13 February 2013

NPC Pricing - Mistakes with NPC Stores

-
---
-----
(hypothetical conversation between a Game designer and me)
-NPCs should sell several kinds of things right? I mean the more things they sell the more likely are Players to spend Gold. If they buy from NPCs, I have designed a good Gold Sink. - Asks a game developer concerned about Gold Sinks
-Yes, NPC shops can be a wonderful Gold Sink. But defining what the Shop should sells is a complex matter. The truth is: what they sell depends on how you want players to experience Shopping and what Reinforcements they should get - I reply
-Could you be a bit more specific? - Asks the developer a bit unsatisfied
-Well, whatever a NPC Shop sells is the Reinforcer in the end of a Token Economy Chain and Gold is the Token. So, in order to not break the token economy, shops always need to have something a player wants. Shops just make players want to earn and spend Gold.
-I think I heard about something like this before... - Ponders the game developer
-Yeah, like in real world you need some sort of degree in Planned Obsolescence to make a viable business. In other words, you need to keep the buying behavior reinforced.
-----
---
-


   Well before I list 10 common mistakes when designing a NPC Store I'll need to define define 3 concepts. Unfortunately I'll have to borrow some of them from economics, a subject in which I'm no expert at all (so please make sure to point any mistake I make):


Token Economy (Psychology)
  • A specific kind of Chained Contingencies:
    • Allows you to reinforce (Increase the likelihood) of a Behavior "1" (like Playing/Farming) with by rewarding it with a Token (something without intrinsic value - Reinforcer A). This is possible because the Player knows he can buy (Behavior 2 - Buying) something he wants (Reinforcer B) with tokens.
    • In other  words setting such economy allows you to use a Token as a Reinforcer.
  • Requires:
    • Something to be the token (Gold in games)
    • Backup Reinforcement  (Things you want to buy with Gold)
    • Specific Target Behaviors (Rules for Gold acquisition)
  • Desired additional qualities: (Refer Wikipedia for detailed explanation)
    • Social Reinforcement (Consensus that the token economy is good)
    • Adequate Shaping (Tutorial on both how to get and use tokens)
    • Immediacy of Reinforcement (Token Trading should be instant for better effect)
    • Learning to Plan ahead and make savings (Token user must be Rational to make best use of Token Economy)
    • Consistent Application (Rules and behaviors must not change overnight)
Commodity (Economics)
  • General sense: "Any marketable item that satisfies a want or need" - (Wikipedia)
    • In other words, something at the end of a token economy that works as a reinforcement for Buying Behavior
    • The specific sense is less useful in this context -  "class of goods for which there is demand, but which is supplied without qualitative differentiation across a market" - but it's applicable in some games
  • Can be either a Good (Tangible - External Object) or a Service (Intangible - Someone's Work)
  • In Psychology we could, without too much imprecision, state that a commodities are Reinforcers. When a Commodity is no longer effective as a reinforcer we have reached a high Satiation level. (To understand what is Satiation check Operant Conditioning - Wikipedia)

Planned Obsolescence
  • Policy of planning or designing a product with a limited life time so buyers need to buy them again, ideally from your company
  • How this life is limited can be done in several ways:
    • The product might Break down (tooth brush) or Run Out (printer ink) after being used several times
    • A new line of products might not have backwards compatibility with the older ones
    • A product might go out of fashion (society and marketing department determines what's fashion)
    • The company adds a way to notify when to change the product for a new one (disposable razors that have a strip that changes color)
    • Another better product might be developed in somehow predictable intervals i.e. you only improve the quality a little bit with each successive release (iPhone version X)
  • In Psychology, by analogy we could say, without too much imprecision, that this phenomena is a deliberate attempt to cause Deprivation - Increasing the likelihood of doing the behavior that satiates this deprivation - Buying. (To understand Deprivaiton check Operant Conditioning - Wikipedia)


