Discord Dev Chat – 1/20/2017

Time again for another Discord Summary.  You can join the Rift Community discord here: https://discord.gg/BQJnPQ6

Since conversations like these can often be difficult to read, follow, and review, I will do my best to summarize several discussion points from the devs on Discord.  Again,  I will not call out dev or player names here.

The purpose of this post is to summarize the general thoughts of Rift’s dev team.  As usual, all details for future changes are subject to change, and should not be taken as 100% guaranteed to happen.

Shoutout to Hepatitis for helping to compile portions of this!

Q: [How do you balance specs for PvP]?

we actually have a reporting tool that tracks any given spread of a spec on an elo (which we reset weekly). So every spec that gets played gets added to that list at exactly 1500 (starting elo), and then the Spec’s elo is tracked as players play, regardless of who plays it. We also see how many people are using the spec.

So a spec that has a lot of players and is bottom of the barrel means that the data is relatively certain and that in comparison to the other specs appearing it performs worse in the time frame.

 

Q: So even if people find an op thing and don’t talk about it on forum etc, you can still track it right?

Yup, It also means we can sift out the hyperbole.

For example, in EU this week no spec was more than 60 points from 1500 (which in terms of elo is statistically insignificant), [meaning] EU was pretty balanced overall.

NA had a single spec that was about 180 from 1500, everything else was within 60

So when the forums complain about “I’M CONSTANTLY GETTING ONE SHOT” the data says:

1) no you aren’t. It happened a handful of times and you remember it because its jarring

2) that isn’t contributing much to their chances to win a match even if they were doing it as often as the forums say

 

Q: What’s the hype about Primalist? people saying they get one shotted and us prims saying it takes a lot of ramp up to reach that

Anything you got planned for that?

The answer is complicated.

Burst specs are balanced around their ability to ramp up, do a bunch of damage, and then cycle down. This works fine in PVE where fights can last long enough.

But in PVP is dangerous if the damage is high enough

Primalists tend to load all that damage into a narrower window. and just nerfing their top end burst makes them worthless so it’s a really hard problem to fix overall

 

Q: So you basically need a find a way to deflate the burst and even out the flow of continuing damage without nerfing overall DPS?

[that] AND not changing the overall flow of the soul

which happens to be the central problem [for Maelstrom].  We could just put a 30% pvp nerf on vaporize, and then buff up scald, eb, flamewalker, SP, by 15% in pvp

but that changes the pacing and feel.

 

Q: [So it isn’t easy?]

I mean, everything you [players] think should be easy is almost never that way. It’s not surprising we say that so often.

Especially because many times it’s when someone is suggesting something that isn’t simple at all and is being incredibly reductive with how much work goes into making games.  Which, were we less composed overall, would be somewhat offensive

If you have ever heard the phrase “robbing Peter to pay Paul”, that’s pretty much what game development is. If we put resources towards one thing, something else suffers.

A good example would be the Timewalking thread from like…a few weeks ago or something?

[That being said], This part is important – Don’t stop giving feedback. It still serves to inspire us in ways that we can accomplish the goals. In the words of Commander Quincy Taggert, “Never give up. Never surrender.”

 

Q: [Why not just use bolstering?]

“Just use bolstering” totally misses a ton of important points about why systems like that work ever.

 

Q: Can you give an example of “ton of important points”?

Keyens: Ok, I do so love this part. Just for a few

  1. It needs rewards that fit into the larger schema of the world and those have to be implemented and handed out in some fair and measured manner
  2. Bolstering doesn’t work anywhere outside PVP, so it would need work to move it.
  3. Sidekicking doesn’t update dynamically as we add new vectors of player power or targets shift.
  4. The dungeons would then need a pass to account for new or adapated tools players have.
  5. The system can’t just replace other content unless we are intentionally orphaning it.
  6. It splits the playerbase up in a way that we then have to account for in future updates and content pushes.
  7. If we ever change the gear targets we have to reitemize it.

There is your baker’s half dozen of things systems alone would have to think about to even consider some kind of “play old dungeons at relevant power” system that I could think of just sitting here

 

Q: But it would be cool if we could somehow play old content at a challenging level, right?

Certainly, but just from an eyeball of those things, that’s several weeks of work to build and those weeks can’t be spent on other things and it would require support from other divisions Like engineering and content.

But that doesn’t answer the other problems that arise there. It permanently segments the queues players are in, It’s now another reward stream we have to manage and continually update, it has to mesh with the progression systems we add in the future. It adds a ton of work we have to forever maintain down the road too. it’s not always just the up front cost.

