Players Will Optimize

I remember back to the early-ish days of TorilMUD, back when I was first getting into groups to do zones.  Doing a zone was akin to what we would call raiding now, where a max size group, sixteen characters total, would set out to fight their way through a series of rooms and bosses, culminating the in the main boss of the zone.

Specifically, I remember doing the City of Brass zone.  It was a popular zone to do for quite a while, one done almost every boot. (See an old post about how MUD crashes were a good thing back in the day.)  It was an older zone, it wasn’t too big, there were a couple of possible drops for class quest items so somebody was always keen to go, and the general loot was pretty decent if you were just starting off doing zones.  There were upgrades to be had and everybody wanted that flaming halberd for an alt.

Back then the approach to the zone was slow and plodding.  Once through the Plane of Fire (you needed flying gear or the spell plus a fire protection item to go on this run) , the group would assemble and prepare outside the first room.  Once spells were up the tanks would roll in, engage the mobs in the room, call everybody else to come in, and we would unload everything to clear the room.

We would then sit down, mem up our spells, then stand when we were done.  When the call “spell up” came again, we would hit all our targets… as a druid I would cast vitalize, a hit point boosting spell, on some of the non-melee characters and maybe barkskin on the tanks and anybody who requested it… then the tanks would move into the next room, call us to come in, and we would burn down the next room.

It was rinse and repeat, taking down every room as a set piece battle.  At boss mobs we would get special instructions.  When spell feedback was introduced, a mechanic that would damage players if two people cast the same area effect damage spell at the same time, there would be some coordinating of who would cast which spell first.  But otherwise it was the same thing every room, and it stretched out the time it took to run the City of Brass into a three hour event.

But as time went along the runs began to speed up.  First, the overall quality of people’s gear began to improve.  This made players more effective at slaying mobs as well as surviving fights.

Then there started to develop an ideal group composition.  For example, whoever was leading the zone would never take more than one druid unless there was an empty slot that they couldn’t fill.  They wanted a caster who could do the “dragon scales” spell on the tanks rather than the lesser spell “stone skin.”  There were classes with buffs that were deemed essential for a run.  Getting the right group comp made runs go more smoothly, especially at boss fights.

And the zone itself became a solved problem.  The efreeti never changed.  There were a couple of random spawns, but otherwise how to do the run was well understood.  There were no surprises, a well defined route existed, and the boss mechanics were old hat.

Finally, there was a big change in how zones were run.  Groups stopped doing each room as a set piece battle.  Clearing up the trash mobs on the way to the boss was now easier due to gear upgrades, so we would roll through all of that with the various casters just keeping critical buffs up, refreshing them at need.  To sustain this, the concept of “mem out” was introduced, where the raid leader would call “mem west” or the like to indicate where the casters could move to refresh their spells while the battle was still raging.

The latter kept everybody busy.  Rooms with trash mobs took marginally longer without everybody blowing their whole catalog of damage spells, but that was heavily outweighed by the reduction in pre and post battle activity.  Only boss fights got the big “spell up” treatment. The time to run the zone, from starting out in to returning home to Waterdeep, approached an hour if everything was going right, and it almost always did.

That is a pretty big speed up compared to three hours from the doorstep of the zone.  And the time improvement didn’t stop there.  TorilMUD, around for more than 25 years at this point, has never raised its level cap.  Instead, it has maintained some semblance of stability by adding in new, harder zones for those at the level cap while re-balancing equipment over time with an eye towards keeping most level 50 zones viable.  That generally means any gear that seems over powered, like the glowing crimson dagger or the haste enhanced grey suede boots, are likely to get a nerf sooner or later.

Still, even with that optimization happens.  Old hands who have run a thousand zones have a bag full of gear so can pull out a set perfect for each task.  I bet if I told long time zone leader Lilithelle I needed something from the City of Brass today, she’d throw together a group of eight to ten people and drag me along, finishing the zone in 30 minutes or less.

As the kids say, “Cool story bro.”  But what am I getting at here?

This is what happens to content over time.  Player optimization alone pretty much cut the run time for City of Brass by two thirds.  And that three hour number was after the “learn the zone and the bosses” part of the process.  Add it some of the usual gear inflation and that time is now down to one sixth the original time, and doable by half a raid group.

This is what happens to content over time, especially PvE content.  It becomes a solved problem.  Players learn how best to assault things and share that knowledge.

Sometimes that is okay.  In MMORPGs where expansions use levels and gear to gate content, it is pretty much expected that older content will be made obsolete.  Often, after enough time has passed, old raids become solo projects that people run to collect gear for cosmetic reasons or to fill out missed achievements.  That is certainly the accepted state of affairs in World of Warcraft.

