Thursday, June 30, 2011

The Trinity of Roles

They are often referred to as the Holy Trinity - Healer, Tank, DPS.  At the moment they are the dominant way to organize character roles in fantasy RPGs but they aren't the only way and I was doing some thinking about the consequences of removing the Trinity and replacing it with something else.  The thing is that specialization is intensely powerful.  We can see this in society today in that I know how to do only a tiny fraction of the jobs out there with any kind of competence and in fact a large proportion of tasks that need to be completed for everyday objects and activities are well understood by only 1 in 1000 people or less.  Do you know how to make a pencil?  Wire a circuit board?  Make gasoline from raw oil?  I sure don't, but the benefits of having a society of specialists is undeniable.  The same holds true for fanasty RPGs; groups that have people that specialize as much as possible in their role are more successful at what they do.  The healers and tanks don't have to do damage as long as they are as good as possible in their role of mitigating damage and the dpsers don't have to contribute in any way at all other than killing the enemies.

In order to eliminate the Trinity we would need to eliminate the way in which it works, which specifically means getting rid of the WOW type of aggro.  As long as somebody can take all the damage from the monsters and mitigate it more than somebody else there will be tanks and as long as anyone can spend all their time healing there will be healers.  Simply normalizing mitigation isn't even enough since as long as people have customization control they will find the best tank and make that person tank even if the difference isn't that large.  What we need here is some way to swap aggro around the group with an element of randomness, remove the ability of players to be only a healer and give everybody the ability to heal themselves.  Imagine that the formula for getting aggro on a random monster is

Probability of getting aggro = Monster Max Health / 4 / Attack Damage Done

and that a similar formula would be used for bosses but the 4 would be replaced with some other much larger number.  The idea is that if you do at least 25% of the monster's health in a single attack you get aggro for sure and you have a proportionally less chance to get aggro by doing less damage.  This means that monsters will generally be hitting people who hurt them the most but who they hit and when they change their minds is somewhat random.  With this aggro formula everybody has to be able to take a hit from the monsters and nobody can reliably 'tank' them.  This means that everyone has to be aware of what the enemies are doing and be ready to defend themselves in some fashion.

In order to get rid of the healer role we would need to get rid of spammable heals that can target others.  Imagine that a healer only had two healing abilities that could affect other people:

Healing Aura:  All healing done to players within 10 yards of you is increased by 20%.
Mend:  The targetted player heals for 5% of their health every 2 seconds for 20 seconds.  You may only have 1 Mend active at a time.

This would mean you could have people with healing abilities that wouldn't be 'healers'.  A character with these abilities brings different things to a battle than a rogue would because they do help people heal better and can put a little bit of healing onto the person getting pounded but they have to spend most of their time using their attacks.  You probably want a healer around but there is every reason to think you can succeed without one.

We would need a comprehensive set of defenses and self heals for every class for this to work but it should make group composition much less rigid.  Any 5 good players should be able to beat any 5 man challenge but there would be some compositions that are better than others.  It would also help pvp dramatically as everyone would have to be keeping 'how tough am I?' in mind for pve and the sorts of things that both activities demand would be more similar.  A pvp focused character and a pve focused character should be much more similar and balancing them should be much easier.  Would it make the game more *fun*?  That is a harder question!  I very much like the idea of flexible group compositions and roles and letting people do many different things depending on the encounter.  I think though that the much greater role of randomness in combat would make tight tuning of pve encounters and balancing ranged and melee classes dramatically more challenging (or impossible).  I think the RP elements would be better as combat would feel a lot more like movie fights or fantasy novels where the monsters don't all stand around pounding on the ridiculously tough guy who can't hurt anyone.

Tuesday, June 28, 2011

No level at all, or just chasing a phantom?

WOW is in many ways a patched together hybrid of 3 different games:  Levelling, PVE and PVP.  There are people who enjoy all 3 parts of course but generally people end up spending the vast majority of their time in one of the 3 areas.  The most vocal players on the forums and on websites tend to be those that focus on PVE and PVP of course because they are generally the more hardcore groups focused on perfection but there is a huge portion of the playerbase that totally ignores the endgame and just levels character after character up to the level cap.  The trouble with the current setup is that people who have no interest in levelling up characters and seeing the various zones and stories are forced to do so at least once and then again any time they want to use a new character class and those that just want to level up alts get to watch the endgame vacuum up nearly all development time.  There are huge numbers of other games that have competitive PVE or PVP content that do not have this requirement (like first person shooters, say) and making levelling mandatory prior to engaging in what is the main part of the game for many people seems like something that isn't necessary.  Could we not create a game in which people can go out and do quests and see things without forcing everyone to do that for a long time before engaging in the rest of the game?

