Data within the game’s files provides a list of raw probabilities for Tokens. It is unknown if this is accurate, since the game’s server is ultimately what tells the client what reward it gets, but it also sends some odds along with it, so I’d advise a grain of salt, but they seem pretty accurate:
The quantity of Tokens you receive is based on the difficulty you play. A Difficulty 10 Boss Chest has a 70% chance to drop 1 Token and a 30% chance to drop 2 Tokens, making the expected average Tokens per chest over time 1.3. Here’s the full table for that.
Difficulty
Expected Tokens
1
1.00
2
1.05
3
1.10
4
1.20
5
1.30
6
1.40
7
1.50
8
1.70
9
2.00
10
2.30
11
2.40
12
2.50
Additionally, the rarity of Tokens you receive scales with Difficulty as well, so you’ll have a tough time collecting Common Tokens in Difficulty 12.
Difficulty
% for C
% for R
% for UR
% for E
% for L
% for M
1
0.400
0.300
0.290
0.010
0.000
0.000
2
0.380
0.300
0.300
0.020
0.000
0.000
3
0.360
0.300
0.310
0.030
0.000
0.000
4
0.340
0.300
0.320
0.040
0.000
0.000
5
0.320
0.300
0.330
0.050
0.000
0.000
6
0.300
0.290
0.340
0.060
0.010
0.000
7
0.280
0.280
0.350
0.070
0.020
0.000
8
0.260
0.270
0.360
0.080
0.030
0.000
9
0.240
0.260
0.370
0.090
0.035
0.005
10
0.220
0.250
0.380
0.100
0.040
0.010
11
0.200
0.240
0.390
0.110
0.045
0.015
12
0.180
0.230
0.400
0.120
0.050
0.020
Multiplying these together can give us an overall average Tokens per run, divided by rarity:
Difficulty
# of C
# of R
# of UR
# of E
# of L
# of M
1
0.400
0.300
0.290
0.010
0.000
0.000
2
0.399
0.315
0.315
0.021
0.000
0.000
3
0.396
0.330
0.341
0.033
0.000
0.000
4
0.409
0.360
0.384
0.048
0.000
0.000
5
0.417
0.390
0.430
0.065
0.000
0.000
6
0.420
0.406
0.476
0.084
0.014
0.000
7
0.421
0.421
0.525
0.106
0.030
0.000
8
0.442
0.459
0.612
0.136
0.051
0.000
9
0.480
0.520
0.740
0.180
0.070
0.010
10
0.506
0.575
0.874
0.230
0.092
0.023
11
0.480
0.576
0.936
0.264
0.108
0.036
12
0.450
0.576
1.000
0.300
0.125
0.050
Which mode is most appropriate for you is heavily based on your speed in battle, but we can see here that the expected yield of each rarity increases with difficulty, except for Common, where the decrease in probability of Commons in higher difficulties lowers yield.
TL;dr If you’re looking for Common Tokens, Difficulty 10 has the best odds, otherwise do Difficulty 12.
EDIT: Medals → Tokens
EDIT2: All tables now account for the base drop rate of 1 Token per run.
With current drop rates, I’ll be able to fully upgrade one mythic every three months. For every one I upgrade, there’ll be 3 more to do. It’s like that “I Love Lucy” episode in the chocolate factory. So it’s one aspect of the game I’ve given up on. Facing level 500 teams with presumably higher drop chances of mythic might make medals less hopeless.
Cedrics at 40% is still well within even the margin of error for even the 95% Confidence Level estimate of what was sampled in the first week or so, I’ll accept it is likely that these are probably accurate to what is happening in game. Even 2% Nysha is pretty close to what sampling was trending toward. I was using the upper bounds estimate of 3% Nysha for test cases to show just how long it would take to elite level a single mythic without gambling medals(as in, bare minimum per)… guess I can use 2% now. And ouch, because that is a huge blow to the “best case scenario average” model I was coming up with.
