Thursday, June 28, 2012

Pathfinder in review

My Pathfinder roleplaying campaign is winding down and we have only 1 session left to go.  Now that I have seen the game played from level 1 through level 10 I have a fairly good idea of what sorts of things work in the system and what don't.  The big thing I have learned is that the numbers and systems in Pathfinder are mostly complete garbage.  There is very little there that I feel like has any real merit from an objective standpoint and the real reason that Pathfinder is so popular is that it is just an extension of the previous system which everyone is familiar with.  Granted Pathfinder is better than DnD 3.5, which is fine and all, but honestly if I was building a system from the ground up I find very little in it I would emulate.

I played a couple of purely tactical DnD 4th edition campaigns with no roleplaying and tons of combat over a shortly time span and I think I can safely say that although 4th lacks some of the flavour and nostalgia of Pathfinder it is a massively superior system in terms of numbers.  Classes are far more balanced and ridiculous and terrible combat situations come up a lot less.  In old school DnD it is entirely normal for specific character types to be nearly useless in many fights, or in fact in all fights, and for optimized characters  to be as powerful as four regular characters and in 4th those things are brought back into line.  There are some common issues that go across both systems though that I would like to see addressed either in DnD Next or in a system I design myself.  In no particular order:

Figuring out your plus to hit should be easy.  Endlessly adding up bonuses and penalties to hit from seven different sources trying to figure out if you missed by one or just barely hit is annoying.

Calculating the results of your actions should be fast.  People want to spend time coming up with fun actions and making interesting choices not doing tedious arithmetic.

Healers shouldn't be pigeonholed into spamming "Heal" for entire fights.  Playing a healer should involve interesting choices and let you heal people as part of that, not doing one thing over and over.

Feats and other static player customization should be reasonably balanced.  I should not have to scour enormous databases to discover 'mandatory' feats that are necessary to be remotely competitive.

Players should have the correct number of power choices.  If fighters only have 'attack' then their game is boring.  If they have 16 different powers the game is overwhelming for new players.

Magic items should not be so powerful that characters are helpless without them.  I want to be a hero, not a vehicle for equipment, nor a merchant trying to haggle my way to greater combat prowess.

One thing that makes me hopeful is that DnD Next actually solves a couple of these issues.  We don't yet know much of what it will be like when we see the whole system but it sure looks like the arithmetic concerns I talk about are much better and the characters won't be overwhelmed with combat options like they were in 4th edition.  I think I will get back to working on SkyRPG and see if I can get some of my ideas into some concrete sort of form.

Tuesday, June 26, 2012

The benefits of trade

When D3 was being designed and tested there was one really major obstacle to figuring out how well geared players would be:  The lack of a functional Auction House.  As I understand it internal testers at Blizzard basically just played the single player game and based their tuning of inferno difficulty on their experiences in single player and this is a mostly useless data point when considering how things work with millions of players trading with one another relatively efficiently.  I was talking with Ziggyny and Gnome awhile ago about this and we were speculating on how Blizzard could have programmed an AH simulator to try to understand what would happen to a player's gear when they had access to the AH.  Ziggyny figured it could be done without much difficulty but I did not; I think you could get a decent first order approximation but rapidly the effects of efficient trade would be really difficult to guess because you would need a lot of data on gearing priorities, stat weights and other such things that are difficult to generate prior to launch.

Rather than try to simulate an entire economy though I think the best way to figure out the effects of efficient trade on gearing is to try to come up with some kind of multiplier to describe how much bigger a loot pool an individual has access to.  If I play the game with four other players, one of each class, and we trade amongst each other with each person looking to maximize their own performance then my loot pool is certainly not five times as big as single player but it is much bigger than normal.  If I am, for example, the Barbarian then any piece of loot with Strength but without any other primary stat will generally end up on my character.  However, there are going to be times when loot with both Strength and Intelligence drops and an Intelligence using character puts it on instead of me.

