Categories

«13456717

[Dev Blog] On Keywords and Omega Items

GazillionLocation San Mateo, CAPosts: 1,445 Staff
edited September 27

Hey everyone!

It’s been awhile since you’ve seen a Mayhem-brand giant multi-page design insight post, and one is long overdue.

Today, I’m going to talk about two projects that I had a heavy hand in pushing for and designing that have come under scrutiny, and by the end of this post, hopefully you’ll have some insight as to why the decisions were made. These two features are the console’s Keyword System, and the Omega Items system.

Keywords (a.k.a Tags)

Since the game’s inception, the design of power and item progression has always been limited by the current PC “damage type” system - Physical, Energy, and Mental. The primary problem with this categorization is that it constricts the Marvel IP into three categories and prevents any overlap between them.

For example, we classified magical-based damage as “mental”, even though there are valid arguments that it could be something else (just because it’s magical fire, shouldn’t it be Energy damage like other Fire skills?). We classified some attacks as either Energy or Physical strictly based on whichever one we felt would be better to itemize for if the attack in question could ride either line, which meant that the theme of the power really didn’t matter.

The problem with this type of system is that it also overlapped with the existing Keyword (tag) system, such as the Summon or Area keywords. These keywords allowed for the few instances of ‘build around me’ power sets among heroes. Collecting items to support a build and figure out a way to make it work is one of the most fun aspects of a loot-based RPG, and those keywords helped to define ways to do so.

What we found when analyzing the item system for console is that we wanted to explore this space even further - why shouldn’t there be items that allow for someone to make a Gun-based build, or a Fire-based build, or a Magic-based build? These builds usually existed already in the heroes themselves, but the only way to build an item that supported one was through a hand-crafted unique that talked to the individual powers in question.

For example, if we wanted to support a standard-issue guns and bullets Punisher build in an item, we’d have to call out each power that we felt fell under that category and buff those powers in some way. This led to the inevitable affix bloat found in the PC Unique items from the ‘power rank’ days, which became overwhelming for new players as they tried to digest 10-12 lines of text on some of those uniques. It also required that hand-crafted items for every feasible build be made for every available slot, since talking to powers had to come on an item exclusive to that hero (any-hero items can’t talk to individual powers, since we don’t know who is going to equip the item). Keeping a regular cadence of those items was a massive undertaking, and we could never keep up with the demand regardless of how many designers were implementing them simply due to the sheer number of heroes in the game.

Enter the Keyword system. We already had all the tech in place to do keywords like Area or Summon, we just needed to create new ones, add them to the appropriate powers, and make items talk to those keywords. An affix giving bonus damage to Explosive powers would allow for players to collect those and make a sweet Explosive build on any hero who had enough of those powers to make it work. An affix that triggers a bonus effect on using one Explosive power could shift a player’s build to make them include just one Explosive power for the bonus.

It was clear once that objective was set out for that we no longer needed the Physical, Energy, and Mental damage types. We had considered a different approach at first - adding those keywords to the damage type lists instead, so you’d deal Fire damage or Explosive damage. Unfortunately, that system only supports one single setting, so if we went that route, we’d be severely limiting the design space available to us. For example, what type of damage should Dr. Strange’s Flames of the Faltine do - Fire or Magic? We’d have to pick just one, and that would restrict the number of items available to Dr. Strange players.

Keywords were clearly the winning case for design space, but it did come at the cost of clarity as players need to have an understanding of what keywords they have access to. We tried to make them as clear as possible in relation to the theme of the power, such as Storm’s powers being Lightning if they use Lightning, Wind if they use Wind, Ice for the Hailstorm, and multiple of those keywords if more than one element was being called in the same power (such as her Thundering Maelstrom power).

After launching the new items system, one thing we all believed we could do better at was to make those build paths easier to see and collect. As randomly rolled affixes, it takes a great deal of time and luck to get a lot of items that are talking to the Keyword(s) you are looking for.

Thus...

Omega Items

Originally, Omega Items were proposed as mostly randomly rolled affixes, with some having power-specific interactions. Admittedly - we jumped the gun on announcing this before the project was fully scoped out. That was our mistake, and we’ve since reacted poorly by not communicating the iterations and pitfalls that were found. I hope this post can make up for that a bit.

