WoW crafting

Now that the WoW token is out and seems to be stabilized in price here in the US, well, as stable as something that osculates at 25% of it's own value about every 24 hours, at least. I'm getting back into crafting to "earn" my account's subscription. To that end I'm building up my Warlock's Garrison. I originally didn't want to do that, but I'm getting sucked in by the lure of gold. I can see Dr. Skinner laughing at me from outside the box.

This isn't a "How to pay for your account with crafting" post. You'll have to figure that out yourself. No, this is a "What I think of crafting in WoD and how can it be improved." post.

Gathering in WoD is done. All crafting materials are obtained by cooldown in your Garrison. I can see why they would do this, it essentially eliminates botting and resource grinding. Sure, you COULD still grind ore, leather, etc. But it's pointless in that few items that you would craft to sell are made from those, instead, they are made from "work orders" and other cooldowns in your crafting buildings.

You may still be able to grind "old school" mats, and there may still be a market for that, but I'm more interested in WoD.

The first thing that you notice is that once you run your Garrison maintenance, and manage whatever auctions you have in the AH, you're basically done for the day. With 2 Garrisons running to support 4 crafting buildings, 30 to 40 minutes is all I need to get everything done. Now, I'm still going to have to trap animals for the barns at some point, but my guild runs "trap groups" once a week or so. Trapping in a guild group is the way to go. It's a fun guild activity and you get all the trapped animals you could ever need.

In my opinion, Blizzard has made a huge mistake with the Garrisons by allowing you to have more than one of each type of crafting buildings per account. I can't have a problem with allowing you to use crafting buildings without the associated crafting skill, as I think that being limited to 2 is somewhat arbitrary. But the crafting buildings should be account unique. You are already prevented from putting two of the same kind in one garrison, extending that to include the entire account doesn't seen unreasonable. If you want to set up “Craft sweatshops” with multiple craft buildings all blasting cooldowns at once, you'll need an account per set of unique buildings.

If you replace “Grinding mats” with “Daily cooldown for mats” then you need to limit the number of cooldowns you can do to one set per account. It comes down to “How many of an item can you make in a day with the cooldowns.” Let's use Blacksmithing and the “Truesteel” line of armors for my example.

If you're thinking “Bah! 640 iLevel! That's junk you can get from LFR!” then you're not looking at the big picture. The 3 iLevel 640 items you can wear are for alts leveling from level 91, and there are iLevel minimums to get into Heroic Dungeons, LFR, etc. That's what these are for. Then, after you get most of a set of raid gear, you can use the crafted ones, boosted with the crafted booster items and rerolled with the stat changing item, to fill in those few slots the RNG didn't seem to want you to have.

So as long as the AH isn't flooded with them from craft farms, they can sell for decent gold. You can do 6 work orders a day, getting 2 Truesteel ingots per when a Follower is working. That's 12. Then you get 10 from doing the daily cooldown at max level. That's 22. On top of that, you can convert Primal Spirit (Along with some of the Mine generated mats.) into a few more per day. Blacksmithing is probably the easiest for that last part in that each Truesteel Ingot made requires only 5 of each of the ores the mine produces. You can typically get 10 more per day from a level 3 mine. That's 32 Truesteel Ingots per day. Per Blacksmithing Garrison.

That would be fine if you could have only one Forge building per account. In reality? I think I'm even being TOO generous with the cooldowns here. Look at my two Garrison setup, 2 barns, both Garrisons have mines and herb gardens. That supports 4 crafting skills with all their related daily cooldowns. And that's just 2 Garrisons. A lot of people have 4 or more.

There is a movement on the WoW forums calling for one account wide Garrison. I think that would be a mistake as well. I think the way to go is to make the cooldowns themselves account wide. If the design intent is for you to have 2 crafting professions at full speed, then limit the account to 2 crafting professions worth of cooldowns.

You could even take it one step further, having only one level 3 Garrison per account. If you did that, however, you would need to allow your “alts” to visit your “main's” Garrison for the special vendors.

The ability to craft things has to cost you something. Either time spent grinding (Which causes RMT and botting problems.) or money in the form of extra playing fees.

I like taxis and teleporters

In pretty much every RPG type game, there is a way to get from point A to point B fast. Usually, you have to "explore that area first" before the warp pad / travel rune / portal spell is available to you.

Then I played Eve Online. No teleporters. You could "autopilot" your ship from one system to the next, but that was stupid in that it made you a sitting duck. At first, I thought "Ok, this is different, I can learn to like this."