----------------------------------
--------------------------------------------------
----------------------------------------------------------
Mistake 1 - Allowing things that shop sells to be easily acquired as random loot
  • The token economy is broken (players won't buy) if they have the option to get the reinforcement as random loot with high chances
    • For example, when a player doesn't need to drink healing potions often, sometimes dropped potions from enemies can provide enough of them. This effectively eliminates the behavior "Buying Potions from NPC".
    • Another possible mistake is making the shops sell only very weak equipment,  so a player doesn't even bother looking at their selection since any dropped good would be better
  • Fix - Remove the chance to drop consumables used often
    • So players always have to buy from shop and they feel that they are good players when they don't have buy often
  • Fix - If possible, make one kind of loot only acquirable from killing enemies and another only acquirable from shopping
    • So players "feel" they need gold to get stronger, since not everything is randomly dropped
Mistake 2 - Not differentiating Buying Behavior from Farming Behavior
  • Killing monsters nets both gold and items but costs playing time
  • Buying nets an item and (free) playing time but costs gold
  • Conclusion: Shop isn't simply an alternative to Farming but a complement
  • Fix - Don't sell only bad gear 
    • The reason FF VII has stores with weak item was because it was a Single Player RPG. 
    • Bad gear can be understood one that provides the least amount of power required to be able to overcome a challenge. It's wise to sell bad gear so player don't get stuck, but it's foolish to sell only bad gear.
  • Fix - Sell Good Gear at high prices 
    • Balance Prices between the odds of dropping random loot, item power and rate of gold acquisition
  • Fix - Sell Random Drops 
    • Gambling (Diablo 2) is a good gold sink if well calibrated
    • You can even tell the player things like:
      • "This gold amount you are gambling will provide you with the best drop you'd get by killing 40 monsters"
      • "This gold amount you are gambling will provide you with an equivalent drop you'd get by killing boss X"
  • Fix - Allow sold items to have customized or controllable properties 
    • So shops provide something different from the random loot monsters drop
 Mistake 3 - No planned obsolescence in gear
  • When a player acquires the best Gear, his Farming Behavior won't be reinforced by the perspective of getting stronger.
  • Obsolescence is not a evil thing, just need to be well tunned to increase a game's longevity.
  • Fix - Use one or several of the examples bellow:
    • Limited use scrolls - town Portal scrolls in Diablo 2 - Single use effect
    • Permanent Buffs attached to Disposable Items like Gems in Diablo series - A gem can't be removed after being inserted into an item, so you need to find another one to insert in the new stronger base item you found
    • New Armor Skins - Guild Wars often released expensive armor that made players farm a lot to make an aesthetic change
    • Durability Display - Diablo tells you when to go to town to buy "Repairing"
    • Expansion Content is usually much stronger making people want to farm again - Any expansion in WoW series
  • Fix - Mix some types of Planned Obsolescence 
    • A Potion that stale after 1 day on top of being consumable item
    • A temporarily single use buff (both time and quantity limitation)
      • Could be more Experience Gain, More Drop chance, More damage dealing
      • Can even mix gameplay elements, instead of having duration based on time, it could be based on behavior like losing a buff when you are hit or take too much damage
    • Ragnarok Online had a Blacksmith that could improve weapon with up to +10 enhancement - That made people buy the forging service again each time they found a weapon with better base stats.

