devtrackers.gg
I was curious about how Mercy stacks up against Ferocity and Persistence, if my goal is to get an adept weapon, so I wrote a quick script to run Monte Carlo simulations of Trials passages of various types. That seemed like the sort of information other folks might also be interested, so here it is!
TLDR for getting adepts: Run Persistence unless you're very good. If you are very good, sure, run Ferocity, or if you're "I could be a professional" good, run Mercy (but if you're that good, who cares, you'll do fine with whatever).
The real TLDR: Run flawed Wealth cards and forget about adepts unless you really, really want the adept version of this week's gun, or you think going Flawless is easy (in which case... why are you even bothering to read this?).
Passage of Mercy
| Win Rate | Average Matches Per Lighthouse Trip | Average Hours Per Lighthouse Trip | Chance of Lighthouse Trip Per Card |
|---|---|---|---|
| 10% | 769,231 | 128,205 | 0.0% |
| 20% | 6,998 | 1,166 | 0.0% |
| 30% | 653 | 108 | 0.2% |
| 40% | 139 | 23 | 1.5% |
| 50% | 49 | 8 | 5.5% |
| 60% | 24 | 4 | 14.7% |
This is the only passage type to include the odds-per-card, since on the others, you just keep trying until you make it. Also, 0.0% is just rounded to the nearest hundredth of a percent, not actually zero (though frankly, it might as well be).
Passage of Ferocity
| Win Rate | Average Matches Per Lighthouse Trip | Average Hours Per Lighthouse Trip |
|---|---|---|
| 10% | 11,628 | 1,938 |
| 20% | 946 | 157 |
| 30% | 226 | 37 |
| 40% | 88 | 14 |
| 50% | 44 | 7 |
| 60% | 26 | 4 |
Passage of Persistence
| Win Rate | Average Matches Per Adept Weapon | Average Hours Per Adept Weapon |
|---|---|---|
| 10% | 613 | 102 |
| 20% | 155 | 25 |
| 30% | 70 | 11 |
| 40% | 40 | 6 |
| 50% | 26 | 4 |
| 60% | 18 | 3 |
A few notes about this data:
- The easiest way to figure out your win rate is to look at DestinyTracker.com. You probably want to look at your win rate both for this season and for some recent weekends, as changes in player population and such can have a big effect.
- As a benchmark, my Elo ratings put me at the 32nd percentile in Survival and the 72nd percentile in Trials (according to DT), so I'm an above average player but not amazing. My seasonal win rate is about 38% solo queuing in Trials, but lower in recent weeks.
- Be aware that all of these cards (including Persistence, wildly) put you in the Challenger Pool when you're still in contention for adept drops, so if you're used to running Flawed cards in the Practice Pool for non-adept drops, you're probably going to have a lower win rate in the Challenger Pool.
- The time column assumes 10 minutes per match, which is around what mine average out to and makes the math easy.
- If you do manage to go Flawless, you can probably farm adept drops on your Flawless card significantly faster than what's listed here, so think of this as an estimate of how long it might take to get your first adept drop on Mercy and Ferocity cards, and then you'll get another one every 20-30 minutes after that, on average (if I understand the post-Flawless adept drop rates correctly, which I might not).
- This data assumes they sort out whatever is causing Persistence cards to randomly get flawed. It definitely seems to be more than just the "win the first game" bug that was reportedly fixed, but I have no idea if the patch may have also addressed whatever else was going wrong.
- After some experimentation, I was surprised to discover that the best strategy for a Passage of Mercy seems to be to reset if you lose once before your 2nd win, or twice before your 3rd win. (I used to just reset on any loss before the 3rd win, but that's slightly worse on average.)
- Since these are Monte Carlo simulations, there's a chance that they might proc a melee charge refresh. :-D
If you'd like to see the source code that generated these numbers, or run it yourself to check some different win rates, try different strategies for resetting Mercy cards, etc, you can find it here: https://github.com/djspinmonkey/tribulations Be aware, of course, that you might get slightly different results, since that's the nature of Monte Carlo simulations, but they should be very close to what's above.
External link →