I think the first step in that direction might be to remove levels completely.  Characters can gain skills and powers without level coming into the equation at all and this would mean that characters would not be utterly unable to fight monsters in much of the world when low level and also would mean that characters could not obliterate the weaker monsters so easily once they are at the level cap.  This also means that when a new set of content is released players do not have the expectation that they can gain more levels and with those levels gain massive increases in power.  You have a choice to make at this point though:  Do you allow characters to gain in power internally or are power gains going to be entirely gear dependent?  If characters get more powerful internally up to a cap then everyone still has to go out and do whatever it is you do to get more powerful until they are capped before they can really get into big group endgame content so we aren't any further ahead.  If characters can't get more powerful themselves but merely can put on better gear then getting into the endgame is very fast and a good player would be able to be very competitive in no time.  Unfortunately that has the side effect that people don't *feel* like their character is improving but rather just that their gear is getting better - they are just a vehicle for their equipment.

One idea I have been tossing around in my head is a rolling, unachieveable cap.  Some MMOs do this now, but none of the ones that are big in North America do, it is a feature found in some Asian MMOs.  The idea is that the level cap is 90 but that nobody in the world can reach it because the last 10 levels each take months to get.  There will be some people who get to 85 or so but the majority of the world is going to stop levelling at 80.  When you want to release new content you just bump up the level cap to 95 and push back the point at which levels become brutal to gain by 5 levels also.  We would also need to remove level from the combat equations.  The idea is that there is always an incentive to go out and do more quests or dungeons because your level can always increase but that there is a fairly small difference between the absolute top player and someone who plays much less.  It would also mean that a new player is weaker than veterans but they can jump right into their chosen mode of play very quickly because their level is not nearly so impactful.

The idea is to dramatically lessen the impact of each new piece of gear and each new level so that a new character has to do a fairly small amount of questing and learning before they can start playing with their friends and trying out all the parts the game has to offer.  If a level 85 character has double the health of a level 30 then the level 85 is much more powerful but at least they could do some things together... when the level 85 has 1,000 times as much health and can defeat anything with one attack then they cannot usefully cooperate.  This would end the tiered raiding style of WOW because gear would not be obsoleted nearly so quickly because it wouldn't scale very much - a group of people might well be able to raid half a dozen instances that could be considered reasonable challenges with useful rewards.  It would also make swapping characters in PVP much easier as getting to the baseline for being really competitive would be much quicker, though becoming the absolute top would remain utterly brutal.

Whether or not this style of game would actually be more popular than WOW is not a question I can answer. Lots of people talk about what they want in a game but they rarely understand their own desires well and they often refuse to play games that have exactly the things they claim to want.  Maybe the mandatory level up grind is actually really helpful in retention... I don't know.  I would like to try this sort of game though as it has a feel much more like Diablo 2 than WOW and that might work out really well.

Thursday, June 23, 2011

Infinite stacks

Way back in the day I played a lot of Heroes of Might and Magic 3.  It is a great game but of course it has many quirks and things that might be called design flaws depending on your point of view.  One thing that defined it was the fact that battles and decisions in the early game were so incredibly critical to lategame success due to momentum.  In the early game you would often be in situations where playing perfectly would save you half a dozen of your units and playing reasonably would lose them and this would snowball as the game went on - having those units would mean that in the next battle you lose only 2 units instead of 6, and the next battle you lose none instead of 10.  Every time you lose any units you lower your maximum army size permanently and that means in every battle that follows you do less damage and will take greater losses.  The main problem with this is that when you play the early game perfectly the late game ends up being really boring as you march around the map with an army that has been accumulating since the start of the game and which cannot possibly be stopped. It still takes a few hours to clear out every single opponent but nothing could happen that could stop your victory.