I assume “expected tokens” is always +1, as in, there is always a guaranteed token added to whatever the “chest roll” is. So a difficulty 12 chest gives 1 token, then rolls either 2 more or 1 more at a 50/50 chance, a difficulty 1 gives 1 token but no others. Tables would be more accurate if they reflected this “guaranteed” token. Edit: wait, that doesn’t make sense either given the tables for difficulties that can only roll 1 token ever… could be just misinterpretation of the way the data is presented, though Edit2: Nevermind, just looked at the tables again, and +1 would make sense here, since even difficulty 2 states it can roll 2 tokens.
Also, thank you for this. This is good info to have. No reason to do somewhat slow difficulty 10 runs now even on alts, I’ll go for the slightly slower difficulty 12 run every time.
All of the token rates in the OP are likely wrong, I assume if anyone got fewer than one token per chest they would have posted a bug report by now. I will say, though, that the token rarities for difficulty 12 match the data I’ve personally collected within a reasonable tolerance, and I primarily run difficulty 12 so at least that number seems reasonable.
It is actually surprising that GoW, which should be by design based on random, has such a secretive random rates of everything. The majority of players have no idea what their chances are at anything, and the majority of the majority apparently does not care. This secrecy of developers (or their lack of understanding) plus pretty lame and really low quality pRNG, partly on the client, partly on the server, plus an attitude of “everything is random, we have so many users, anything can happen”, are the things that create this type of so-called data mining, which is irrelevant and wrong. Look, I found something in the code, lets post it in the forum. No matter what it is, accurate or not, somebody will figure it out.
OP, please kindly edit your initial post to put some clear disclaimer that the information is wrong and irrelevant.
Player based is so information-starved that anything that has numbers and claims, regardless of them being unsubstantiated by anything, already is considered a blessing.
Sirrian, or whoever is in charge of random and drop rates should have told everything to everyone long time ago.
I think in all honesty, level 12 explore should guarantee 3 tokens. It’s a double whammy, of crap 99% of the time and then only getting 2 tokens to try and avoid crap.
RNG on the tokens, RNG on the upgrading really as the higher value/rarity tokens are so rare. How long has it been out, 2 weeks? I’ve got 4 Nysha tokens and 5 Anu tokens. I’ve explored on average for 50-55 hours in those two weeks.
The drop rates are so poor, layering a further level of RNG is taking the mickey.
I’m not interested in levelling all troops with the medals, but levelling even 1 Legendary in 2 weeks, c’mon, it’s a mickey take.
I’d love to see the variation with Difficulty clarified, but huge thanks to @EliteMasterEric for tracking it. I hope we can get the Token-count rates verified, too. Everyone assumes 2-3 means 50% of 2 and 50% of 3, but that sure ain’t how Diablo III does it with Rift Keys.
Anyway, I’ve been collecting my own data, with help from Black Dragon Court. Haven’t been tracking Difficulty, but I can say the following:
Overall drops from Common to Mythic Tokens for 4,168 drops (probably mostly D12):
18% / 24% / 40% / 12% / 4% / 1%
My personal D2-only estimate over the Vault weekend (52 Tokens):
23% / 25% / 33% / 13% / 2% / 4%
My personal D12-only estimate exclusing the Vault weekend (155 Tokens):
19% / 27% / 38% / 10% / 5% / 0%
Obviously, the aggregated numbers are a lot more meaningful.
Has Sirrian said anything official yet, like he promised?
That’s why I just stay on difficulty 10, less headache. Still 2-3 token guarantee and if you follow the mythstone bonus around the map, you still get a boss every 2nd run as well.
I tried timing a bunch of teams at different Difficulties, but got bored.
Mostly, D12 is most efficient for Tokens, Gold and Arcane Traitstones, while D2 is most efficient for everything else. Some detail here: http://paulius.50webs.com/gow-explore.html
Oh, and touch screen is about 25% faster for most teams than mouse.
This is correct. I had mistakenly believed that Difficulty 1 Explores dropped 0 Tokens, and wrote this post with that in mind; this is not the case.
All the values in the “Expected Tokens” table are off by 1. So Difficulty 2 has an expected 1.05 Tokens, Difficulty 12 has an expected 2.5 Tokens, etc.
This also affects the Tokens per rarity per run table at the bottom, but I’ll need to recalculate that later.
The second table is unaffected, the Mythic drop rate is 2%.
. Still 2-3 token guarantee and if you follow the mythstone bonus around the map, you still get a boss every 2nd run as well.