Of course having an enormous pool of players helps in other ways.  In the previous five person example when two sets of good Strength pants drop only one gets worn and the other gets discarded but in the real game both get worn.  When levelling up I often find that I get a new drop and realize that it is literally five times as powerful as my old item simply because I have found no decent bracers while at the same time I discard my fifth good helm.  I imagine these two scenarios to find some kind of equivalency:

Scenario 1:  I use the AH and buy gear of value X for each of my 13 slots.  13 drops to fill up to a total value of 13X.

Scenario 2:  I fill all 13 gear slots by getting gear corresponding to a random slot with random quality that varies.  All slots start with a value of .5X and I stop adding gear when my total gear value hits 13X.

Loot quality varies from .5X to 1.5X:  21 drops.
Loot quality varies from .8X to 1.2X:  28 drops.
Loot quality is always 1X:  41 drops.

I think that the third scenario is very unlikely.  Getting more random drops in a single slot definitely improves that slot slowly.  .5X to 1.5X seems extreme the other way since it is highly unlikely I will find gear 50% better than my current gear and certainly not on a regular basis.  I think the 28 drop number is pretty close so I will assume that being able to sell gear when I get multiple drops in a single gear slot is worth getting twice as much loot.

Figuring out the numerical value of having the right stats rather than just good stats is harder.  Gear that has only 1 of Str/Dex/Int is only good for 1 or 2 classes but there is more than that.  Some gear gives bonuses to specific skills, some gear is class specific and some builds use different stats than others.  On one hand we have barbarian only belts that are almost never good because they are only good if they get barb friendly stats (high loot wasteage, little trade benefit) and on the other hand we have witch doctors who pretty much *must* use witch doctor only daggers (low loot wasteage, massive trade benefit) and as such the benefit of being able to get them from other people is immense.  I think I can confidently say that even if an item is designated "good" I only have a 20% chance of being interested in it regardless though this number is clearly just pulled out of thin air.

Given one number that is mathematically sound and one that is made up I come to the conclusion that having an efficient AH available multiplies the loot available to a character by 10.  That means of course that Blizzard would have to do some pretty hilarious testing if they wanted to see what it would be like for someone who farmed Inferno A1 for 10 hours and to get a sense of their gear.  Here are 250 random rares - have fun!  Having an AH is monstrously powerful for gearing up people quickly and I suspect that Blizzard really didn't anticipate just how powerful it would be.

Thursday, June 21, 2012

Magic Find math in 1.03 D3

I did some runs through D3 over the past couple days trying to figure out the best way for me to farm.  Act 1 is extremely easy to blow through and completely safe, Act 2 is mostly safe and reasonably fast and Act 3 is brutal; I die a lot and killing things is very slow.  I am sure Act 2 is my spot for now and I want to figure out how much I should be valuing more Magic Find on gear.  My current strategy is to simply clear out the act from start to finish since there isn't much difference in drop quality from mob to mob but ramping up to a 5 stack of Nephalem Valour takes awhile.  Given that situation I mostly want to know how MF behaves when I am clearing out packs with a full stack.  Right now with a full stack of Valour I have about +200% MF and I get about 2 rares per pack.  1 rare is guaranteed and the other comes randomly and in bursts and droughts.

If I add on 15% more MF by getting a new piece of gear with MF on it how much more loot do I get?  I improve my random rare drop rate by (315/300-1)*100=5% and those are about half my rares so 100% MF gives me approximately 2.5% more rares in total.  To get 15% MF on gear I generally have to give up another decent mod so I would probably be sacrificing 100 Vitality/Intelligence, 50 Resist All or some equivalent stat.  Giving up 100 Intelligence lowers my damage by about 7%, 100 Vitality supplies about 10% of my health and 50 Resist All makes me take 2.5% less damage.

I suspect that 7% more damage is significantly more useful than 2.5% more MF at generating loot because I will not only kill faster but also die less.  It is harder to compare Vitality and Resist though because they have low value until not having them gets me killed and then they have a *lot* of value!  I do die while running a2 and most of the times I die it is not getting blown away with no chance (where more defensive stats would make little difference) but rather just running out of health and cooldowns to use against particular enemies (defensive stats help this situation significantly).  I probably shouldn't underestimate the dps value of being invincible though... if I can't die I can spend more time attacking and looting and less time running and hiding and that is important.

