24.7679

Compared this to the average dry spell (all of them, not just the longest one per batch of 200) of 6.6667 as it should be

Eddiemon

HaywireII wrote:

I just opened my 29th legendary token in a row without a 5* cover showing up, all at the new 15% draw rate. If I'm doing my math right the chances of opening 29 in a row without getting a 5* are 0.8977%.

They aren't the odds, and you're not a 1%er.

The odds of getting head or tails is 50% when you flip an unbiased coin. If you flip it exactly once you have a 50% chance of getting heads. But if you flip it 4 times, your chance of seeing heads is no just 50%.

You didn't just open 29 tokens and see 29 misses in a row. You opened at least 34 tokens and saw 29 misses in a row. Which has totally different odds.

simonsez

Eddiemon wrote:

They aren't the odds, and you're not a 1%er.

You're right, the odds of a runout of 29+ in a run of 34 pulls is 1.6%. I'm sure the OP feels MUCH better knowing he isn't getting screwed as badly as he thought he was.

Awesome nitpick, dude...

HaywireII

No, it's cool. I wanted some people who know more about math/odds/statistics than me to comment. I had thought that the chances of 29 misses in a row would be .85 ^ 29 which MS Excel told me was 0.008977 but I'm fine if someone corrects me on a mistake.

Hadronic

It depends on exactly what question you ask.
If you want to know the odds that you will miss 29 times on exactly 29 pulls than the chance is indeed .85^29

However, if you want to know the odds of missing 29 in a row out of 34 pulls, thats slightly more complicated. It's slightly higher. In fact the odds of missing 29 times in a row increases with each pull. This should be obvious by looking at the extreme case, what are the odds of missing 29 times in a row in 10 billion pulls? Pretty close to 100%.
You can reference binomial distributions to find out more about it.

dsds

If we are talking about rng, even if the first person who opens 50 LT in a row and gets no 5* characters is still within the 15% draw rate. It's very very very unlikely to happen, but it can happen with rng. Let's get to the bottom of this. RNG is tinny kitty for progression.

Just get rid of it. Or put some kinda counter on it. So it's 10% draw rate with a counter. So once you hit 20 LT and you haven't pulled a 5*, you automatically get one. This is roughly about $300 when you do buy clubs to get a guaranteed 5*. Chances are you will likely get a 5* with the 10% rate anyway. Thus this lowers the rates (with the 10% draw rate) for people who get extremely lucky (eg. pull 6 out of 10) but then have a bad luck streak (0 out of 30). I'd prefer this so I know on my 20th pull, no matter what, I get a 5*.

Yes I am on a dry spell too. But I don't think I've pulled 20 LTs yet. I got a tiny bit lucky in the beginning and now I am paying for the past luck with a bad luck streak. It's frustrating not to know when the streak could end. I could end up pulling 30LT and get nothing. I am not a hardcore player so I get an LT from time to time. Maybe 1-2 a week or less.