Crystal Project

Crystal Project

48 ratings
Optimal Stat Growth Guide
By Jale
How to optimally select for stat growth levels based on a *HIGHLY BIASED* interpretation of how stats affect gameplay.
3
4
2
   
Award
Favorite
Favorited
Unfavorite
Stats & Growth Adjustment
Crystal Project has a pretty unique and cool stat system in that it allows you to control the stat growth of any of your characters at a very micro level. Want a Reaper with 60 levels of warrior? Maybe a Summoner with 10 levels of Assassin and 50 levels of Mimic. All things are possible.

Each class has its own unique growth pattern. If all stats were equal these would be roughly equivalent to one another in terms of raw value, with each class gaining between 24 and 26 total stats at each level. You can see this growth pattern expressed on the class selection screen via a series of stars next to each stat.


Stats have two functions. First they provide a bonus unique to each stat:
  • HP - Alived-ness amount
  • MP - Pew pew pool
  • Strength - Defense pierce
  • Vitality - Defense multiplier
  • Dexterity - Crit chance & Crit damage
  • Agility - Accuracy & evasion
  • Mind - Magic defense pierce
  • Spirit - Resist multiplier
  • Speed - Turn Timer reduction
  • Luck - Status application, streak modification
Second they drive class ability damage, with each class using a different core damage stat. There is quite a bit of overlap between these damage stats that can be generalized as:
  • Warriors use strength
  • Rogues use dex in melee and agility at range
  • Mages use mind
  • Healers use spirit


Stat growth can be changed in the capital city just south of the Gaea Temple, in the leftmost building in the middle city tier


Cost scales based on the number of level conversions being purchased, but in general you can expect to spent about 7.5G for 60 levels in a single class.
Starting Point
Things get really spicy when you realize that stats are not all of equal value. The impact of increased base HP, base damage or TT reduction when compared to something secondary like defense pierce is pretty huge. This gives us a solid starting point that we can use to drive stat growth decisions.

The complexity of this system, with different classes getting different amounts of stat but assignable in any quantity to a character regardless of their leveling experience, has led me to the conclusion that the best approach to optimizing growth revolves around managing a series of priorities. The rest of this guide is going to be about around narrowing in on the specifics of this strategy.

Priorities:
  1. Speed for more turns
  2. Base HP & MP for being alive and/or useful
  3. Base damage stat for bigger number
Caveats: HP & MP Growth
HP growth loses a fair amount of value in the late game due to a few factors:
  • Late game areas and bosses use far more AoE attacks, so having one unit with a much higher HP pool doesn’t do much to change how many resources need to be consumed to recoup after taking a hit. You are going to use a group heal of some sort.
  • There are a multitude of very powerful ‘Survive one hit’ style passives available to any character. These mechanics are absolute and don’t depend on variance.
  • Even a fully kitted and specced tank isn’t going to survive two big attacks in a row without healing. This is due to how most bosses used mixed damage and tanks can only really successfully itemize against physical attacks.
Because of this I personally believe the best stat growth pattern selects for HP growth only up until a certain point. How you define that point will vary based on your own personal risk preference, but I tend to try to ensure every character can survive one boss AoE without consuming their survival passives. I generally use 1.1k HP as a hallmark for this, but it is dependent on having end-game gear so you may need to adjust your targets accordingly. It is also worth noting that most plate wearers will have HP available on their BiS gear items, which further lessens the value of HP via stat growth.

MP Growth patterns are pretty similar in principle, with the summoner as a notable standout. (Summoners tend to focus much more single mindedly on MP due to the singularly powerful, consumable and non-scaling nature of their spells).
Caveats: Defense & Multipliers
Looking at the description for stats, you might think that “defense multipliers against static pierce!? I can be immortal!,” I know I did. Defense as an archetype, however, relies pretty heavily on the dominance of the holy trinity (Tank/Healer/DPS) and therefore on the power of threat management.

We have the tools to manage threat, but the frequency and frank brutality of boss magic AoE’s sort of breaks the rest of the strategy. Having a tank maintain threat becomes pointless pretty quickly when the crew dies every other turn to area attacks. It is even further trivialized by how poor resistance gear is for tanks - you will find that it’s not ‘everyone but the tank’ who dies to boss aoe, but just everyone.