None of this math above takes into account the value of regular magic items.  Since magic items can only be worth anything if they are weapons, rings or amulets and are *extremely* unlikely to be worth anything even then I don't think it changes the numbers all that much.  The great majority of my valuable finds are going to be good rare pieces and that is what I should focus on maximizing I think.  It looks to me like I don't really need more MF gear so much as I need gear with more stuff in general.  More damage, health or MF is going to let me get more gear in a useful way and if I get powerful enough I can just jump up to Act 3 and start farming there.  My 125% base MF seems like a good place to be at the moment.

Tuesday, June 19, 2012

D3 1.03 patch - invest in gold

The 1.03 patch for Diablo 3 is live.  There are plenty of changes to mechanics but the key things that really shake the game up are things that change the economy.  Repair bills have been increased massively - I don't know the exact value but I think it is something like a 400% increase in cost.  This is going to be a really serious money sink and will drain gold out of the economy fairly efficiently.  Of course people are running to the forums to complain about how they can't make any money because they die constantly; the solution is to fight in an area where you don't die constantly!  Blizzard clearly feels like they should arrange the game so that optimal farming is done in an area where you are challenged but aren't graveyard zerging the enemies and I think they have succeeded at this.  Coming along with this substantial gold sink is a big nerf to pots and jars and such so that just farming them for gold is going to be completely garbage; people will be farming up items from monsters instead.

The other key economic change is the introduction of powerful items into earlier and easier parts of the game.  The rough breakdown of ilvl 63 (the best, previously only in a3 and a4) item drops:

a1:  2%
a2:  4%
a3/a4:  8%

The drop tables for ilvl 62 and 61 gear roughly correspond but are more likely.  This means that if you are slow to clear higher acts and die a lot you can get exactly the same gear more efficiently from farming a1.  It looks like my Witch Doctor can comfortably and quickly clear a2 but less so a3.  I know I can clear out a3 but I will die a lot and be slow - in particular a WD has a incredible plateau of farming potential dictated by Zombie Bears.  If you can safely farm with bears you run in and explode the enemies and when you can't you have to kite with Darts and kill them very slowly.  The drop tables make the decision pretty clear to me for a WD:  Pick the highest act you can efficiently clear out with bears and do that.

The big upshot of this is that people will be getting a lot more higher tier loot and spending a lot more gold on repairs which will result in massive deflation and a price crash on the AH.  When everybody can do a stress free a1 run and get a few ilvl 63 pieces the cost of high dps weapons is going to plummet.  The same applies to a somewhat lesser extent to other gear - the top end stats available to people farming easy content is going way up.  The absolute top end gear is definitely going to cost a ton but any given piece of gear will see a serious drop in price over the next few days or weeks.  The average person will make far less gold and far more stuff and basic economics tells us what that means!

I recommend liquidating all of your sellable gear and crafting materials asap.  The gold costs on crafts are very high and with less gold floating around crafting gear is simply going to be a big loss.  Personally I plan on selling everything I find in the next few days and hoarding gold to buy stuff as the AH fills up with gear and people realize they need gold to pay their bills.  Ride the wave of deflation to massive profits, or so I hope!

Thursday, June 14, 2012

Magic Find issues

I have been doing a bunch of farming in D3.  Initially the way I farmed was to look for the most powerful gear and then go kill monsters but this changed recently when people I was playing with talked about running away from enemies to go put on Magic Find gear before the kill.  I should have thought about this before of course; the best way to get gear is to have one set of killing gear you use until the enemy is at 4% and then open your inventory, click madly on an entire second set of geared stacked with MF, and finish off the mob.  I have been doing this lately and the difference in drop quality is really noticeable.  I go from roughly +110% MF to +300% MF on the switch so presumably I double my drops.  There may be some kind of diminishing returns on MF (this has been hinted at but not confirmed by blues) but even then the difference in how much I get from a fight is immense.