 

Q: ok, i got it that the pvp bolstering system would not work for pve.  if the hit buff is working… why shouldn’t it work if you give 20k str,dex,int,wis to every char. (caster would just benefit from int and wis etc.)?

Just want to understand why HIT is possible and other stats are not.

Stats aren’t the only vector of power that has to be accounted for. And a buff that gives +flat stat is wildly different than scaling your power down to what it was at X point in time.

 

Q: does it have to be a new reward system? couldn’t it just be a “legacy expert queue” with same rewards as the current expert queue?

then time to complete becomes a huge issue we have to worry about. People will only grind the easiest thing available when it comes down to doing something repeatedly aren’t we already doing that anyway? skip all trash and kill bosses asap?

Yes, and if older dungeons take longer or are harder, nobody ever runs them, wasting all the time we spent working on it [See: Rhaza’de Canyon].

 

Q: then merge them into one queue or make it like pvp: have a pool of some dungeons and rotate every 2 or 4 weeks

People just requeue when they get one they don’t want. [Again, see Rhaza’de Canyon].

Intrepid is generally simpler. Trying to scale power town down to a specific target is one of the most difficult design processes full stop.  In terms of production time, it is MUCH easier to up level something than it is to bolster folks down to a target level.

 

Q: we got IDD and we’re getting Gyel. Is there a chance that you can give the player base a poll to choose a possible next intrepid dungeon?

I would guess no, because not all old dungeons are created equal

 

Q: if you internally say like “okay, we could do dungeon A, B or C next”, couldn’t you ask the player base to choose between those?

That’s a trap in most cases, because those polls devolve into “WHY DON”T YOU DO THIS ONE”.  We did [a] poll when we redid Deepstrike and it won. Course, we also had some bad actors skewing the poll too

 

Q: Any plans to create new low level chrons/dungeons?

Making dungeons or chronicles are similar development time sinks, and lower level ones don’t make them any cheaper. It would be an incredibly expensive time spend for a part of the game people spend a super tiny amount of time in.

But at that point why not just make 70 dungeons with that same time, and encourage people to play in the same place, rather than creating more different queues

 

Q: What about reworking older zones?

the TL;DR on touching low level zones is that the time is almost always better spent working on other things.  Remember that any given dev isn’t interchangeable with another

Keyens: My time working on souls/balance can’t just be converted to making dungeons, [and] I don’t have the skills on our editors to do that. [In the same way], artists can’t help make more souls :p

 

Q: Why not just allow people to mentor and queue for all of the lower level dungeons? New people wouldn’t feel like the game is dead, and veterans would have more content to do.

because time to complete and path of least resistance becomes a huge issue. If we put relevant rewards on them, people will gravitate towards only doing the thing that is fastest. Then we get in to fixing sidekicking to account for every power spike in the game, which is a very complicated problem.

 

Q: but then why not make the force mentoring 2-3 levels below to make it less easy?

because that doesn’t solve the problem. The core issue with downscaling people is that the L70 is not a static number. So we are trying to make a moving value move to a static value. As your gear increases, you get more power points, more passives, more anything, your power changes.

So we either have to constantly update the scaling rules (which is a horrible use of dev time) or build a completely new system that lets people scale more accurately (which is incredibly expensive)

Even just max level at current has a huge gradient, as gear goes from l70 green to full expert purple as a scale, with fragments as well.  [Given this, should] we use a full expert 15 fragment person? Anyone less than that is too weak for the content.  If we use less than that, then people eventually vastly overgear it and the same problem we have now arises

 

Q: So, out of curiosity, how does Final Fantasy 14 do this “downbolstering” for dungeons?

FF14 is built in a way that if you “LFG” queue for an older dungeon, then it down-levels you and your equipment to that dungeon’s level, but it never feels like you’re overpowered.

They probably spent a crapton of time developing a system specifically for it

FF14 is built with that entire premise as a core part of the game, because dungeons and raids are mandatory as part of the story quest which is what lets you move forward in areas

They also almost never add new power vectors for players. They also turn off abilities you wouldn’t own when you down sync

 

Q: I guess I’m looking to understand what more can be done to quickly provide content for l70 players, even if it meant faceroll old dungeons as a daily.

Then to answer your question, there isn’t really such a thing as quick content.   Intrepid Gyel is about as fast as stuff could be added. Content is complicated and expensive and even with a ton of testing it can break

 

Q: what you are basically saying is, that we should hand them fluffy in tutorial, make them 60/65/70 and throw them asap at end game content, because that is what everybody wants?