In other games it can be problematic.  In EVE Online optimization is an ongoing battle for CCP.  Without levels as a gating mechanism any new PvE content is pretty much solved immediately.  So, despite there being something like four thousand NPC missions in the game, the PvE is generally considered boring and is subject to pretty extreme levels of optimization.  This goes especially for null sec anomaly running, where titans are the latest high yield ratting option.

Only the Abyssal deadspace content isn’t completely solved, and that is only because it has a random aspect to it.  Once you start one you are committed and cannot go back and refit if you have chosen poorly.  And even that is only an issue for the level five runs.  CCP last said that the percentage of Abyssal deadspace runs that ends in a PvE death is very, very low.  I cannot find the number at the moment, but 3% springs to my mind.

Then there is PvP content in New Eden, where The Meta constantly strives to find the optimum ship for given circumstances and CCP is constantly tweaking ships in order to try to bring balance to the force, only to find that suddenly every big alliance is now focused on a specific hull for its main doctrine while the small gang and low sec forces have a new favorite of their own.  And then there is suicide ganking in high sec.  That has become one of the few PvP solved problems at this point, something CCP needs to shake up somehow.

And so it goes.

The thing is, a game’s core player base will always optimize.   But outliers and new players tend to get left out of that.  If a studio focuses only on the core, a game can become impenetrable to new players.  But if you don’t focus on the needs of the core your most loyal fans may get bored and walk away.

8 thoughts on “Players Will Optimize

  1. anypo8

    There are two standard solutions to the PvE staleness problem. Both were tried pretty successfully for MUDs, but MMOs haven’t made great use of either.

    1. Sophisticated procedural (computer-generated) content. Minecraft is probably the modern poster child for this, but roguelikes are where it’s really at to this day. The best ever is probably Dwarf Fortress. As you said, EvE is doing a bit of this with Abyssals, but it just isn’t complex enough and is thus still generically solvable. The fact that The Maze is not procedurally generated astounds me to this day: it would be so easy, and would drive folks plumb nuts.

    2. User-contributed content. Again, this was a common thing in MUDs, and Minecraft makes great use of it. This is not an easy thing to get right, because user-generated content has a host of issues: Asset generation and management is expensive, there are sometimes IP problems, users can be malicious, poor out-of-the-box QA. EvE doesn’t really do this, but the huge and sophisticated third-party software stack built around the EvE API (also WoW’s) is evidence that it could work. Signal Cartel’s *A Dead Parrot* wrote and runs a web-based “Allison” personal assistant that CCP should just buy and adapt. It’s really amazing.

    If I owned an MMO today I’d be investing a lot of money in a solid PvE editor and adding some mechanic that allowed users to experience their hand-built content in the game world. I’d then add in-game rewards and big fame for the players who contributed and moderated that content as it was added to the “official” game. Ideally the editor would have good procedural generation facilities so that user-created entries could have randomized content. It would be a huge and expensive project, but it seemsto me like the way forward in a world where PvE content is so expensive to make and maintain right now that new MMO expansions are rare, limited, pricey and go stale quickly.

    Like

  2. Azuriel

    That’s a great write-up.

    It’s kind of a tautology to say, but optimization cycles are self-reinforcing too. If one method is faster and less risky, then everyone following it will be rewarded faster and more consistently. Anyone who was fine with the old system will be left in the dust, even if the viability of their class/strategy has not changed. And since these are social games, one cannot simply pretend the meta does not exist.

    Like

  3. Bhagpuss

    I think all older content originally aimed at groups and raids in PVE MMORPGs should be designed with the specific intention of encouraging smaller groups, duos and solo players to use it at later stages of the game’s life. This is content many of those players won’t have seen when it was current high-end and it will absolutely count for them as bona fide new content as and when power creep and optimization brings it with in reach.

    Some of the best times I ever had in EverQuest came when Mrs Bhagpuss and I hit the point where we could duo stuff like the Planes of Power, Gates of Discord and Omens of War expansions. We came back to EQ a few expansions later and once we’d levelled and geared up we spent almost six months doing old group and raid content. Similarly, some of the most fun we ever had was when we outlevelled Mistmoore in EQ2 by enough to do a full dungeon crawl there. I’m currently using my insanely overpowered Berserker in EQ2 (upwards of 60 million hit points now) to do old raid zones I’ve never seen before, which is great fun and also very useful for some of the really nice-looking armor and weapons that I can use for appearance, and the vast amount of Heirloom raid gear I am going to be twinking my now-level 88 Bruiser with once he hits 90.