The problem is that this is really annoying and immersion breaking.  I like fighting monsters and getting stuff but desperately running away when victory is imminent to change clothes feels terribly gamey.  I struggle with this because I like optimizing systems and one of the joys of playing D3 and other games like it is trying to find the best ways to improve my acquisition of gear or other rewards.  I *want* to use every tool in my toolbox to win but I *don't* want those tools to be crappy things like full gearswaps.

There are other problems with this situation beyond simply immersion or annoyance.  People are using macros on their keyboards or mice or even third party macro programs to swap gear with single keyclicks.  This means that they gain a massive advantage over anyone not doing it and Blizzard's stance on these sorts of techniques is not clear.  Using bots and other extreme hacks is obviously against the rules but mouse macros are in a grey zone; usually my rule is that if the action is questionable then I avoid it but simply clicking a button to automate a very annoying process would be really nice.

The solution to my mind is not to worry about macros.  I don't even know that detecting them would be feasible and policing the process would be a giant challenge.  Instead what Blizzard should do is to change the time when the loot on a monster is generated.  Right now loot is generated as the monster dies but if it were generated at a random time during the fight there would be no way to abuse the situation.  You cannot, of course, change the loot generation to 'when the mob is spawned' because then people would just lead off with MF gear and swap out as soon as the mob is on screen.  Generating a % randomly and rolling the loot when the mob's health goes below that % would make sure that the only useful way to get more MF on gear is to actually wear it throughout the combat.  It would mean that if people want to swap gear around for whatever reason (like putting on a shield for a tough opponent, say) they could do so and anyone could macro big gear swaps for fun but nobody would feel 'obligated' to do obnoxious gear swapping during fights.

What I really want is a game where I pick my gear, pick my spec and then go wreck some demon faces.  Right now gear swapping is a fly in ointment and I hope Blizzard finds a way to fix that.  It will lower my drops for sure but I don't even care about that - I just want the game to be fun when I am playing optimally.

Tuesday, June 12, 2012

Gaming environment

A lot of games these days make a huge focus of setting up their players' social networks in an effort to use social connections to create sales and keep customers.  Managing your friends list in an online game has become a major and expected part of the experience and we are becoming used to meeting up with people from game to game and maintaining those networks as people change what games they play.  Blizzard is certainly no different in this regard; they made a very big deal out of their RealID program and really worked hard to get all of their players into guilds and part of social webs to keep them in the game (I mean, improve their experience!  Right!)

All of this is why I am so very puzzled by the communication blockages in Diablo 3.  D3 obviously was designed to make joining other players' games quick and easy and at this it succeeded.  Unfortunately the communication between friends is really quite clunky because we don't have proper guild support or chat rooms.  While I can jump into games with friends I can't ask people to find out if they have certain recipes, offer gear, or even comment on the state of my favourite celebrity's secondary sexual characteristics.  Given that D3 made such an effort to be a group game and Blizzard is trying so hard to preserve our social networks I am really pretty shocked at how effectively they blocked our ability to communicate with each other.  Blizzard isn't new at this; they have a very customizable chat window in WOW and good guild and channel support.  Why none of this translated over I really can't say.

There are a few other features that didn't make it into D3 that puzzle me.  In particular I can't figure out how the trade window is so completely abusable.  If you change what you have in the window right before clicking accept there is nothing stopping you from scamming the person you are trading with.  Granted we are generally supposed to use the AH for selling and buying goods instead but a simple counter that blocked clicking 'accept' for 5 seconds after anyone changed the goods in the window would secure this very cleanly and that feature isn't at all new.  You can't just set up a system where people can use click scripts to scam others and expect people to not get burned by it and of course they have and are.

There are plenty of whiners on the Blizzard forums complaining about things that make no sense and complaining about things that are *extremely* difficult to test in a small, secret environment.  That certain chest and goblin spawns were too easy to get to is completely understandable and unavoidable but things like player communication and security of trades are very easy to understand and test well before the game goes live.  These sorts of issues need to be sorted out both for Blizzard's good and my own.  I will definitely play the game more and longer if I can chat with my friends effectively.  I like having a group environment and I loathe my memories of trying to avoid scammers and bots and other undesirables in D2.  These issues really should have been sorted out before launch and need to be addressed asap.

