Rico Skill Analysis & Commentary
Passive: Obtain a Tarot Card every 3 attacks, up to a maximum of 3 Tarot Cards. Each Tarot Card increases her damage by 15% for 5s. Active: Consume all Tarot Cards and fire a shot that will not miss, dealing 3x damage do the furthest target and possible critical damage. Additionally, deal additional damage to enemies within a 1-unit radius around the target. The additional damage is equal to 0.5x the number of Tarot Cards consumed, up to a maximum of 1.5x damage.
Update, 2021-April-15: all tables/text have been updated to reflect Rico's stacks being multiplicative.
Rico's skill has two components. With her passive, she builds up to 3 Tarot Cards, giving her a constant 52% damage increase if she maintains maximum stacks. When her skill activates, she shoots the furthest enemy with a miniature bamboo shot (that can crit), plus explosion damage to nearby enemies. The high uptime passive, explosive bamboo shot, and 45% critical damage SPEQ make Rico a very powerful rifle.
One idea I've heard mentioned a few times is turning her skill to force manual and letting the stacks build up - as her active ability consumes all Tarot Cards, it lowers her damage afterwards. Today, I'm going to investigate the value of this, and how she stacks up with and without her active component. A third option is to have her skill on force manual, activating it at specific times, but that's nearly impossible to have a general discussion on and would be a nightmare to collect good data for. That said, it is theoretically always the best option as long as you're willing to manually control her skill activations - you can then emulate either of the above strategies or something in-between, depending on the situation.
Simulations First#
I'm pulling these numbers from this sim which is more up-to-date on dolls, averaging the results of 100 trials. For the full team numbers, it's M14 mod, 5-7, Calico mod, and P22. I'll be using the same dolls for the target practice tests.
On their own, autoskill gains a substantial lead over leaving it off:
In a team with M14, Rico has a slight advantage with autoskill and about ties with no skill activations:
These are all with 1 enemy group, so it shouldn't be adding splash damage. On the other hand, potential overkill from her charged shot is counted. So: autoskill gives higher theoretical DPS, ignoring splash and overkill. This holds when she's on her own and in the team above.
Let's consider for a moment why this is. To help figure this out, I've got some tables on relative per-shot damage for Rico's first few attacks under some different circumstances.
First, the unrealistic assumption of her being able to maintain maximum stacks permanently.
| Starting at max stacks, never decaying | |||
| Attack | Auto | Manual | Event (auto) |
| 1 | 3 | 1.52 | Nuke |
| 2 | 1 | 1.52 | |
| 3 | 1 | 1.52 | |
| 4 | 1.15 | 1.52 | Card |
| 5 | 1.15 | 1.52 | |
| 6 | 1.15 | 1.52 | |
| 7 | 1.32 | 1.52 | Card |
| 8 | 1.32 | 1.52 | |
| 9 | 1.32 | 1.52 | |
| 10 | 1.52 | 1.52 | Card |
| Total damage | 13.9384 | 15.2088 | |
Under these assumptions, attacks before her nuke and after the auto setting climbs back to 1.52 will be the same either way, so I've left those out.
The damage used per attack is relative to her base FP. Accuracy is ignored, so her sure-hit skill will become better with evasion. Anyways, if she can always maintain max stacks her DPS looks pretty good. But is that a realistic assumption?
Now, using the power of observation, I've let Rico shoot the target practice dummy 13 times with and without her skill (left), recording her stack count/per-shot damage each time. On the right I've done the same, but with Welrod+Makarov+TEC-9+P7 combining to give her a constant 80% ROF buff and no skills to influence matters.
| Unbuffed | With constant 80% ROF buff | |||
| Attack | Auto | Manual | Auto | Manual |
| 1 | 1 | 1 | 1 | 1 |
| 2 | 1 | 1 | 1 | 1 |
| 3 | 1.15 | 1.15 | 1.15 | 1.15 |
| 4 | 3 | 1.15 | 1.15 | 1.15 |
| 5 | 1 | 1.15 | 1.15 | 1.15 |
| 6 | 1 | 1.32 | 1.32 | 1.32 |
| 7 | 1.15 | 1.15 | 3 | 1.32 |
| 8 | 1.15 | 1.15 | 1 | 1.32 |
| 9 | 1.15 | 1.32 | 1 | 1.52 |
| 10 | 1.32 | 1.15 | 1.15 | 1.32 |
| 11 | 1.15 | 1.15 | 1.15 | 1.32 |
| 12 | 1.15 | 1.32 | 1.15 | 1.52 |
| 13 | 3 | 1.15 | 1.32 | 1.32 |
| Total damage | 18.22 | 15.17 | 16.55 | 16.43 |
Autoskill has the lead for damage dealt in these first 13 shots. If you want to ignore the last attack on the left where her skill triggers again, it still adds up to 15.22 (auto)/14.02 (manual). This number of attacks is entirely arbitrary.