For this reason I have heavily de-prioritized mitigation stats and prioritized in their place raw HP. This is both more universal (eg doesn’t require an equip passive to utilize well) and has a better synergy with the variety of survive-one-hit passives and auto-res buffs.
Caveat: Agi & Accuracy
Accuracy scales 1:1 with a characters Agility and is compared to your opponents level when calculating hit chance. Values over 100% hit chance are wasted, and values under that amount can be applied to your average output as a percent reduction equal to the chance an attack misses. All of these values assume you are in combat against an equal level opponent - deviation in that level comparison one way or another doesn't change the rate at which agi converts to damage but instead increases or decreases the amount of agi needed to hit that 100% mark.

The interesting thing, however, is that this conversion scales directly with the primary stat multiple on a typical attack. As a point of reference, most physical attacks vary between 0.5x to 1.5x of the units primary stat. The end result of this is that:
  • Non-agi characters with low primary stat conversion will get about 0.22 bonus damage per agi added
  • Non-agi characters with high primary stat conversion will get about 0.44 bonus damage per agi added.
  • The break even point (eg when 1 point of agility is roughly equal to 1 point of the units primary stat) is around 3x. I think the only ability this applies to is the Reaper's Guillotine.
  • Agi characters benefit from bonus hit chance in the exact same way non-agi characters do, which means high converting attacks like Snipe get something like 3.5 more stat value than any other class.

Caveats: Non-Dex Crit
Adding Dex to a class that doesn't gain base damage through it, but is still inherently physical, will still benefit from the additional crit chance and damage. Upon analysis, however, the total value gained is pretty unimpressive. Though the actual damage contribution per dex changes slightly depending on your total dex, it averages out at about 1 bonus damage for every 9 dex. It's not negligable, but it's not worth investing in.

For classes that do use Dex as a base damage stat (Assassin, Fencer, Rogue, etc), the additional factor of Crit more than doubles the relative value of adding Dex over any other stat. This simply reinforces the rest of the advice here.

Caveats: Speed & TT
Speed is an interesting stat. Increasing the speed of a character reduces it turn timer, which influences both the order and number of actions in a combat. On paper it is the most egregiously overpowered stat in the game. However, the relationship between TT and speed is lossy; the more speed you have, the more it takes to change the turn timer. Or, to put it a different way, the more speed you have the less valuable getting more of it is. Compared to something static in value like base damage, this points us in a pretty clear direction for stat growth: Identify the saddle point in speed growth curve, get that much speed and then stop.
  • Under 100 speed you get more than 1 TT reduction per stat.
  • Between 100-140 speed the TT reduction scales down to 0.5 per stat.
  • Between 140-160 the TT reduction scales down further to 0.33 per stat
  • Finally between 160 and 220 it scales down again to 0.1 per stat.
If you zoom in far enough on the delta between these relative value changes, it appears as if the saddle point sits around the 0.33 reduction per stat, or at 160-165 Speed.

(I have included a link to the data table and sheet that I used to find these values near the end)




***NOTE*** This kind of flies in the face of conventional wisdom on the game, and 165 speed is absolutely trivial to hit. If you are using a class that has any of amount of inherent speed you likely won’t have to pay attention to the stat at all. If turn order (rather than turn efficiency / action count) really matters to your comp (Hello Dream Hunters & Coup De Grace debuffers), you will need to take the time to find the average speed of bosses you face and then make sure you are above that when setting your speed target.
Revised Priorities
This now leaves us with a refined priority list.
  1. Speed up to 165
  2. Base HP up until you can survive a hit from bosses you face
    **Some units (Sleep Hunter / No threat rogue) may want to remove this priority
  3. Base MP up until your mana pool persists across a single boss combat
  4. Base damage stat
Putting it all together
So at this point we have a series of goals / priorities, and an understanding of how much of any given stat a level in a class produces. The process now becomes about identifying classes which produce the maximum number of stats that fall into our prioritization, and ignoring all other stats. Google sheets makes quick work of questions like this. Let's run through an example.

Imagine we are building a Reaper DPS unit. Reapers use Strength as a primary damage stat and don’t use MP beyond a few infrequently used buffs, so we can say the stats that matter are: HP, Str, and Speed. If you add the stat per level contribution of the classes again, after culling the irrelevant, you can see that until your priorities shift, levels in Beatsmith are the most efficient.



Once the first stat priority goal has been achieved (Speed >= 165), we can cull that stat from the evaluation and re-assess the total contribution from each class. Thus, our next class would be Samurai.



Finally, once both the HP and Speed goals have been met, Reaper, with 5 Str per level, should finish out the path.