 Mistake 4 - Using a durability mechanics just because it makes sense
  • Durability is just a Immersion feature. The designer makes the rule of how long the virtual item lasts and when the durability is used.
  • Durability can be an powerful gold sink, but it's aversive due Repairing being Negatively Reinforced Behavior. In other words the only reason players repair their gear is to not lose it or die. As a rule of thumb you should avoid setting Negatively Reinforcement Contingencies.
  • Losing Durability (usually from playing he game in any way) is a Positive Punishment. No one likes positive punishment.
  • Fix - Remove durability by changing gold input - Make the quantity of gold a player earns by killing monster be proportional to their level difference.  
    • This is much better than making the gold monsters drop and gold required to repair scale with both Monster level and Equipment level.
    • Basically when level increases both Repair Costs and How much gold a monster drops you are effectively making players earn less gold anyway
  • A game doing it right - Farmville uses something exactly like durability but with another intention and much better effect:
    • In this game Plantation Crops get rotten if you don't collect them after 24 hours (or some other time interval)
    • Players can avoid losing their crops if they collect their crops in time (Shaped behavior: Playing often) - Provides a way to avoid the punishment
Mistake 5 - Not regarding services as a valid commodity
  • Like I was trying to point in the introduction, any commodity is a valid Reinforcement, and NPCs can sell services
  • The more things a player wants to spend gold on, the better currency gold becomes
  • Fix - Here a few options you may or may not know: (check which could be used in your game)
    • Charging Gold to Repair item (Score 2/10)
      • This is just a creative way to put player under a Positive Punishment Schedule (Forcing them to lose durability and therefore Gold)
    • Charging Gold to Replenish Mana or Resource (Score 2/10)
      • It's very unpleasant  to pay to use a resource to play
    • Charging Gold to give temporarily buffs (Score 6/10)
      • Make sure that the buffs aren't required to be able to play or too powerful/useless
    • Charging Gold to improve a potentially disposable equipment (Score 9/10)
      • Make sure game is balanced around the improvement
    • Charging Gold to Gamble Gold (Score 4)
      • This is pure gambling like in real world - well at least you can make sure the game is honest
    • Charging Gold to providing a opportunity to participate in a run (Score 5)
      • It's an obvious gold sink
      • It's pointless if the Run gives more gold than it's required to enter (unless your game is designed to make players fail often)
        • Fix - Make the runs with gold requirement reward no gold and have exclusive drops
Mistake 6 - Not regarding time saving services as a valid commodity
  • It's possible to make gold buys something usually earned by playing
  • Players respond well to using Gold to save playing time (Grinding/Farming) tough implementing it into an existing game might be not as well received
    • Charging Gold to Give a Random Drop - Gambling Gear (Score 10)
      • This is the perfect conversion of Gold into Playing time
      • Effectively makes gold into money since items are the most sought after commodity
    • Charging Gold to make a map (Score 7)
      • Essentially a specific type Buff
        • Can be useful for maze or randomly generated dungeons
        • Alternatively can provide information like monster/chest positioning
    • Charging Gold to grant Experience Points XP (Score 6/10)
      • Make sure playing behavior is reinforced by other factors like Drop chances
      • Balance your game around the fact that veteran player can spend xp (allowing players to create different characters each needing XP is a good system)
    • Charging Gold to providing a opportunity to get a rare drop - Usually some sort of Monster Arena (Score 6)
      • Essentially a improved version of the idea above and a improved random item gambling
    •  Charging Gold to transport players to a specific area inside the game (Score 4)
      • The only way to make this work well is to make the very experience of getting to a point inside the game unpleasant
 Mistake 7 - Not making good use of purely aesthetic products
  • Players like to look cool or pretty
  • People find it fair to spend money on good looking clothes
  • It's prestigious to wear expensive clothes
    • Higher prices actually make buying Aesthetic more prestigous
  • Fix - Use an algorithm making price increase or decrease based on rate of acquisition per item interval
    • I'll show how to make such algorithm in the article: NPC Pricing - Prices based on Collective Behavior
  • Fix - Program new different looking skins every so often
    • You can even charge real money world money instead (League of Legends)
 Mistake 8 - Not making good use of Limited Edition Goods
  • In real life Limited edition goods make people:
    • Pay more - so they will have that exclusive product
    • Buy quicker - so it won't run out
    • Buy more units - so they can use it more times
      • Fix - Make limited edition goods with planned obsolescence 
 Mistake 9 - Making bad consumable items
  • If a player sells a consumable item to a NPC instead of using it, then you made a bad consumable. Or the decision to drop that consumable as random loot was pretty bad.
  • Regardless of the case a sold consumable must have a coherent price and effect
  • Fix - Make dropped consumables always an advantage to use
    • Healing Potions are a bad drop because you can buy them at shop, and need to use them frequently. But a potion granting +25% experience points 5 minutes is never wasted
    • Use dropped consumable as a free sample to convince players to buy it from shop - in fact make dropped potions have no gold value
  • Fix - Make consumable grant optional buffs
    • Healing potions are perhaps the worst used mechanics in all RPGs. They are as bad as repairing, when you consider the player behavior - Drinking Healing potions is a Negatively Reinforced Behavior (Drink to avoid death).
    • A Game doing it right: Guild Wars made restoring health a gameplay element instead of a passive "potion alcoholism" we see in Diablo series
Mistake 10 - Being careless with gold distribution
  • The reason so many gaming economies get bloated with gold is the bad balance between gold rewards and gold sinks - especially NPC shops
  • Fix - Make an algorithm changing Gold gain
    • Gold gain doesn't need to be static based on monster level, for instance it could also be based on the difference between monster level and player level
  •  Fix - Don't make all enemies or Runs reward with Gold
    •  Even tough making everything potentially reward with gold makes gold a good estimator of how long a player "Farmed", players can farm for other reasons
  • Fix - Make global gold reserves variate over time
    •  Well, real life currency has varying buying power, why can't you variate the gold inside your MMORPG behave like that?
      • Call it "Inflation" and reduce  all current gold reserves from all players and Gold gains by 5%
      • Make the opposite sometimes so it becomes more like a real market
      • Heck you could make such adjustments using some algorithm taking Playing Hours into account
  ----------------------------------------------------------------------------------
--------------------------------------------------------------------- ---------------------------------------


   Well, with these solutions and possible problems cleared for next time expect an article about Algorithms for Individual Player Behavior and then another for Algorithms for Collective Player Behavior.


References: