Cornucopia Devblog 1: Strategic Balancing

A Hotfix for updating Cornucopia to v1.0.0.4 consistency is in the works. Expect it sometime in the next couple of days. v0.3.1 still has a lot of work to be done for it and ultimately we’re still not entirely certain of everything we want to do for it. We’re not putting out a timeline for its release, but my personal estimate would be ‘not as soon as v0.1 and v0.2 released relative to the mod’s announcement/v0.1, respectively.’


The Cornucopia project is a complex one. To some it may seem as though we’re making decisions entirely arbitrarily and so far we’ve simply had good luck with our changes. To others it may seem as though this is the ‘community choice’ mod in which we just slap your feedback into the mod and call it a new release. But if either case is closer than the other, there is yet more going on behind the scenes for the mod. In these devblogs, we hope to open the curtains a bit so that you, the community, can better understand our approach to rebalancing Grim Dawn. Perhaps you’ll learn from it for your own modding adventures. Perhaps you’ll teach us something more to better the quality of this mod. Or perhaps we’ll just eat up more space in the modding subforum. The possibilities are endless!


Does this image make you feel uneasy? Does this suddenly make you want to uninstall our mod? Don’t worry, we have countermeasures:

If it wasn’t obvious enough, this devblog is all about buffing and nerfing things in the subjective name of ‘balance’. Specifically, this post is all about the strategies involved in moving forward to such a point.

If you’ve been following along with Cornucopia so far, you’ve probably noticed we’re buffing the hell out of a bunch of things and leaving anything that was already pretty strong more or less alone. On the surface, this probably garners as much endearment for the mod as it does backlash; some will laud it because it will make their favorite skills/items/etc. better whereas others will turn away from the mod because it will make the game easier. But to both our proponents and our opponents, we are obliged to offer you this message: we’re not done.

You see, we and Crate are presumably working toward the same goal. Presently, we’re just taking different paths. From what we’ve observed, Crate’s manner of balancing involves nerfing the overpowered into the ground all at once and then buffing the underpowered back up to par over time. Our approach is the opposite, and one only we – as modders, not the developers of a published product – are allowed to enact. We’re buffing all the ‘junk’ to levels which may be reasonable or far from it, and then tuning the absurd down accordingly. You’ve seen these actions by Crate and by us already; the popular discussion for Crate’s methodology at the moment seems to be centered around Devastation, Rend, and Shard of Beronath, three things which were nerfed. On our side, we’ve buffed Cadence to godtier status, Doom Bolt is reaching damage values unparalleled, and Bloody Pox has grown to be the subject of much debate as to whether or not its overpowered.

But, again, we are not done and chances are neither is Crate. Recent hotfixes on our side of things have seen Cadence being nerfed downward and downward toward more reasonable levels. This trend will repeat itself almost universally with time. To put this all into perspective, I’ve made some graphs comparing Cornucopia’s intended burndown toward ‘balance’ relative to that of Crate’s.

Here’s an exaggerated graph of how we each respectively handle a skill/devotion/item/etc. that’s overpowered, with 0.5 on the Y-axis being ‘balanced’ and the ‘Balancing Iterations’ being completely arbitrary.:

And here’s one (still exaggerated) of how we each respectively handle a skill/devotion/item/etc. that’s underpowered:

Notice a trend? Cornucopia tends to be erring on the side of over-poweredness until balance is achieved whereas Crate tends to be building from the ground up toward balance. There are a multitude of reasons for this and when adoomgod and I first started talking about this mod – before we even had a name for it – the option of which approach to take was a topic of significant thinking and discussion. Ultimately our reasoning is thus: we’re sick of seeing so many things being so vastly unplayable or, at the very least, not being fun. So, for the time being, we may as well make everything completely usable, if overtly broken. Crate can’t do this. They’ve sold a product to over 400,000 consumers and if everything was way out of whack their product would be boring and likely acquire substantial negative reviews; hence, Crate takes the ‘less fun’, harder, but equally respectable path.