The problem with the original proposal was that it was simply too large to create an interesting, diverse set of affixes due to the nature of power-specific interactions, and what it takes for us to implement those power-specific interactions.

Like Uniques, these affixes would have to be hand-crafted for each hero. At first, we’d outlined only 5 per hero, but this quickly was proven to not be enough, as if any power was buffed by your affixes and you had opted not to use that power in your build, there were no other avenues available to you. 10-15 became the new standard, with that amount needing to grow over time - again, for each and every hero. We’d have to patch in another 2-3 affixes per hero on a regular basis in order to keep things interesting, or the item game would become invalid once you’d completed collecting the ones you wanted. It was not a sustainable model and would have put us in a case of needing to overhaul the system not that long after releasing it, which was not a position we wanted to end up in. It was the exact same pitfall that we had with Uniques in the first place, and we knew it wasn’t sustainable.

Around this time, we’d already been discussing the Keyword system and debating its benefits and detriments. If we went ahead with the Keyword plan, suddenly the landscape would change for Omega Items. We could focus on Keywords, which are abundant and not specific to any one hero, and create interesting items around those keywords. We could now come up with a great “Throw” item affix that lets players create a Black Panther knife-throw build or reinforce a card-throwing Gambit build, or a sweet Wind item affix that really impacts both the Storm Wind powers and Star-Lord Air Elemental Gun powers.

Using Keyword-driven affixes instead of hero-specific power affixes was a much more sustainable model that we could grow over time in a reasonable time frame once the feature had been launched. It would get us a unique item system that played into hero builds without the overhead and scope problems the previous feature had. We began the Keyword conversion and put Omega Items on hold until we’d finished the pass.

The new ‘Omega Items’ looked very similar to the previous model, except the affixes now spoke to Keywords instead of individual powers. One problem that was immediately recognized was the stacking of those items. A tech limitation of our item system is that multiple equipped items of the exact same type do not stack their benefits under certain circumstances. A proc effect with a cooldown, for example, is not written ‘onto the item’, but rather onto your player. All powers stem from a character. Thus, the cooldown for that proc effect is tracked on the caster of the power, not the item that granted the ability to use it. In short - equipping multiple copies of that proc won’t make the effect trigger more often than if one copy was equipped.

Since some of our designs (both in the old system and the new system) called for interesting new procs with powerful effects, this was a serious problem. A player wouldn’t know to not equip multiple copies of the same affix if it fell under these conditions, and we’d need to figure out a way to solve that. We thought of restricting the affixes to only roll on specific slots, but that further restricted the build potential for each hero, and would force us to make even more affixes to help cover those blind spots.

The solution was to simply make all of the affixes fall under the same ruleset - if we couldn’t make every case work when stacked (and we couldn’t), we needed to make none of them stack, so we could make that functionality very clear to the player, and preserve the balance of the items.

In doing so, the engineer working on that request pointed out that this was the same functionality that prevented us from doing Set items when it had been pitched once or twice in the past. Items have always conveyed their listed benefits to you, and making that no longer true under certain conditions (such as having multiple of the same type equipped) required new code implementation. With just a little bit of extra work, he could add an entire Item Sets feature.

With that in mind, our design space blossomed into what became the final feature for Omega Items. Suddenly, instead of making extra duplicate affixes “dead” and unusable, we could make them stack into new, additional bonuses!

This was a dangerous road to travel though, as we know Sets are loved by some and hated by others. There are two major pitfalls associated with Sets that we set out to avoid when designing them:

  • Not having so many that you can’t collect them all without an extreme amount of inventory/stash management and space to do so.

  • Not making it feel like you had to collect an entire set before it started to work out in your favor.

To help support the first point, the loot tables for Omega Items are constructed per-hero, so that Wolverine will never see the Dormammu’s Embrace set, which is all about Fire powers. Wolverine doesn’t have or care about Fire powers, so he never needs to see that set. He’ll never see anything regarding Intelligence, Energy, or any other stat that he never could feasibly use. This reduces the number of available sets per hero down by at least 50% from the total sets available in-game, and in most cases, reduces by more than that. NOTE: This is claimed to not be functional in the current console patch, but will be fixed as soon as possible (long before it hits the PC build).