On the other hand even small losses would slowly chip away at your army's strength over time and a bit of bad strategy or bad luck in the early game could wipe you out easily if the scenario were challenging.  I don't think that the early game should be irrelevant of course because that would make it rather boring but it seems like in most hard scenarios the player ended up spending the last half of their playtime demolishing trivial challenges.  Recently Sthenno told me about a very similar game called Kings Bounty:  Armored Princess that solved this issue very neatly.  (I find the Armored Princess title to be utterly hilarious... who thought that up?)  The solution involved limiting army size based on characteristics of the hero.  In HOMM3 any hero can lead any amount of troops so losing troops at any point permanently shrinks your final army but in Armored Princess that isn't the case because army size is capped by the leadership score of the hero involved.  This means that you still have to win every fight and you can't go around losing *too* many troops to each battle but losses in the early game can be made up for by luck or strong play in the later stages of the game.  This mechanic balances out the critical points in the game much better and allows the designers to create significant challenges in the later stages of the game that cannot be trivialized by good play in the early going.  In addition it prevents the situation where a player does badly in the first hour of play, plays well for 3 hours, then discovers that he simply cannot win anymore.


HOMM6 is coming out fairly soon and I look forward to it with much trepidation.  HOMM3 was a treasure, despite that it did have a few flaws, but HOMM4 had really terrible gameplay and was quite a failure.  HOMM5 was basically a reskin of HOMM3 but with huge problems with AI pathing to the point that large and complicated maps were impossible - the game just locked up your computer.  Had that not been an issue though HOMM5 would have been a tremendous game and very enjoyable.  Here's to hoping that Ubisoft figures out the magic formula:  People don't want that much innovation.  Just build HOMM3 with better graphics and more cool stuff and people will love it and pay for it.  Mindless summer blockbuster films prove this over and over... you can add nothing of substance to the genre and just rehash an old classic as long as you rehash it *well* and people will shovel money your direction.  Just ask Madden Football, NHL Hockey or any other 'prettier and with some new toys' game franchise!

Tuesday, June 21, 2011

I need combat logs

WOW spoiled me.  For the longest time I had access to the best tool available for maximizing your character's effectiveness in an RPG - the combat log.  I could find out whether or not my models were correct, how game mechanics actually worked and get real data on average monster statistics.  I find myself wishing for all of this while I play Diablo 2 and DnD.  DnD at least has few enough actions that I could record everything if I really wanted to but it does have the problem of a very relevant gameboard and tactics that are very difficult to record unless you have extreme amounts of time available.  D2, on the other hand, has no such recording possible at all.  I suppose I could eventually teach myself enough about coding to write a mod that captures all combat results and gives me good logs of the data involved but not only would I have to write the mod to do the data collection I would have to write a data analyzer too.  D2 has complicated enough mechanics that parsing the reams of data and displaying it usefully would be awesome but would be a huge undertaking.

The iconic example is crushing blow.  Crushing blow in D2 works roughly like this:

1.  If you hit a monster, check your Crushing blow % to see if it procs.
2.  If it procs, the monster loses a % of its current health that depends on weapon type and monster type.
3.  The % health lost varies between 25% for a melee weapon and a regular monster with 1 player in the game and 1.4% with a ranged weapon on a boss with 8 players in the game.
4.  Once the crushing blow occurs the weapon's regular damage is dealt.

So how much of my damage is coming from crushing blow and how much is coming from my regular damage?  I have no idea!  I know that having 100% crushing blow is good because certainly it makes me kill things more quickly and it isn't that hard to achieve but I can't effectively test how good it is in any way.  Taking off all my gear that provides crushing blow means my character doesn't work properly for other reasons and it would compromise the data even then... and that even assumes I can figure out how effective I am against a wide variety of targets with different tactics, locations, defenses and numbers.  It just isn't possible to say with any reasonable amount of certainty how good crushing blow is and how much I should prioritize it.

This makes me crazy.  I have just as much fun poring over gear lists and endlessly iterating on my best possible gearset for my character as I do killing monsters.  Hell, when I am levelling up I actually spend half of my time in town enchanting, gemming and tweaking my gearset instead of just going out and blow up the enemies.  Every 5 levels or so some new equipment becomes available and I go back to tweaking and enchanting again to reoptimize for encounters that can't possibly be hard.  It clearly isn't an efficient way to level but it is tremendous fun.  I want to optimize perfectly when I do get to the endgame in particular but there is so much data I have absolutely no way to acquire.  This makes me wonder what will end up becoming of my paladin spreadsheet for WOW.  I have been updating, tweaking and maintaining it for years now but I am losing interest in doing so due to not playing WOW anymore.  That spreadsheet was a huge project, something I poured myself into, and soon it will lie abandoned, slowly becoming obsolete.

