This post was inspired by @freegenz 's post (The cost of acquiring all mythics - MtGPQ General Discussion - 505 Go! Official Forums) and @Volrak 's previous analysis of fat packs even before the drop rates were known. I can’t trust myself with doing the math so I relied on a Monte Carlo simulation (10,000 random Premium Pack pulls using the rates posted on the booster pack). I’m a statistics nut who also knows R programming so writing the simulator was not much of a problem. Only risk was once I get into it, I have a tendency to forget to finish my ToTP (my 60 crystals, nooooooo).
Based on some of the feedbacks I got from earlier work and the comments from @freegenz 's post, I made the simulation a bit more realistic. Rather than running it until I got all the mythics, I ran it starting from zero, counting how many median pulls were needed before the first mythic was pulled. And then I ran it again starting with the cards I already pulled until the second mythic was obtained, and so on and so forth. By the way, I used the median instead of the average because bad RNG has a tendency to inflate the average. If lucky, the sim can get the next mythic on the first pull, if extremely unlucky, it could take dozens of pulls before it sees the next mythic. The minimum is obviously capped at one, but the maximum theoretically has no limit, so the average will get worse as the sim runs into more and more terrible RNG. Also the median is very easy to interpret. It splits the population exactly into the 50% of the lucky ones, and 50% of the unlucky ones. For example, the median for the first mythic was 2 pulls. I got my first mythic after 6 pulls so I know I definitely had a bad run. The average was 3.35 but obviously, half the population got it in 1 or 2 tries so between a median of 2 and an average of 3.35, that’s a significant difference that’s only going to get larger as I run it for the rest of the mythics, so I picked median. By the way, the worst run the sim encountered was 24 pulls before receiving the first mythic. So if you think your luck was rotten, talk to the sim.
My first version of the code did not consider MP as a mythic so it will continue pulling even though MP’s are obviously a lot more valuable. Maybe I will fix this in the next iteration but for now, MP’s are rare enough that I don’t think it would over-estimate the results. However, I did run into some results were the sim pulled 3 MPs before finally receiving the first mythic 
. Come to think of it, it could be the same run that took 24 pulls! As you might expect, the average will be affected by this, but certainly not the median. Which reminds me of another median advantage, the median is always a whole number. If I tell you 50% of the population only took 2 pulls, then you can expect to do the same half the time. If I tell you it took 3.35 pulls, would you do 3 because it rounds down to 3 or would you do 4 pulls to make sure you get your first mythic. You wouldn’t just open a third of the pack and abandon the rest would you?
So here are the results:
Starting from nothing, half the people will spend 640 crystals or an equivalent of $21.33 (based on price of $99.99 for 3,000 crystals) to get their first mythic. Along with that are 25 common cards, 20 uncommon, 5 rares and 315 orbs for that one dupe. Wait what? 315? Since the results are the median of the 10,000 simulation runs, these numbers are the combination of the results of all people whose results are neither lucky nor unlucky. The way to read the result is to take each number individually. How many common cards can I expect after pulling my first mythic? 25. How many uncommon? 20. But don’t expect to get exactly 25 common and 20 uncommon.
From there, it will take another 3 pulls to get the second mythic along with 15 new commons, 13 new uncommons, 5 new rares and another 855 orbs. The owned cards are cumulative, but crystals and orbs are not. Here’s what it looks like if we keep track of the running totals instead.
As you can see, it could take 152 pulls, or 48,640 crystals, or $1,621.33 to get all the mythics (along with one MP) but you should not do that! Not because it’s a lot of money but because at the vault, for 25 bucks, you can buy a color bundle that comes with 50 jewels. For 200 bucks, you can get yourself 400 jewels and get a guaranteed mythic or MP from the Elite Pack. Why spend another $437.33 to pull that last mythic from a PP. Boom, you just save $237.33. No, I’m kidding, that’s too much money people! Seriously, at mythic #14, you would have accumulated 40k orbs (and possibly 400 jewels from playing AX and ToTP) to craft/pull Elite Pack the remaining 6 mythics. The sad part though is you’d still have to pull 52 times or spend $554.67 to get to that point. But it all depends on how much you’re willing to spend or how much crystal you’ve hoarded.
Another interesting point is that the lowest you can spend, if you’re willing to spend is $21.33 (see mythic #1 cost). If a mythic is on sale for under that amount, then you know that’s a good deal, if you are going to spend money. If you have a couple of mythics under your belt already, then pretty much all mythics for sale that you don’t have yet and want are good deals.
TLDR: It’s not the best looking chart, but I’m still figuring out google docs and it’s the best I could do
Opening PPs are only cost effective at the beginning (hence diminishing returns). Once you start crafting or pulling Elite Packs, you are reducing it’s probability of getting your next new mythic. But what I find surprising is that, it doesn’t really get bad until you’ve pulled mythic #16. It’s a gentle rise as you can see on that green curve. Sure, $40, $50 is a lot for some, and at mythic #16, the cost of $85.33 is definitely more than the cost of a new console game, but I know of a few people who have dropped hundreds in this game. To those people, I hope this analysis helps you find your stopping point.
-Brigby