I made a janky tool for myself when going through this process. Anyone who is interested is welcome to make their own copy of it from here: https://docs.google.com/spreadsheets/d/1D1rI_oXStA4aylZ2Xf2Z4KhCLA779EylNpbcYlJO8ik
TLDR
Summary
  • Speed loses value fast, only take it up to 165
  • Survival passives are better than stacking HP/Def
  • Leveling for accuracy on non-Agi units isn't worth it.
  • Beatsmith is the best class in the game because everyone cares about HP and Speed
  • Find the class that adds 5 to your damage stat and try to use as much of it as possible

Build Archetypes
Strength leveling path:
  • Beatsmith until 165 speed
  • Samurai until HP target is met
  • Reaper with rest

Dex leveling path:
  • Beathsmith until 165 speed
  • Assassin with rest

Agi leveling path:
  • Beathsmith until 165 speed
  • Samurai until HP target is met
  • Hunter with rest

Mind leveling path:
  • Dump it all into Shaman. It’s universally good.

Spirit leveling path:
  • Chemist until 165 speed
  • Rest into Cleric
18 Comments
slartifer 2 Sep, 2022 @ 6:43pm 
I came to the comments section looking for pretty much exactly what Niflthaena posted. I don't understand why the damage stat is treated, in this guide, as if it doesn't have diminishing returns, when that's inherent to linear scaling stats. Seems like an update is needed.

That said, I appreciate a lot of the other suggestions and advice in the guide, and the inclusion of graphs is a really nice plus.
Niflthaena 21 Aug, 2022 @ 10:20am 
Speed and Turn Time are deceptive because Turn Time doesn't give flat returns, it's the inverse of a term that does - Turns per Time.
Doing the math based on Jale's spreadsheet, every 50 points added to speed gives you 7/6 as many turns as you last had.
This means once a flat-scaling stat (Strength, Spirit, Mind) hits 300, every point in speed will give more DPS than a point in the damage stat.

This has several caveats:
* Non-linear stat growth (Agility or Dex classes) will usually scale better at 300 points, so their break-even point will be higher.
* Charge casters (Cleric, Scholar, Wizard, etc) add FLAT time on to their turns. This has a much steeper relative effect the quicker your turns are.
* Utility casters (Warlock or Chemist, or debuff Rogue/Fencer) may not need a damage stat at all. Buffs are buffs.
* Utility receivers (such as a heavily buffed tank or assassin) generally consume their buffs based on turns; the slower they are, the longer they stay buffed.
HeraldOfOpera 17 Aug, 2022 @ 4:39pm 
My reading was that not only is the 16 speed not worth it, the 46 speed above 165 you have isn't worth it either.
BadMedium 11 Aug, 2022 @ 12:57am 
I'm level 57 on my Dex char, and if I put all 57 levels into Assassin, I end up with 211 speed. If I go by your comments at the bottom and hit the 165 speed on Beatsmith before putting the rest into Assassin, I end up with 227 speed, but with 22 less dex comparable to full Assassin.. Is the extra 16 speed worth it over the 22 dex?
Jale  [author] 2 Jun, 2022 @ 9:42am 
@BBK you can use the chart
Jale  [author] 2 Jun, 2022 @ 9:41am 
@Peng this is super interesting, thanks for modeling that out. I'll apply to to the existing crit calcs and see how it changes things.
BBK 31 May, 2022 @ 6:00pm 
Is there a chart to see how the main class re-adjust stats when picked or can we go off the growth chart to get an idea?
Peng 30 May, 2022 @ 9:09pm 
Luck basically gets you extra variance rolls. luck factor = (user luck - target luck / 2) * 0.75, and every 100 luck factor gets you a bonus variance roll, which replaces your previous rolls if it is higher. I simulated 1M rolls at varying luck differentials for 1000 damage, 75% variance, and the results are available here [docs.google.com]
Jale  [author] 16 May, 2022 @ 8:21am 
Luck would be interesting to model but the game doesn't give you any indication of how an increase to luck impacts variance damage. It's pretty easy to go about modeling things like hit and crit when you can see flat %'s with each stat add but I don't have the patience to test enough to dig into luck.

If anyone wants to go get me a ton of data (luck, hit min, hit max, hit actual) I'd be happy to model it tho.
Iconmaster 15 May, 2022 @ 3:44am 
I'm curious as to the difference in damage output Luck makes, especially if going with variance attacks such as Beastmaster or Lucky Dice Scholar. The latter is especially interesting because Luck both effects its raw damage and your chance at getting a high damage range.