Friday, June 17, 2011

Too much stuff to do

Rift, it seems, is starting to lose its lustre.  It was, like nearly every other MMO in the past 6 years, touted as the WOW killer and by far it seems to be the one that came the closest.  It has less than 10% of the subscribers of WOW of course so it isn't like it really had a chance but a very large chunk of that has to be due to the fact that just like Facebook people play WOW because their friends are already there.  Just because WOW is the biggest is plenty of reason to play it because you are likely to find people you know already there and being social is a huge component of the game.

One of the big things that lots of people yelled about on the forums when Rift was new was the tremendous speed of development.  New patches came out very rapidly and if I understand it right (not having actually played the game myself) there have been 3 major content patches in as many months - a release schedule drastically more rapid than the one WOW players are used to.  It has been regularly noted that Blizzard pays out enormous sums to Activision shareholders and notes that the money simply could not be put back into the game in any useful fashion.  Players complained about this because although everybody knows that if you double your production budget you don't get double the content you still get a lot more content than you had before.  It is clear that since a new company with a new game managed to put out a lot of patches that Blizzard could do the same if they wanted to since they have access to truly stupendous funding so we are left wondering why they do not.

I think we have the answer now.  Rift players are complaining that in fact the pace of new content is *too fast*.  They feel like there is no point in grinding out a dungeon when the next dungeon is going to come out in two weeks with better loot and harder challenges.  The gear grind in MMOs is a strange beast and no one can argue that it is hugely, though obviously not solely, responsible for people raiding.  Everyone wants to have the best gear and twink out their character to the max and for many people those rewards are the reason to raid.  The trouble with very rapid patching and new content is that everyone knows that their hard earned rewards will be obsoleted right away and they feel like the effort involved is simply not warranted.  This can be countered by making the content very easy such that people can get their rewards quickly but then the people that play for challenge get bored and the rewards themselves lose their shine because they were so easy to acquire.  People value that which they worked hard to acquire much higher than that which was given easily and this seems to be a powerful moving force in MMO design.  Players evidently need a certain amount of time to bask in the glory of their acquisitions before they are ready for new challenges and new goals and if forums and bloggers are to be believed Rift has gone too far in the rapid production direction and is losing a lot of players as a result.

Of course no matter what schedule a company picks lots of people are going to be bitter about it.  Companies need to figure out what the bulk of their playerbase wants and try to cater to that.  Given the relatively long life and immense success of WOW I think we can safely say that Blizzard did pretty well overall; they have made their share of big blunders for sure but there are so many people that have raided for so long they must be doing something right.  We could avoid this whole argument by losing the gear grind level cap style of fantasy MMO for some other design but I have no idea what the next big thing will be - somebody will eventually break that mold but for now every new fantasy MMO game pretty much follows in WOW's very large footprints.

Tuesday, June 14, 2011

Twinking out

Despite having aims at redesigning D&D 4th edition I actually have played it very little.  I ran in a couple of fairly short campaigns fairly early on in the game's life and then didn't play at all for a few years.  Last weekend we had a hack and slash type campaign running through the Thunderspire Labyrinth module and had a blast.  We talked a little about the changes in the game since launch and it seems clear that DnD is moving in the right direction.  The new classes seem really interesting and much less ridiculous than the ones in the book at launch - they are going far beyond the old archetypes of rogue / warrior / mage / cleric but the numbers look a lot better.  The rewrites for clerics take the very good tack of nerfing their overpowered abilities into oblivion and the warrior rewrite (which I am playing) is much better.  I lack the excess of 'damage without a roll' abilities and am much better at controlling the enemies and bashing people.  It looks like now if people choose to ignore my tanking abilities I get to do a ton of damage and be really tough and if they just beat on me instead then I do some damage and get to focus a ton of enemy damage through the tough guy... which is also good.

The trouble is that there are some mechanics issues that they can't just modify slightly to fix.  We still have lots of big problems to deal with, namely the brutal straightjacket that the current stat mechanic imposes on character design.  It is obvious that any character that does not maximize their one important stat is simply going to be terrible and that every class has a tiny number of viable builds due to this situation.  I don't mind there being bad ways to play but I sure don't like it when there is only one possible way to be good.  There aren't very many different styles of characters or impactful choices during character creation for some classes and this is a poor design.  Thankfully combats are pretty interesting affairs anyway - even though my class had very few activated abilities I had many interesting choices in terms of where to go, where to stand, who to hit and how to use my abilities.  Just the simple mechanics of flanking, shifting, terrain and attacking generate lots of fun dilemmas so playing the tactical game is good fun regardless of the stat mess.