But there were two overarching problems. First, if you needed something from the trade hub (Jita was closest to me, at 12 jumps away.) you had to travel for up to 30 minutes to get there. What happens is most people "live" near the hub. With the hub itself being a ridiculous mass of chaos. Going to Jita the first time wasn't so bad... but after a while it was a time sucking pain even in my fastest ship, and if I had to take the Orca, youch.

Taxis and teleporters just allow you to get to where you're going fast and without stress. Idiots can't set up traps for you to fall into along the way, and you get back to what you were doing in a hurry. Can you imagine having to take a land route in WoW to Stormshield from your Garrison? What if you had to do that every day? You'd go nuts!

Making a better market

I want to analyze "markets" in MMORPGS in an effort to design a better one. I'll point out specific traits of some games, but I'm not going to go through the list, review style. I'm more interested in the player's needs and motivations.

First up, you need a center of commerce. Whether that is an "Auction House" or a open air market in the center of town or whatever, you need a place players can go and quickly find everything they need, and sell the stuff they have. You should have all items of a single type in a single place, be that a single vendor or a single interface to a larger market system. You should not have to go to 5 different places to see all the offers for a single item you want. Selling your items should be easy and you should be able to get a fair price for them. Again, you should not have to go to 5 places to see all the offers to buy your item.

Ok. Now let's look at the players. Most players just want to buy and sell their items quickly, easily, and for a fair price. Some people just want to get rich. Ok, that's fine. If you can accommodate these players while not harming the first group, that's all good. But if there is conflict, the first group is deferred to. Lastly, there is a group that simply desires control over a specific market. These people are toxic to both of the other groups in some way, and should be discouraged.

Before I continue, it's important to clearly specify why the “control the market” group is being treated so roughly. There are two ways to “control a market”, monopoly or flooding it with cheap goods. Monopolists are easy to stop, but there is nothing you can do if someone is dedicated to flooding a market, and they have the ability to do it. A flooded market is great for people trying to buy the goods, but is terrible for people trying to sell their goods. You can slow them down, but not outright stop them.

Here are two examples of that. Someone wants to control the raid flask market in World of Warcraft. So they mercilessly farm herbs (Or use a bot to gather them.) then undercut everyone else, possibly even to the point of selling the flasks for less than the herbs would have sold for by themselves. A counter to this would be to put in a way to “salvage” the crafted item so it can be converted losslessly to it's component parts. Of course, if the component parts are being botted into the same state, the botting would have to be addressed as well. The second example, also in World of Warcraft, involves crafted armor. You craft them with components made with daily cooldowns in your Garrison, but you can have 11 Garrisons on a server, each with their own crafting buildings. The counter to this is to limit a player to one Garrison, or to one crafting building of the same type across all Garrisons and one cooldown of the crafting mat per account. Of course, if the player starts multiple accounts to have more crafting buildings and cooldowns, there is nothing you can do about it. Some people are just that driven by the need to control.

Back to the market. The next thing you need is a known normal price for every item. Eve Online does this by region, with an average transaction price for every item across that entire region. When you attempt to buy or sell an item you see the current average price. Very nice! World of Warcraft is the exact opposite, you see the “vendor price” that was assigned to it when the item was added to the game, which is almost certainly absurdly low. You are then faced with a complete lack of knowledge about the item. The exception is if you're a dealer of that item, or if there are enough of them in the market.

Next, there is the problem of unlimited craftability. If your system allows you to make as many of an item as you want, and doesn't allow that item to be salvaged back to it's component parts, then the result is markets flooded with those items. The way to make this even worse is to require the player to make a lot of items to increase their crafting skill. I like crafted items. I think most if not virtually all items in game should be made by the players themselves. I think it's absurd that some npc with the apparent intelligence of a rock living in a cave somewhere can miraculously have access to equipment superior to anything your civilization can produce.

Last, the interface. I want to have a search window that lets you select two categories, like “Armor” and “Leather”, then a text search that refines further. (Use the text search at any time if your search phrase is specific enough.) Then each item is in the form:

Item name Average sale Qty Best buy price Buy button Display details

Item name is also a mouse over that shows a tool tip detailing the useful features of the item (stats, armor values, materials it's composed of, etc.) Average sale is the average of recent sales of that item. Qty is the number of them currently for sale. Best buy price is the the price of the lowest price item in the sell list, Buy button pulls up the “Buy” dialog, where you can buy the cheapest or set up a buy order, and the Display details buttons pulls up the list of recent sales that the average is based on, the lists of buy and sell orders.