One critical factor this doesn't represent: her skillshot basically fires independently from her other shots. As a result, these 13 shots happen a fair bit faster with autoskill on. After reviewing my footage frame-by-frame, with no ROF buff: autoskill fires 13 shots in 15.23s, manual in 17.07s. With a constant 80% ROF buff: autoskill fires 13 shots in 8.20s, manual in 9.33s.
Different levels of ROF buffs and skill timings will change how much of an impact this has, so I can't really assign it a specific value.
In Conclusion
Simulator prediction is that Rico performs better with autoskill on. If she could maintain 100% uptime 3 stacks manual would be better, so long as the enemy doesn't have crazy evasion or come in a massive swarm. However, she can't manage that even with a constant 80% ROF buff. HG buff skills could help while they're active, but those don't last forever. With how her skillshot works, she will also get more shots out with autoskill on.
Target Practice#
To find if there's a difference between Rico's skill settings, and which is better if there is, it's time to collect some experimental data. Rico will be kept in the same generic RFHG squad as used above, and I'll record data for various enemy+fairy combinations. For some middle ground between my sanity and the reliability of the results, I'm running 20 trials with each combination.
Most of the analysis uses box-and-whisker plots to represent the target practice data. The central mark in each box is the median, and its top and bottom represent the 75th/25th percentile values. Whiskers extend to encompass nearly all the data points, while outliers past that are marked individually.
Strength 5, except for the Wraiths which are put to 4 (they walk past otherwise). Twisted Faith is fought at night to make it slightly more challenging, Aegis & Deer is the same either way, and the Wraiths were set to day.
There are hard ways to analyze these results and there are easy ways. Generally, if the boxes are overlapping a lot, there is not likely to be a statistically significant difference in performance. To make it more precise, we could add notches to the box-and-whisker plots above.
Then, we can visually compare the notches and see for which configurations there is a statistically significant difference in performance, and which way it leans. Now, the sad news is, these notches can look pretty funny if their width extends past the main box making it awkward to present. This becomes more exaggerated if the data is skewed to one side of the median, a particularly large problem for the damage taken charts when the median damage is near zero. Keeping them mostly secret from the world, you can preview this look in the album here with some extra text pointing out the significant changes. These recommendations are made with 95% confidence.
In short - overall performance is quite close between auto/manual skill on Rico, but most significant differences observed favored autoskill. No significant differences in the team's damage taken (perhaps the most important metric) were observed. Rico herself seemed to contribute as much, if not slightly more, damage with autoskill compared to manual across the board. In the nine configurations tested, Rico cleared faster on autoskill in 4 of them, slower in 1, and was about equal in the remaining 4.
Conclusions#
While there are far more possible fights out there and team compositions, there's not much of a case for ignoring her active component so far. Anyways, who can say no to a surehit explosive shot?
Some further considerations:
- Holding her active until 6 seconds in or so for a new enemy to enter range or for HG buffs to activate may be beneficial. This will be very fight-dependent.
- Enemy evasion, density, armor, and so on are important factors in determining her skill's effect, as is team composition and buffs. More tests with different teams and enemy groups would be useful.
If you're not as concerned with her specific skill modes and just want to know about Rico in general: she's amazing. I would strongly recommend raising her. She can fit in like a FP selfbuffing rifle in most echelons, and can be significantly better than one in some cases (like vs Dog/Archer compositions). Her SPEQ is very much worth enhancing, too - 45% crit damage is crazy.
Update: More Target Practice#
(Added on 2021-07-26.) Target Saving is now a thing (and has been for a while), so it's about time to test a few more echelon compositions against some tougher enemies.
Team 1 (left): Rico/Lee/Calico mod/Stechkin mod/P22/Warrior. It's possible to fit some more ROF in an echelon, but this is on the high side.
Team 2 (right): Rico/Carcano M1891/Python/Calico mod/P22/Beach, a fairly standard Pink/Python setup.
Clear times were usually about equal, but there were two cases where the clear time was significantly faster with autoskill and zero without.
There were no significant performance advantages either way, but depending on your confidence level, Rico with autoskill may have lead to less damage being taken against the Brute/Strikers in each team.
Autoskill was equal or better in most cases, but Rico did do significantly more damage in the left team against the Aegis/Hydra group.
"Relative damage contribution" is Rico's damage divided by the other rifle's damage - while not a perfect metric, a better-performing rifle is likely to contribute more total damage and normalizing against the other rifle helps keep the numbers comparable.
The previous trends observed seem to still hold - the overall performance is close, but you're more likely to get a better outcome with autoskill on compared to only using her passive.
More like this
4 Shiki SPEQ testing
Does it matter?