As mentioned, ultimately, we do intend to wind up at the same place. If Crate gets there first, this mod is moot, and we’d have no qualms about that. Until then, my friends, look forward to the next release of Cornucopia. :wink:

Would it be a little arrogant of me to suggest that every change we made is in fact, not arbitrary, I’m going to go ahead and risk doing so. While the numbers we chose might be within a slightly arbitrary range based half on relativity to skills we feel are balanced and half on our own instincts and experiences, I think every choice we made was with intention, regardless of whether said choice ended up falling short of the intention or not. I can at least personally say that nothing I’ve pushed so far has been entirely arbitrary. Well anyway, on to matters of actual substance:

I really enjoy the topic of balance, and now that we’ve begun to do some solid work with the theme of “balance” in mind, I feel much better equipped to talk about it. First of all, what is “balance”? What does it mean?
(The first half of this post is just going to be on establishing a common definition and understanding to avoid semantics for the reader, the second half will be about our approach).

Let’s look at the definitions that I’ve chosen:

  • A condition in which different elements are equal or in the correct proportions.

I chose this definition because blatantly displays the issue: The ambiguity of the very concept. We know that in an ARPG, it’s not about making everything "equal of the game would be bland. Everything has the same speed, attack, hp, etc. It becomes a survival game where there are no advantages or numbers to play with, so clearly that doesn’t apply.

With “in the correct proportions” you have human interpretation. Unlike chemical equations which were established based on patterns in nature, the balance in games is and always will be, purely subjective.

One player may say, “my build doesn’t kill fast enough” while another argues the same build kills things too fast. This applies to every number/feature in the game. Enemies drop too many/too little legendaries, the game is too hard/easy, the base run speed is too slow, the feet on male meshes are too big/small. No developer can please everyone, nor is it good try. ARPG’s are balanced on the players becoming stronger than the opponents, and ARPG’s are about balancing the peaks and troughs in power you obtain. It’s also about balancing it so you can become powerful in different ways, and that each one of those “powerful ways” is more and less powerful in different situations and against different enemies.

Complicated? Well that doesn’t didn’t even begin to scratch the list of things a developer may try to “balance” in an ARPG. In many ways, every single number you see in the game, is being balancing in some relation to every single other number. Think about that shit.

And that’s why I write this because so often I see players type “this is too strong, this is took weak” without any real explanation. I mean it is helpful feedback because the developer can than look at the complaint themselves and decide if they agree with the perspective.

There is no right or wrong, even if you love or hate it. One build may kill a boss in five minutes and a player will complain it is too slow. Another player may complain that even a 10 minute fight would be too fast. Who’s right? Neither, the developer is. Because what a lot of players fail to realize, is (and especially with indie games) that it is the developer’s game. Regardless of what demographic a developer may target, game developers want to make something they’re proud of. Especially indie developers. They often have skill-sets that could enable them much safer consistent jobs, but they want to make something fun, an experience and a project of passion.

This brings up what I think is one of the single most important concept a player could learn in terms of taking their feedback to the next level and understanding balance: points of relativity. A point of relativity is exactly what it sounds like: a point/example in the game, that you relate other things to in order to discuss their “balance.”

For example, in another thread someone was comparing one celestial power to hawk, using falcon swoop as the point of relativity to claim that the other power was too weak. I countered with something like, “If you are using hawk as your point of relativity I’d argue it’s less of that celestial power being too weak and more of falcon strike being too strong.” This is because numerically, and in practice, hawk was out performing other constellations on the same tier by a large margin, on average. This is why we must think carefully about picking our points of relativity… and ofc it’s still subjective.

Another example would be on our balancing passes on doombolt. Doom bolt was tricky to think about at first because there was no real good point of balance. Raven’s lightning strike came to mind in terms of similarity, but that’s part of a whole other entity, it’s not player controlled, it has a stun, it’s really another ball game.

What it did compare to though was devastation, loosely. Even though they function differently, they are both skills that are contained within themselves (no skill modifiers, just 16 points to invest), and are both meant to be nukes on cooldowns. The problem being that Devastation hits in an AoE and over a period of time, where as doombolt is more of a single target nuke. You have to consider things like how devastation does have the drawback of not being able to do it’s full damage to a target if the target walks out of the AoE radius, where doombolt does it’s burst dmg all at once. Consider that doombolt is on a third of the cooldown. I also considered that devastation was vastly over-performing (It was more than doubling my dps while active, I could kill pretty much any boss within one cast of it in combination with my other skills). In the end we agreed that balancing doombolt as purely a single-target nuke was too hard (similarly we removed lightning strike from raven in favor of lightning storm) and maintained the theme of single-target damage but gave it the side-effect of spawning a pool of chaos upon killing the target hit. This allowed us to distribute some of it’s dmg in a damage over time AoE, which takes less pressure over balancing the burst dmg to feel valuable by itself. I stand by this as an extremely healthy choice, and that doombolt now feels like it’s in a very nice place, worthy of investment.