So far, this is like Eve Online, but with a slightly different interface. And shorter orders.

To sell an item, drag and drop it into a box at the bottom. That pulls up the sell order interface, where you select Start price and price rate drop. All sales have a 48 hour duration, and can be extended for 24 more hours after the first 18 hours. You can extend them forever. The cost to put in the sell order is 1/4% of the item average price and to extend the sale, 1/4%. You pay a fee of 5% when the item sells. If you cancel or allow the listing to expire, you get the item back. The defaults are Start price is average price + 10% and the start price rate drop is 5% every 24 hours.

When you extend the sale, which you must do once a day if you want the item to stay listed, you can either “let it ride”, reset to default (!0% over, 5% drop per 24 hours.) or enter new values. You need to actively monitor your market orders by checking in on them once a day. (Note: If you want to just change the settings without extending the sale, you can do that any time, but it still costs you 1/4%.)

Buy orders work similarly. Except the default is 10% under average value and drift is 5% per 24 hours up. Otherwise, the same rules apply. When you put in a buy order, the maximum possible sell price is taken and goes into escrow. There are no fees for buy orders.

If a buy order and a sell order drift into each other, a sale is automatically made and both orders are removed from the market.

If you're coming in blind and just have an item to buy or sell (The majority of people.) the best strategy is just accept the defaults and wait up to 48 hours for a sale. Of course, most buyers want their item now, and will likely accept the best current price if it's not too far from the average. If you're selling items for a living, and you know the market, you can put in carefully thought out initial values and then extend the order once a day for the duration you expect it to go.

If you start a buy or sell order with the defaults and it doesn't transact, then the market is either stale (Very few buyers.) or it has drifted off average too far. Resetting to defaults with adjust to the new average and extend the order.

If the defaults are tuned right, and the fees tuned right, the result should be a system where the majority or people trust the “average price” and either buy / sell at best price or put in a contract at defaults. The result should be a self tracking average price tracker that is difficult to manipulate.

Back to the future, paying for time used.

How would you design a F2P system that actually works? One that avoids the moral hazard of gold selling and also avoids the problem of market speculation?

Obviously, real world money has to change hands at some point, you can't run a game for free, so someone has to be paying for it. At the same time, the possibility of playing for free or close to free is appealing.

Naturally, there is a trial period where you download the game for free and are granted a trial period, say 2 weeks, where you can play free under some limitations. The intent is you can try it, and if you like it, you subscribe.

Once you subscribe, which basically costs you a one time fee for the first month of game time and removes the trial account restrictions, you have the option of continuing to pay a monthly fee, or use in game resources to pay for your account. Essentially, you introduce a gold sink into the game that you can, at your discretion, use to keep from paying real money.

Obviously, the game is not truly “Free to Play” because you have to, in essence, pay for one month of time to unlock the game from the trial version. Being unable to buy game time in game is one of the restrictions of the trial account.

So, how do you buy game time? You buy it from the web site. You never allow “partners” to sell it with the exception that, if your game is massively popular, to sell “game time cards” that are redeemed at the web site. But those cards never have a “multi-month” discount. This is your income from the game, people buying game time from the web site. Or you buy it with in game gold. A percentage of your players will be playing for free after that initial purchase. How much you charge for each “month” of time, and how much subsequent time costs will determine the percentage of free players.

To buy game time with in game gold, you just buy it from an in game interface. There is no token. There is nothing to speculate with, there is nothing in your inventory. You go to the interface, you select the months of time to buy, you enter payment information, you click Ok. Done. When the transaction is cleared, your game time is updated.

It is just that simple. Except for one thing. RMT. Some entrepreneur is going to come along and pay someone to farm the hell out of your in game resources and sell them to other players who want to use the gold to buy in game time. Or another entrepreneur will come along and figure out how to build bots to do basically the same thing. This works because once you pay for a month of game time, you can mercilessly grind resources 24/7 for the same cost as the player who plays a “normal” amount of time. Or, set a bot loose to do the same thing. Tracking and banning this is a huge Customer Service cost.

The instant you have any “gold sinks” in the game, you're begging entrepreneurs to abuse them through RMT and botting. Both of these can be fixed with the game time system.