Keyens: It turns out (at least in my experience on MMOs) new players flowing into the game only want to get to the newest and shiniest stuff as fast as possible. Back when I was with cryptic we spent ungodly amounts of dev time on new player experience and tutorial and levelling content to update, and we noticed zero change in new player dropoff or behavior.

The best way to keep new players (as far as the industry has found so far) is rapidly push them to the same part of the game as everyone else so they can play together.

 

Q: What kind of reasonable incentives (if any) could be given to people to get on and test more often when it’s needed? At least in regard to stress testing things like Sieges.

Testing incentives is a complicated thing to do for a lot of reasons:

First, we don’t want something like that to add power because then its mandatory (which isn’t a great state to be in).  However this immediately excludes a large set of people who would test.

Second, making rewards does take dev time.  So does rewarding them and tracking who tested (all of this is actually quite expensive).

Third, even with incentives, not enough people would be on to reveal volumetric problems that come with live servers, so it largely wouldn’t reveal the kinds of problems that cropped up with sieges and the LFR queue.

 

Dev to players: out of curiosity, why do you think that isn’t difficult? I genuinely want to understand why people think “Make a new thing you only get from X” is a simple thing to implement.

I’m genuinely curious how long people think things take. I’ll even simplify it down to something I can answer with 100% certainty:

Q: How long do people think it takes to make a new soul from the time Vladd says “Keyens, we are making a new soul, start now” to “Its on live for use” Just a single soul?

Player estimates were all over the map, from  8 weeks to 4 months

The actual answer is about 6 weeks from starting paper design to its in a pretty much ready to launch state. Although this has some variance based on how well the mechanics work from the paper to first implementation stage.

There is a caveat however: thats 6 work weeks where no other tasks are assigned to that designer.  FX and Audio don’t take nearly that long obviously, but they work on other stuff all the time, so their time tends to be booked up

 

Q: 6 weeks design, and 2 weeks each for sound and fx?

The 2 weeks of work that sound/fx do can happen during that 6 weeks. So depending on a work week, that’s ~240 hours. In total that is 10 weeks work cost wise.  [that’s] almost 3 man months of work per soul just from a human budgeting perspective.

 

Q: with the average workload of a soul-designer-dev, how long would those 6 weeks become?

Keyens: with other tasks I generally manage (and I only budget for myself because I know my implementation speed well) is double an estimate. [so 12 weeks]

This accounts for “oh god the game caught on fire fix it now” time and “that mechanic can never work, gotta rebuild some core elements”

But yea, the TL;DR of all this is that Systems Designers (what Vladd and Keyens are) aren’t just idea people. Ideas are easy, and everyone has a few good ones. A systems designer is effective at also considering consequences and actualizing on a nebulous desire (such as “We want to drive players to do X kind of content more frequently”)

 

Q: [Can you give an example of things being done to push players to do X kind of content?]

LFR is an example of pushing people to do content they might otherwise not have done. It removes some barriers to players who might not have engaged in that content before.

 

Trivia time!

LFR is probably one of the largest, if not the largest and most complicated system that we have ever made in RIFT.  Planar Fragments and the current Wardrobe system areclose seconds.

Original internal title for Warchanter was “Pragmatist”

“Mystic” is the working name for the Primalist support class

There was a Lunar Lancer soul floated around to mirror the Trove class

 

2 comments

  • Q: What kind of reasonable incentives (if any) could be given to people to get on and test more often when it’s needed? At least in regard to stress testing things like Sieges.

    Testing incentives is a complicated thing to do for a lot of reasons:

    First, we don’t want something like that to add power because then its mandatory (which isn’t a great state to be in). However this immediately excludes a large set of people who would test.

    Second, making rewards does take dev time. So does rewarding them and tracking who tested (all of this is actually quite expensive).

    Third, even with incentives, not enough people would be on to reveal volumetric problems that come with live servers, so it largely wouldn’t reveal the kinds of problems that cropped up with sieges and the LFR queue.

    Dev to players: out of curiosity, why do you think that isn’t difficult? I genuinely want to understand why people think “Make a new thing you only get from X” is a simple thing to implement.

    “Just recolor an existing mount, any one that has “glowing” effects and REWARD it to anyone who does whatever you want them to do. Example: make a “glowing” GOLD color Ki RIn mount, make a “glowing” GOLD squirrel mount, ETC.

    Is that really that hard??? JUST CHANGE THE COLOR OF AN EXISTING MOUNT!”