With all that food for thought said, I will now discuss our approach a bit, and while I think what Ceno typed was quite valid, that it may mislead people a bit on crate’s approach.

Let’s address the latter first: Grim Dawn was not planned to grow this big. By it’s original intention, it would have been “finished” a little over 2 years ago. Think about how far the game came in 2 years. That’s because at the time the ARPG market was kind of dead, but all of a sudden in mid GD development, you had shit like Van Helsing 2 and Diablo 3 coming out (which I laugh at now). Regardless of how poor those two games may be as “classic” ARPG’s, they had a lot of hype (especially D3) and pushed expectations for the ARPG market.

Medierra and team felt they had to improve the game way beyond it’s original pant-size in order it to stand up to other titles (god bless them). So consider that the game was supposed to only have 3-4 classes, and that perhaps the original max level was going to be lower. Yeah.

They had to take things they never originally thought were going to have to scale so high… and scale them that high. But that’s easier said than done. The more difficulty you add, the more content and levels etc. the more you begin to see that some skills, in how they function mechanically, are hard to make feel good from level 1 all the way to 85. I think Crate gets less credit than they deserve, because people don’t realize what it’s like to try to take something that was designed to be one way, and then try to rapidly expand it to be so much more. When the max level was 50, there was a completely different meta, and it would have been a completely different thing to scale everything to that “end-game.” As Ceno said they also have people actively buying their product, so they feel an obligation to deliver on quality through and through the whole experience.

Where as, the demographic Ceno and I are trying to appeal to is basically… Ceno and myself… and whoever decides they like what we like. We don’t have financial backing so we only have responsibility to ourselves. And we share a slightly different mentality than Crate.

Our mod is not for people who play the game once and stop, though they’re welcome to it. It is for people who specifically love trying different build combinations and making them work. Not making them perfectly competitive, no no, just getting them through ultimate well enough and pushing their limits numerically to see how far they can go. Our mod is for people who want to try and make every skill shine, for people who like finding hidden synergies, who like testing to see if a single legendary is impactful enough to completely change the face of a build (in rare cases).

We feel normal difficulty (including veteran) is mostly a wash. If you’ve played the game through normal even just 3 times, you’ve found enough farmed stuff that you can pass down to make the game much easier, and because of the lack of skill points and gear your character is never as interesting to play in normal as they are in ultimate. We are balancing around the end-game first, the mid-game second, and the early game last, if at all. I can’t tell you how many times I’ve said to Ceno when testing my Bloody Pox changes, “well it seems OP at low level… but I’ll see how it scales to 85… yeah it’s really shitting on early game content… but I suppose so does Devouring Swarm and DS does it even better with it’s life steal… okay as long as it’s not easier than using DS I suppose it’s fine.”

And so we go to our bottom up approach. Because, unlike Crate (to no fault of their own, they are EXTREMELY busy making content. It’s more time consuming than I think anyone out of the know imagines) we’ve both individually put over a thousand hours into the game, we more or less know the mechanics well, and even if there are play-styles we favor and play-styles we avoid, we could easily pick up what we’re not used to playing with and understanding the numbers behind them and how they compare to other things in the game because of the vast experience we’ve accumulated.

As we’ve said before, we were just so very very tired of seeing so much cool looking content in the game fall out of existence because it just didn’t scale into ultimate. This is why we felt it was fine to drastically tone up anything we felt was flat out “bad” even if it begins to border on OP. We try not to take it too far, but we’re fine with doing so. (take my first attempt on the meteor shower constellation, Ulzuin’s Torch… I need more feedback on it guys, I nerfed it by like 30% last time) For us, at this juncture in time, just seeing people pick choices in the game that they never picked before equates to a huge fucking win.

Every time I see someone post that they can’t go back to vanilla because they love the build diversity we’ve already accomplished in this early stage of the mod, my heart goes warm and fuzzy.

But the nerfs will come, and some Cornucopia players will whine… and I’ll mostly roll my eyes. This is not out of contempt, but simply the promise I make to you all that anything we lower is not to make it fall out of play, but to make sure it doesn’t feel like “the only correct choice.” And regardless of saying this ofc some people will like the nerfs and some people won’t no matter what. This is just a heads up that if you think we’re only going to continue buffing everything you’re… not going to be happy.