Thursday, June 7, 2012

Go where the loot is

Players will go wherever the best loot is pretty much regardless of how much fun that activity is for them.  In D3 there were a huge number of people farming up money by breaking vases in the Royal Crypts in Act 1 of Inferno.  The Royal Crypts is full of vases that have a lot of cash in them and because the monsters there are utterly trivial and few in number the farmers could safely wear nothing but Gold Find gear while they destroyed the mighty and evil Vase over and over again.  Of course they complained that smashing furniture for gold was just causing inflation and wasn't fun (but they did it anyway) and Blizzard lowered the amounts of gold that come from vases in response.  Now people complain that Blizzard is evil for nerfing their favourite farming method... ack.

The intent originally was that people would kill packs of elite monsters to build up their Nephalem Valor buff and then go beat a boss to get a big bonus from the buff.  This was intended to stop people from simply doing endless single boss runs like was the norm in D2.  20 Mephisto runs / hour, anyone?  Unfortunately this seems to have had the effect of locking people in to doing only the areas before the most farmable boss over and over and making those that can't beat the boss or who take a long time doing so feel left out.  The new solution coming in the next patch is to lower the minimum number of rares for a boss kill (assuming a full stack of Nephalem Valor) to one and give all packs of elites a minimum of one rare drop.  If you happen to really like killing the Butcher you can still go get him but it seems that by far the best strategy is going to be to just roam the world murdering groups of elites as fast as possible without leaving the game.  I think this is a good decision.  It will be very reasonable from a loot optimization perspective to fight in all kinds of different areas and that will keep the game feeling more fun I think.

The other thing they seem intent on doing is flattening the loot curve.  This is especially important for weapons which are pretty much entirely defined by how high the damage bonus on them can be for a simple blue drop and it looks like people in earlier acts will be able to get higher end weapons.  Given that the monsters in earlier acts die drastically more quickly and with fewer deaths I expect this will encourage people to farm in easier areas and make it more profitable to do so.  It should definitely have the effect of letting those stuck in a1 or a2 eventually grind up their own loot to progress even if the time taken to do so is still quite large.  The other factor that is going to make farming in easier areas profitable is that the penalty for death is going way up.  Instead of 800g / death we are likely going to be seeing 4000g / death instead.  That isn't a tremendous amount of money but dying multiple times to each elite pack is going to become a sizable drain on income and will encourage people to play in areas where they won't die much.  This might even mitigate the outrageous farming advantage that ranged classes have at the moment because a tough melee character who can clear an earlier act in complete safety is going to pay nothing in repairs and a glass cannon ranged unit is still going to misclick and die regularly, much to their detriment.  I think Blizzard really wants to send the message that the intended way to play the game is not to play in areas where dying constantly is the norm.

Initially I was pretty disappointed with crafting in D3.  The jeweler seemed completely useless beyond his ability to take gems out of gear and the blacksmith was a total waste of space.  I think though that this is changing rapidly in the endgame.  The best recipes are extremely high level rares with 6 properties and crafting them is going to be very much worth the investment.  The trouble right now is that there are very few of those recipes floating about and the lower level recipes create items that can be much more easily found on the ground or in the Auction Hall.  The crafting costs for everything but the highest level recipes could be drastically slashed and the only effect would be to allow people who play solo to more reasonably make their own gear - I think that would be a positive change.  Same goes for the jewelcrafter - every gem combine below Flawless Square should cost 0 gold and 2 of the lower quality gem to allow people to actually crush their way up to something nice.

Tuesday, June 5, 2012

D3 Inferno tuning

Blizzard has stated that they intend on addressing balance in Inferno difficulty.  In particular they talked about how damage in Inferno was supposed to be a drain on your health and how mitigating that drain was a really important point.  They want to reduce damage spikiness in Inferno in the coming patch.

