Saturday, January 12, 2013

Blizzard Pays For Overemphasizing Item Level

Via Blizzard tweet, we learn that WoW's patch 5.2 will remove the ability to upgrade the quality of endgame epic gear which they added to the game in patch 5.1.  The studio was forced to backtrack not due to the design merits of the system, but rather because they have spent years emphasizing the importance of item level (ilvl) in the game's incentive structure.  They have no one to blame but themselves that players are now over-reacting to anything that affects this meta-statistic.  To understand why, we need to look at both the history of ilvl and how it has been used as an incentive.

Making a book-keeping stat into an incentive
Item levels were originally an internal number used to determine how many stats each item got.  In principle, you could use this meta-stat (after accessing it via third party database or UI add-on) to compare gear (or to argue on the forums that a specific item should have higher stats based on its ilvl), but in practical terms it was irrelevant to players for years post-launch.

Ilvl shot to prominence midway through the Wrath of the Lich King era when players began using a third-party add-on called "Gearscore" to add up the ilvls of a player's gearset as a quick and dirty way to check whether the player's gear was plausible for the content your group intended to complete.  Gearscore was not without controversy, but Blizzard brought it into the game's base UI anyway, adding the average ilvl of the player's gear to the UI and using it as a screening mechanism for the random group finder in patch 3.3.

This change happened comparatively late in that expansion cycle, and applied primarily to content that players were already routinely and trivially completing (easy 5-man heroic dungeons).  Going out to get gear specifically to meet a minimum ilvl to be allowed into content came later, most prominently with the addition of the raid finder in Cataclysm's final patch.  During the same window, we've seen each raid expand to three separate tiers of loot - differing primarily in ilvl and associated minor bump in stats - that only further emphasize that increasing ilvl is an incentive in its own right, and not just something that happened incidentally as you upgraded your gear.

Comment Update: Commenters point out that ilvl scaling is actually less linear than I remembered, making the upgrade system too powerful, not too weak.  The rest of the reasoning in this post about why focusing on ilvl as an incentive is a bad idea stands (and actually makes more sense with the correction).
A Technical Increase In Power
The new item upgrade system in patch 5.1 crossed an important line - increasing the ilvl of gear (along with marginal bumps to its stats) was explicitly used as an incentive.  The intent was to extend the benefits of continuing to earn valor points (from daily quests, random dungeons, etc) by allowing them to be used for small upgrades to certain gear.  A bump of 8 ilvl's may sound significant until you consider the context.

In the launch game, ilvl's corresponded roughly to the level at which the player obtained the item.  However, max level group content at each of the game's level caps led to dramatic inflation of item levels, such that gear at level 90 is approaching ilvl 400.  As a result, the new system was only a 2% boost to the ilvl of one of your pieces of gear (out of 15-16 depending on whether you use a two-handed weapon).

Even if you did eventually get this bump across the board (I'm not familiar with whether every single slot could actually be upgraded this way), a 2% increase in ilvl for all of your gear does not directly lead to a 2% increase in DPS or other output.  Other factors, including your inherent base statistics, scaling rules for increased combat ratings, and most importantly actual player performance are also going to impact character performance.

In short, this mechanic was technically an increase in character power, but functionally small enough that it's all-but cosmetic.

Running into the "optional" debate
Blizzard's defense of the item upgrade system hinged on two arguments.  First, they state that players will quit if there is not stuff for them to do, and that many players will not do any in-game activity that does not increase their character power.  I think we can agree that this is mostly sound reasoning, with the caveat that most players would really prefer that "stuff for them to do" take the form of new content on a more frequent basis, rather than incentives to run the old stuff into the ground.

Having said all of that, Blizzard tried to have it both ways by saying that the upgrades, while intended as an incentive to convince players to get valor points, are optional.  This argument is sound from a game design perspective for all the reasons I discussed above about the actual significance of such a small boost to ilvl.  Unfortunately for Blizzard, the correct design approach lost a battle of perceptions brought on by their own decisions to emphasize item level in incentive structures.

Players were already complaining that they felt daily quests were not optional because the resulting reputations are required to purchase entry level gear to start raiding.  Allowing item upgrades exacerbated the situation because players eventually either run out of reputations to grind or else acquire superior gear by other means.  I don't believe the system worked on all gear, but it worked on enough gear to make these players feel that they were now obligated to continue doing stuff they did not want to do in order to get valor points to pay for the upgrades.