And one kind of comment I find grating is “don’t nerf that OP thing, just buff everything else up.” No. Fuck you. Fuck that. That’s not how good design works. If one single thing is steam-rolling content, I’m not going to raise everything else up to steam-roll content. Content is not meant to be always steam-rolled. PEAKS AND TROUGHS PEOPLE!

Take their falcon swoop change. Frankly I agree with nerfing it, I just disagree a tiny bit with the implementation. I think the reducing the weapon damage to 60% was totally fair, but the slight cooldown nerf doesn’t feel good. It hurts the “fun” of the skill being a constant spam of birds when there are other ways to tone it down. I like the %weapon dmg nerf, but I think I’ll also reduce 1 projectile from it’s high end scaling to lower it’s shotgunning power, but keep the cooldown low. But perhaps I won’t like how this feels and I will directly use Crate’s change >.>

I’ll re-emphasize Ceno’s final point: We get to do a more fun approach of just going absolutely crazy with shit in the game at times, and throwing it out to the public for feedback because we are not obligated to any form of quality control. This affords us much more room for error and change.

I also want to state one thing about something in the very late-game we have in mind for the mod: challenge dungeons/side-content balance.

We’ve had a brief discussion and agree that there’s no real reason to make the main-content of GD that hard. It’s story-line stuff and we want all players to feel like it’s not too hard to get through it. We feel completely differently on optional side-content. It’s side-content. You don’t need to beat it to beat the game. You want to play it anyway? Get ready for some ass-whooping brow sweating challenge.

While it feels like one of the new bosses in the hidden path 1.0.0.4 content is a bit bursty, I like that they’re trickier than pretty much all of the main content. I want to keep them that way, and in some cases make them make them even harder but have more counter-play. I want to make new dungeons with Ceno, with 2 real types in mind:

  1. Challenge dungeons that are like a puzzle with a slight gimmick. More difficult than normal content and requires you to build around their gimmick, but once you do they’re not that bad and very manageable.
  2. Challenge dungeons that require you build around them a little less, but are just straight up hard for as many builds as possible. Just difficult heavy hitting content that will push the player’s ability to play.

I also want them to be ultimate only in most cases :]. Crate can’t reasonably do this because… only like 3% of their player base really plays ultimate continually!

So uh… yeah I’m done now. Don’t worry so much about the nerfing stuff. While we are toning down stuff we raised too high, we’re still in a buffing phase over-all. After .3 is done, and not .31, i mean after .31, .32, .33 we will start looking at individual monsters and fights and also at the OP shit on player side and bringing them down to “really good.” Like the new devastation. Give or take 5-10%, it’s fine where it is. If you don’t feel the same way, you’ll be unhappy with things I do later >.>

Go on…

Soon™ tomorrow

I think the reason crate is always balancing upward is because the game had reached a relatively stable state of balance when Veteran was the highest difficulty available. Elite and Ultimate difficulties are relatively new compared to how long Normal and Veteran have been around, and remnants of veteran’s balance are still visible to this day (blackwater cocktail and stun jacks receiving heavy nerfs which still affect their viability). Some older items haven’t caught up to recent changes yet, like the movespeed and attack speed cap which were recently implemented. On the other hand, ceno and doomgod have been balancing with ultimate in mind, whereas Crate are only recently beginning to do so. It’s good to have both approaches, and hopefully good will come from both as well

Updated my post.

For shizzle. We’re always excited for Crate’s changes because we KNOW we’ll probably like a good portion of it and incorporate that portion directly into the mod. (#lazydevs, crate does the work for us in some cases)

Well, to be fair, that’s only an issue if the buffs are player-side only. If enemies were also buffed to match when everything was brought up to the level of the overpowered thing, the result would be the same as a nerf to said overpowered thing. The only differences being the more positive psychological effect of buffs vs nerfs in general, the slightly more refined gradient that higher numbers would allow…and the completely unreasonable amount of work it would take. When the choice is “nerf one outlier” vs “buff literally everything else”, I guess it’s a bit absurd to suggest the latter. Especially since such a method would almost inevitably just create new outliers anyway.

The only thing i want to see change is Trozan’s Sky Shard skill if possible make the middle projectile hit where u aim and the two other ones random since sometimes they all miss where u aim which is pointless.