What exactly do they mean by *spikiness*?

Everything in later Inferno 1 shots me on my Witch Doctor aside from weak aoe attacks like Electrified bolts, a tick or two of Desecration or junk mob melee attacks.  Generally those things that don't 1 shot me kill me in 2 hits instead though so there is virtually no forgiveness.  The thing is though that there is no spikiness anywhere.  *Everything* hits that hard.  I would describe damage as spiky if I was able to take lots of hits from monsters but occasionally be 1 shotted by a particular effect - that effect is probably out of line.  In this case though the damage isn't spiky, it is just really, really high.

Another issue is the way in which various classes deal with damage.  On my Witch Doctor I have 20% damage reduction and a ton of Int which means I take about 30% less damage than a similarly geared Demon Hunter.  The WD still gets 1 shotted regularly though and that is even with him stacking defensive stats really, really hard.  It isn't even possible for my DH to stack enough defenses to live through a hit because I would need to gear like the WD, overcome a 30%!!! deficit and then *still* get 1 shotted.  Given that I do what everyone else does which is gear purely for offense and accept that I die to any hit that lands.  It is no problem to progress through the game under the assumption that any hit kills me and gearing to survive multiple hits in a row is effectively impossible so everybody plays a pure glass cannon.  This is exacerbated by the fact that the DH gets a lot of Dodge from stacking Dex so I already avoid many attacks flat out.  My best strategy is to kill the enemies before I run out of room to kite or cooldowns to hit.

I am not complaining about Inferno being too hard.  In fact I think it is good that it is hard and it is entirely possible for everyone to progress through it with a lot of farming - you might well think that the amount of farming is too high but time is the only barrier.  Note that melee have it much worse than ranged in terms of gear requirements and I don't like that much but I don't see a good way around it either.  There has always been a huge problem balancing melee and ranged in twitch games and D3 is no exception.

There are some real problems in the game that I hope are addressed in the next patch.  I have not played a Monk or Wizard yet so I will mostly leave those alone but I have high level characters of the other three classes and there are some things that need looking at.

Witch Doctor:  Mana is a disaster.  If the WD had much better base regen they could actually use their spells in some way that made sense.  Zombie Bears would be horribly overpowered if they could be spammed without using Vision Quest though so they would likely need a nerf.  Pets aren't particularly functional either because they just die instantly to any monster.  Making them nearly invincible introduces all kinds of other problems but I have a potential solution:  Remove the cooldowns on the pets.  They would die constantly but they could be brought back if the WD is willing to spend the time and mana to do so.  This would mean that a summoner could maintain a packs of dorks but there would be a real cost to doing so - I would imagine that is a much better situation then now where they are a wasted slot on your bar.

Demon Hunter:  Nether Tentacles is absurdly overpowered.  It does outrageous damage in both AOE and single target settings, far more damage than anything any other class has available.  It needs a punishing nerf and the most likely way to do so is to give it the AOE damage penalty that other similar skills like Ball Lightning have.  At that point it would be comparable to other classes' abilities.  This single ability is mostly responsible for DH damage being so much higher than anyone else's.

Barbarian:  Rage is not working right.  The rage spenders are simply so underwhelming in comparison to the generators that it is far better to never spend rage aside from 2 minute cooldowns.  The passive that gives you 25% more damage when at max rage cements this in place.  To my mind using that passive should be a niche choice, something people do when they choose not to use rage for some strange reason - a way to make up for avoiding rage spenders.  When the standard and optimal build for the class is to completely ignore the resource the class uses something is wrong.

There are of course lots of other less pressing issues.  Every class has tons of abilities that are hopelessly bad but I don't know how pressing that is as long as two main things hold:  First, that every class feels right and people's intuitive ideas about how they should work are mechanically supported.  Barbarians should spend rage, WD should be able to cast their spells, and WD pets should be functional.  Second there should be a variety of ways to build a top tier character.  There are bad ways to build a character (Angry Chicken!) but as long as there are a number of good ways I will enjoy trying them all out and people will have fun arguing over which is best.