    As for anypo8’s suggestion for player-made PvE content, that sounds almost exactly like what Neverwinter did, down to the rewards to get people to use it. Tipa made some nice EQ -themed dungeons that I tried. Other MMOs have done something similar too. The problem is, players tend not to take content made by other players as seriously as they do “official” content and there are always complaints that the developers are greedy and/or lazy in getting players to do their jobs for them.

    Like

  4. Shintar

    Ever since SWTOR introduced level scaling, all endgame content pretty much remains relevant to players forever, whether it was part of the launch package or introduced yesterday, and it’s fascinating to me how people still find new ways to optimise it. For example there’s one dungeon with a fairly tricky pull before the first boss, and in my most recent runs I suddenly found people climbing a ledge on the wall to bypass it. This has always been there but apparently it’s only been noticed and is starting to see widespread use now, seven years into the game, which is funny to me.

    It also makes me wonder how this will affect WoW Classic. They may not have got every single detail of the experience right, but generally speaking the private server community has had over a decade to “solve” Vanilla WoW, which has led to all kinds of weirdness that we never saw back in the day, such as people gathering buffs to boost their dps to the point where they can clear the entirety of MC in less than an hour.

    Like

  5. Wilhelm Arcturus Post author

    @anypo8 – The flaws with user generated content, in addition to quality, tend to be related to rewards. In a progression driven game, players won’t spend much time with content that doesn’t advance them in some way. But if you give players any direct control over rewards and advancement as part of the content they create, then things quickly get out of hand. SOE had a whole dungeon creation tool in EverQuest II that was quite powerful, but you could gain XP from it and so the most popular player built content was dungeons that delivered the maximum amount of gain for the lowest amount of effort in order to power level people. In the end SOE/Daybreak had to turn all of that off and it mostly faded from view at that point.

    As Bhagpuss said, Neverwinter has some decent tools. Cryptic invested in that. But you still end up with players not being able to accomplish their vision due to limitations in the mechanics and assets available, as well as the usual quality problems that plagues such endeavors. (We ran some WoW recreations at one point in Neverwinter. )

    In the case of EVE at least, there is also a loud contingent of PvE players who complain that the missions they run are boring but who also won’t do content unless they know the rewards up front and are essentially guaranteed to get them. For PvE content, the Abyssal deadspace content has a very big following from PvP players, in part because they are more loss tolerant.

    And, of course, as has often been said by developers, a smarter AI is easy. You could make a boss that always killed the healers first. That would be the smart boss reaction. Likewise, in anomalies in New Eden, just have the NPCs nuke fighters/drones first and swap ammo to find where you have a hole in your resists are and you would make people work harder. But would it make things better? The trick is to make an AI that makes you feel like you and your party have been through the wringer without just murdering you right away if you aren’t perfect.

    @Shintar – We’re getting close to something that will be in my New Year’s predictions post, but I will say that every retro server ever has been almost comical in the speed with which old hands at a game (who make up most of the players for such servers) will jump in and burn through the content because they know all the tricks. There won’t be a huge work-up to something like Molten Core. Some guild will likely do it on their first try because they will know what to expect and how to gear and staff for it. All the optimization will have happened, this will be the victory lap.

    Liked by 1 person

  6. Raziel Walker

    Wait, ratting titans are now a thing? Guess nullsec security has improved and people don’t get jumped that often? Plus I guess modern weapon systems are able to deal with subcaps better?
    so many questions from a single line not even relevant to the main subject…

    Like

  7. Wilhelm Arcturus Post author

    @Raziel Walker – You just boson all the way. I think I linked to a Reddit article describing it at some point earlier this year. CCP reduced the spawn rate of Havens for the December update allegedly to cut back on how lucrative it could be relative to the effort.

    Like

  8. Telwyn

    As others have alluded to this optimisation behaviour kills user generated content very quickly. I saw what happened in EQ2 from the sidelines, but we witnessed from close-up how the promising start at Neverwinter devolved into a mess as the optimizers gamed the system to make their storyless killfests the most popular content. Also the bot operators swamped the ugc instances to farm the profession nodes we could place as user-placeable loot. Compare that to the other extreme – the carefully crafted, lore-rich multi-dungeon arcs some players were creating even within the tools limitations. UGC can work certainly but how to prevent the optimizers from optimizin the fun out of it for the rest of us (throuhh forcing ever stricter restrictions on functionality to combat exploits)?

    Like

Voice your opinion... but be nice about it...

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s