I don't think Blizzard's position was wrong from a design perspective.  However, as I wrote a few months ago, I think they lost this battle for the same reason that they lost the battle over dungeon difficulty early in Cataclysm.

Failure of the skinner box
Blizzard is still in the business of trying to sell a service to customers, and you can only get so far by telling a customer who is dissatisfied that the merits of your design trumps their preferences.  If you tell the customer that they need ilvl - directly through minimum requirements for the raid finder and indirectly by using ilvl as an incentive - you should not be surprised that they believe they need ilvl.  Once that has happened, it is natural for these customers react poorly when told that they have to do something they do not want to do in order to get the ilvl they think they need.

As Tobold points out in a post examining the motivations for botting, MMO's have increasingly misused incentives as a means of pushing players into trying other forms of gameplay - daily quests, dungeons, PVP, etc - that they do not enjoy.  History has shown time and time again that this CAN change player behavior, but DOES NOT change player preferences.  The numerous unsavory reactions - joining PVP matches only to try and sit AFK for the currency rewards, requiring gearscores far in excess of what content was designed for, and threatening to cancel because you feel "forced" to do daily dungeons - are a natural response.

I recognize that studios are struggling to produce enough material to keep people paying, and that they need to get each player to use as much content as possible.  I recognize that incentives can be the difference between enough players in the dungeon queue to fill groups in a timely fashion and leaving newcomers high and dry.  In the long term, though, reducing the game to a "Skinner box" activity that you do only because it is a prerequisite for something you actually want to do (e.g. raiding with your friends) is a recipe for burnout and churn.  This particular example of pushing an arbitrary number far beyond its functional significance was an extreme case, but it is by no means the end of the issue.

13 comments:

  1. "Even if you did eventually get this bump across the board (I'm not familiar with whether every single slot could actually be upgraded this way), a 2% increase in ilvl for all of your gear does not directly lead to a 2% increase in DPS or other output."

    This is not actually correct. It's not linear. Due to the way itemlvls work, every 13 itemlvls (a "tier") offers a 10% stat inflation. 8 itemlvls is thus 6% stat inflation. That's worth MUCH more than 2% total performance.

    ReplyDelete
  2. "Even if you did eventually get this bump across the board (I'm not familiar with whether every single slot could actually be upgraded this way), a 2% increase in ilvl for all of your gear does not directly lead to a 2% increase in DPS or other output."

    This is not actually correct. It's not linear. Due to the way itemlvls work, every 13 itemlvls (a "tier") offers a 10% stat inflation. 8 itemlvls is thus 6% stat inflation. That's worth MUCH more than 2% total performance.

    ReplyDelete
  3. Yeah, what rodolpho said above.

    The problem was that because of quirks in itemization it was often better to hold older pieces and upgrade them than it was to get new pieces. It worked really weirdly with trinkets and often made broken sets. That, and it made inflation of items go even faster than it already had been. It was a poor idea that had to go - but not because it was trivial, but because it was too good.

    ReplyDelete
  4. Disregarding the corrections in detail above, which soar right over my head as a non-player of WoW, that was an excellent summary of where we are and how we got there for MMOs in general, not just for WoW.

    Fortunately I do think we have hit the high-water mark of this particular design orthodoxy. WoW, although it was and remains far and away the most successful western MMO, is no longer growing and replicating its success has proved beyond the wit and skill of the industry. The prevailing winds are blowing in other directions now and WoW may be left to plough its own highly profitable furrow (on some kind of Steampunk sail-assisted tractor if that metaphor is anything to go by).

    I often feel like the proverbial mouse in the wainscot, living in a house whose function and form I barely apprehend, using only the parts that fit my purpose and remaining happily ignorant of the rest. I played WoW for three months, finishing just after the Dungeon Finder was added and I didn't even know iLevels existed. I never even heard of them until maybe a year after I'd stopped playing. I wonder how many of the supposed 10m WoW players still there are in that happy boat.

    ReplyDelete
  5. Replacing a green magic staff with a golden one of different design is much more interesting than increasing the itemlvl by 5%. This is especially true the more often the replacing happens. Numbers become boring much faster than art.

    In the past an item was a combination of numbers, art design, location where it was found, rarity, (green / blue / violett / orange).

    Today items are dominated by one single number: itmlvl. All the rest doesn't matter.

    Iinstead of immersing the player in a simulation, Blizzard today, shows him the technicalities of the gameplay as clearly as possible. They thus make the upgrading process and gear itself much less interesting.