The other issue I have was made very clear:  Magic items are not working in a way that I approve of.  This is less of a clear cut issue than the stat issue since there are probably some people who think that magic weapons should be absolutely necessary to move along and that high level characters should be helpless without their magic weapons.  I think that magic weapons should be interesting and impactful but it shouldn't be in a way that wrecks the numbers.  It would work well if magic weapons simply added damage or gave characters interesting abilities because then lacking them wouldn't make things untenable but when they add 6 to attack rolls it means that lacking a magic weapon makes fights pretty much unwinnable.  You can't go from hitting 60% of the time to hitting 30% of the time and have the game play even remotely the same.  I like magic items having fun one shot abilities, adding to skill checks, giving new combat options or other bonuses but large bonuses to the most basic role in the game is just the wrong way to go about it.

Thursday, June 9, 2011

Bang! Precisely one of us is dead.

Imagine, if you will, a table.  This table is nothing out of the ordinary itself but rather has three extraordinary people sitting around it.  The reason the these people are extraordinary is that they are planning on pointing a revolver with one bullet in it at their own heads and seeing who dies.  The question becomes how do we design this revolver / maniac game so that it is intuitive, fair and quick?  The entertainment value of Russian Roulette is pretty much guaranteed to be there or not depending on the observer so my job is entirely to make sure that the game is the best one possible.



Usually the way the game is described a single bullet is put into the gun, the chambers are spun to randomize the location of the bullet and people put the gun to their heads and pull the trigger one at a time.  The trouble as I see it is that you sometimes end up with somebody knowing that the only remaining chamber has a bullet and it is their turn to point and fire!  While it is quite a bit nutty to be involved in a game of Russian Roulette in the first place it requires even more moxie to complete the game when you know you are the one to die.  Clearly cheating in a game of RR is going to have some pretty serious consequences so even if you have a 50/50 shot it seems very reasonable to try your luck - it isn't likely a bunch of people willing to be in this game are going to let you get away with chickening out or shooting somebody else.  However, once you know that you have a 100% chance to die I expect that the prospects of a bunch of homicidal, vengeful RR players looks not too bad after all.

You could avoid that certainty by simply rerandomizing the gun after every pull.  Each time somebody pulls the trigger the chambers are spun again before the next lunatic takes his turn.  This avoids any certainty of death for sure but has the issue that the first person to go is at a substantial disadvantage.  You can keep it random by simply having the players draw cards to see who has to go first but that seems like an inelegant solution because it introduces another random element into the game from outside - functional, but ugly.  I think the best solution is to rerandomize after each round rather than after each pull.  Have each player pull once and if the bullet is in the unused chambers you spin them again and start from the beginning.  This way there is no disadvantage for going first but nobody ever ends up certain that they are killing themselves either.  It has a decent likelihood of taking awhile if the bullet ends up unused a number of times but I figure since the game isn't going to be played a lot of times we can probably afford the extra time spent.

So there you go, if you want to play RR you just make sure you have less than 6 players and have everyone take a turn before you spin the chambers again.  Now you know!

I wrote up a little simulator to figure out your approximate likelihood of dying based on your position from 1-3 when the gun is randomized after each shot...  the lesson is don't go first.

1:  39.4 %
2:  33.0 %
3:  27.6 %

Tuesday, June 7, 2011

Elitist Raiders or Tobold being wrong.

Tobold recently made a series of posts about how raiders in WOW are elitist and don't want anyone to join them in their raids.  It is a revealing sort of diatribe against a group of people that Tobold obviously wanted to join but couldn't for a variety of reasons - scheduling, dedication, skill, take your pick.  Not that I am saying that all raiders are lovable bunnies who help everyone, because they aren't, but rather that they have the usual distribution of niceness and selfishness that you find all throughout humanity.  Becoming a successful raider has a number of barriers to entry including a substantial time commitment and playskill and it can certainly be discouraging to try to join that group and find out that you simply can't make the cut for one reason or another.  It is certainly disheartening to try to do something new and find out that you aren't good enough or don't have enough time to play as much as is necessary to match up with your group but that doesn't much mean that those you are trying to join are somehow deficient.