To help support the second point, we budgeted the affixes on each tier to be equal - the ‘item power’ granted by a tier 1 bonus is roughly the same as one granted by a tier 5 bonus. And, each and every Omega Item has a Tier 1 bonus. Some can argue that on some of the sets, the best affix is the first one in the list - the one you get immediately just for equipping the item by itself. In many cases, the first affix of the set is one that is usable by the widest number of heroes, such as a triggered effect on a cooldown that happens when you use a certain type of Power. The affixes further down the tier usually talk to a more honed, focused build, such as one that almost exclusively uses Powers of that type.

What this means is, your perfect desired build of Omega Items may very well be 5 different sets’ Tier-1 bonuses, and that’s an absolutely acceptable choice. Some of the best-in-slot builds that are discovered may use five different Tier-1 bonuses. Some may use a 2-2-1 setup, or a 3-2, 4-1, etc.

Furthermore, the growth of this system is now modular. What I mean by that is, perhaps down the road, we will allow Rings to roll at Omega Rarity, pulling from the same set pool. Now, 2-2-2, 3-3, 4-2 setups open up as possible avenues since there are now six available slots instead of five. We could add a Royal Grandeur-type item which bumps you up a tier on all equipped sets, further increasing build options. We can do these things because the system is open-ended and not hero-specific, so if we were to make either of those changes I described, every single hero would be affected by the new options, without us needing to hand-craft hundreds of affixes when we want to vertically scale the system to offer more power. It’s flexible and scaleable, and something we can reasonably grow for years with the implementation capacity we have available to us.

Conclusion

I know these types of changes might be hard to process, but I hope this post has brought you some clarity on these decisions - why we are pursuing this route, and what it will do for our future in opening new build paths and new itemization to you, the players, faster than we could before.

I’m really excited for these items. There’s so many affixes that are build-defining, from ones that will skyrocket stats into the stratosphere, to ones that reward the most veteran players via Infinity points, to ones that will focus players down new thematic paths. We have many more planned, and hope that they end up being satisfactory once they are finally in your hands.

Until next time!

Powers Designer of Many Faces - https://twitter.com/MikeMayhemDev

Posts: 4,092 Cosmic

Can you talk a little about the transition between the old stat system and the new stat system ?

Thank you for this devblog.

Total hours played on Steam in 2016: 17 474 560,2
Total hours played by the whole community in 2016 according to gaz : 21 517 247
17 474 560,2 / 21 517 247 = 0,8121187715138465

Posts: 168 Incredible

This was awesome insight. Thank you.

Location Buff Psi CrushPosts: 1,531 Fantastic

That's what I saw...when the Omega items were previewed by a community member from the console forums. As such, I felt hopeful.

Just don't muck it up with hidden ICDs, bugs, opaque proc scaling, and level-1 omega prestige, and this might actually work out.

Thanks for the post. If it was split up, it would look a lot like monthly status updates.

Phoenix PoweredPosts: 1,011 Fantastic

@MichaelMayhem said:
Today, I’m going to talk about two projects that I had a heavy hand in pushing for and designing that have come under scrutiny, and by the end of this post, hopefully you’ll have some insight as to why the decisions were made. These two features are the console’s Keyword System, and the Omega Items system.

I know these types of changes might be hard to process, but I hope this post has brought you some clarity on these decisions - why we are pursuing this route, and what it will do for our future in opening new build paths and new itemization to you, the players, faster than we could before.

Thank you very much for this post, and I personally found it easy to process. Being a programmer myself, I can see the many benefits of the design you described to both the designers and the players. I look forward to seeing this come out with November's Omega update.

Askani'sonLocation Austin, TXPosts: 2,564 Cosmic
edited September 27

Love me some @MichaelMayhem!

My only question is what kind of tags will Cable's psimitar powers fall under?

Location GibraltarPosts: 2,951 Cosmic

Thank you, Mayhem.

My super fantastic custom sig
True BelieverPosts: 495 Amazing

I like it. Finally some explain and I understand. Everyone are waiting for the "End of Uniques" to Omega items... this explain why take too long.

Still not "full happy" with the keywords... but I will fine with the classic "no-keyword" damage type, to ones who like make hybrids builds... But if the "not used" keyword or even Omega powers will not show, will be better.

But I still want to know what will happen to the Uniques "don't toucheds"... because a lot of people are farming this ones to go to Omega... and without "no changes" they will became useless? Os their "not used anymore" keywords (Physical, Energy and Mental) will still works, even in background?

«13456717
Sign In or Register to comment.