    ReplyDelete
  6. Absolutely. Don't get me wrong, I think the item upgrade system was a bad idea. It was obviously designed to stretch out content, and delivering NEW content more quickly would be a far better way to go.

    Thing is, Blizzard development is almost legendarily slow. They aren't Trion; they can't deliver new content every 3 months. Given those constraints, it's easy to see why they decided to implement item upgrades.

    At least the implementation is sound, unlike item reforging, which forces players to reforge all their gear after every single new piece (or item upgrade) to avoid overcapping two hard-capped stats.

    ReplyDelete
  7. Imagine a feature that increases content for a lower cost of production than item upgrades. I would argue it is impossible. Making a new item with a new skin, while case-by-case a better feature, is far more expensive than allowing all items to be more valuable, given time and effort.

    You may say that reforging and upgrading make the endgame more tedious. You could also say it's successfully make it a longer event. All those players that reached BiS (or their personal equivalent) and left now have a reason to keep playing week after week. A good enough reason? Not for some. More for others. Enough for me.

    ReplyDelete
  8. But the aim of a business isn't simply to lower production costs. If a games company, particularly Blizzard with its reputation for quality, chooses an inferior route because its cheaper the profitability of the product may be reduced.

    They've always used the argument that content is expensive and I've never quite believed it. If it's too expensive to make new dungeons for a game which has 10 million people paying an average of, what?, $7 a month then how to games with 10,000 players produce new content? But they do.

    Also some of WoW's new content has had a very lazy feel to it. The second ZulGurrub was pretty much just the first one edited to put an extra 0 on the end of all the numbers. Instead of hitting for 200 the mob hit for 2000. Same art, same AI etc. So, as a proportion of millions of $15/month subs plus Chinese fees how much did doing that one cost? 0.00001%?

    ReplyDelete
  9. Upgrades are a simple answer to a bigger problem: what to do with VPs after you start raiding?
    When you make all the VP-buyable gear worse than raiding gear (so that it's not "needed"), you have a problem to keep people running dailies and dungeons.

    Of course, the problem runs deeper (as you have written). The problem is that we're seeing a "return" to cross-pollination across activities, i.e. you need activity 1 to perform in activity 2. Not only this is bot-bait, but it's also a design failure. WoW had been advancing towards activity separation in previous expansions, but did a step back in this one. Other MMOs seem to follow suit (or more precisely, they never advanced in the activity separation). The genre will not implode simply because it still provides very good entertainment for the cost (even more so with F2P).

    @Bhagpuss: The prevailing winds are blowing in other directions

    Ah, really? All I see in the new F2P trend is an even more grindy gameplay, which can of course be convenently skipped by paying....

    ReplyDelete
  10. Also posted in the MMO champion article linked to in your article is this:

    "we all know that even on live things can still change [...] so if a change is urgently needed, we do our best to apply it as fast as possible"

    If Blizzard has finally decided that the NPC upgrade vendor has to go, perhaps they'll remove him even earlier than 5.2. What would you do if you woke up after the next restart to find him gone?

    I have some VPs in my pocket that I was saving to buy a new item. I think I'm going to spend them tonight on upgrades instead, while I still can.

    ReplyDelete
  11. Bashiok has confirmed that it's just going because they'd rather people spend VP on new items for the new raid tier. The item upgrade vendors will be back in 5.3.

    http://us.battle.net/wow/en/forum/topic/7592242035#1

    ReplyDelete
  12. "But the aim of a business isn't simply to lower production costs. If a games company, particularly Blizzard with its reputation for quality, chooses an inferior route because its cheaper the profitability of the product may be reduced."

    And if Blizzard devoted itself completely unto quality over cost, they would become Copernicus and WoW would never have been. Saying cost should not be a factor because cheap things are bad is as false as saying quality shouldn't be a factor because people need tons of content.

    The answer in game design, as in food service and airplane parts, is the happy balance of cost and quality.

    ReplyDelete
  13. If it's not worth upgrading a 496 to 500 & 504 subsequently, there is no need to pursue Heroic 502 gear, is there?

    Upgrading is worth far more than a 2% increase or heroic content wouldn't require a higher gear score. Furthermore, players wouldn't expect a smaller margin for error in the event itself - which is the exact point of harder content.

    ReplyDelete

Comments on posts older than 14 days are moderated and will not appear until manually approved because the overwhelming majority of such comments are spam. Anonymous commenting has unfortunately been disabled due to the sheer volume of comments that are defeating Google's spam filter.