In fact the raiding community has an incredible selection of articles and tools available to those trying to join it and many volunteers from the raiding community (myself included) go to extreme lengths to assist the new and inexperienced in joining the raiding community.  Though in the best guilds in the world there is definitely the expectation of independent knowledge and research I have seen plenty of the medium guilds that are very happy to assist people in learning and improving.  New recruits definitely have to show some initiative and drive but there are an endless list of opportunities where the only expectation is time spent on research and the ability to do a simple google search.  There is a reason that recruitment notices for guilds nearly always ask for people who are dedicated, self motivated, easygoing and tenacious as more important than book knowledge or track record.  Hell, many or most of the mid level guild advertisements explicitly state that they are happy to train anyone who is willing to work hard and put in a lot of time trying to improve.

This 'raiders are elitist' attitude is nothing new or special.  People generally reserve their greatest enmity for those that are close to them but have some particular difference - just see divorce proceedings for evidence that those closest to you are those who end up being the focus of your greatest hatred.  Raiders dislike casuals, casuals dislike raiders and this relationship is true across the spectrum of hobbies and sports.  Not everyone falls prey to this of course but it is an undeniable tendency.  It is a little sad though - both in the general case and the specific. Tobold is so often a voice of reason and sense on the internet that it makes me sad to see him ranting and raving against raiders for no particular reason.  Raiders are jerks, raiders are nice, raiders are people.  Mostly they are just people who have lots of time to play video games and really like to kill gigantic dragons and take their shiny, shiny loot.  You might dislike those particular traits or you might love them but there really isn't much more to the group as a whole than that.

Friday, June 3, 2011

Remembering to eat

Diablo 2 has my firmly in its savage clutches once again.  I am playing the Eastern Sun mod and having a wonderful time doing so - I highly recommend it over playing the regular game online and indeed over doing nearly anything else.  It has given me some very powerful insights into game design as I play through it too - it isn't all about killing things and taking their stuff after all.  (Yes it is.)  One of the most substantial differences between ES and Classic is that Classic has extremely bumpy gear progression and little flexibility.  In Classic nearly every item is utter crap but there are a few unbelievably powerful items that are obviously the best thing you can wear.  The trouble with this in a game where you progress by killing random monsters and hoping to get a lucky drop is that you end up with a few incredible items and a few junk items that you don't care about at all.  When you can find an item that will give you 50% more damage why bother trying to figure out which of many shoddy items could give you 1% more than the other?  ES has a much bigger list of items of nearly every category including runewords, uniques, sets and others and takes much better care to make sure that the quality of these items is not so easily divided into 'crap' and 'awesome'.

This has many consequences.  One of the most obvious is the constant upgrading of items.  Instead of just putting on anything at all while I wait for the insane drop I regularly look for upgrades and rearrange my gear to make sure everything balances out nicely.  More of everything is good but I need to make sure I have enough resists, health, mana, pets, damage, magic find, etc. and no item does everything you want so there is a constant shuffle in and out as I find more things.  One of the other great consequences of this design is collecting.  Since many, many items are relevant I need to have a monstrous collection of items on hand to make sure every character can get an excellent set of gear.  As stamp collectors, rock collectors and any other kind of random collector can tell you there is a real sense of joy in completing a collection and having a huge inventory of things on hand even if others dismiss the collection as useless.  Finding all kinds of things and keeping them, sorting them and eventually using them is wonderful fun - I love making a new character and finding all kinds of great stuff they can use hidden away in my stash.

The other big thing that ES does that seems great for a game long term is tiny improvements.  Once you have an Harlequin Crest for Classic you never look at that piece of gear again.  Nothing can improve on it.  In ES that is almost never the case as you can collect jewels to crush together into better jewels which can crush together into even better ones which can be used in formulas to make even crazier pieces of gear.  Every piece of gear can be tinkered to have more stats, forged to have more stats and dstoned to have more stats all in different ways and by collecting different things.  That constant experience of finding simple things and having them slowly combine into improvements to your character makes for a feeling of accomplishment and whets your appetite for even more slaying and gathering.  That combination of substantial upgrades that come periodically and small upgrades that come constantly is a recipe for addiction.  Addiction, in this case, being good.