While you are in game, you are doing one of three things, Typing in the chat channel, traveling from place A to place B for whatever reason, or using in game skills and abilities to perform useful tasks. Every skill and ability in the game has a period of time associated with it's use, and that cost goes against your purchased game time. If you use your time in low efficiency tasks like grinding or botting, it costs you more than you can get back. Making it pointless and stupid to bot. Activities that are “economy neutral” like questing or exploring, have a low cost, activities that have a high economy cost, like gathering resources for crafting, have a higher cost. You can't limit people from an activity with the heavy hand of stopping them from doing it, you need to just make it more expensive for them to do the things that are more expensive for you.

When you buy game time, you buy the amount the average player uses in a month. If you are doing low impact activities, basically using the game as a time sink like watching TV, you will be using your time up more slowly. If you're furiously grinding resources, you use your time up much more rapidly.

Another mechanic of note is the “daily cooldown” to obtain crafting components. I like daily cooldowns. They give a way to allow producing an item with a rarity that should require many hours of effort, but can be made quickly. This gives a crafter a starter income and promotes crafting, but it only works if you limit the cooldown to once per account. In WoW, you can have 11 Garrisons, and with them 11 sets of cooldowns going at once. An optimal system would give you a daily cooldown, then allow you to farm for more, but the cost of the “additional materials” is greater than the initial amount gained with the cooldown.

You might be thinking “But isn't this essentially pay to win? With richer players gaining an advantage?” Well, that's unavoidable. All you have to do to “Pay to win” in any online game is have multiple accounts that work symbiotically with each other. However, when you have multiple accounts, you have to pay for more subscriptions. Think of the increased fees on the heavy users as the same as playing several accounts to get the same thing done.

Post update change.

Adjusting my posting schedule slightly.

No posts on weekends except for special occasions.

Will be back Monday!

Retro skill based system

I'm trying to design a system where instead of having a single level for your character, you have a level for each skill. A "skill based" system, if you will. I know this has been done, and with some degrees of success.

So, my first thought is you can't have hit point inflation. You can't start with 100 hit points at level 1 and end up with... say... 200,000 hit points at level 50. You would have much less at level 50, say 150 hit points. There needs to be SOME increase, if for no other reason than to telegraph to a stranger what your highest trained skill is leveled to. So let's start there. You have no race or class, all characters start with 100 hit points and gain 1 point per skill point in their highest skill. Just for kicks, let's cap the skills at level 50.

When you start out, you talk to a series of trainers that give you starter skills for free. You have the skills to kick any level one npc's butt. Let's just arbitrarily create some skills: For melee combat, we have main hand dagger, dodge, and parry. For spells we have damage spell (keeping it generic, here.) and spell resist. We'll have all skills use the same resource pool I'll call "stamina" that starts the same as your hit points.

Ok, so you kill the low level mobs with your dagger, leaving the damage spell and spell resist at level 1 because you like to stab stuff. You're now at level 3 in daggers, level 2 in dodge, and level 2 in parry. Dodge and Parry didn't track with daggers because you don't always use them. If you always choose the easy fights (enough to get skill in daggers, but not enough to risk great injury to yourself.) your dagger skill will always be higher. The other 2 won't catch up until you start doing harder content.

So you're ready to tangle with the npcs with level 4 dagger skills. But what would happen if you cross the street and start something with the level 4 spell caster npcs? If you could get into melee range, you could get a really good hit rate because they have no defense against daggers, but at the same time, you have no defense against their spells.

All kinds of choices open up now. Do you tune the npc's to expect cross trained players? Some players would just charge ahead with one skill until something stops them, others would cross train from an early point. Of course, each skill gets harder and harder to train as you go, and the support skills become more and more important.

I like this simple start. You train skills as you use them, your hit points are never more than 150% of anyone else's, forcing damages down across the board. What keeps you from getting into a fully one sided battle is that you'll never hit, and they always will and for maximum damage (Albeit a maximum damage of about 50 per second.) You can learn any skill you want, but they all use your limited stamina pool, preventing you from from doing everything at once.

EU WoW token launch

Previously, I expressed my opinion on RMT and my hypothesis that the WoW token was not really subject to supply and demand as they are normally applied. That was a week ago, go back and read those posts if you haven't. You'll need that info as a starter point for this one.

Again, I'm referring to the WoW token as two separate tokens. The "RMT token" that you buy for cash in the Item shop, and the “Game Time token”, which you buy for gold in the AH. All the price fluctuations take place on the AH side. The RMT token's price is fixed.

Just under 2 days ago, the WoW token was released in the EU markets. As with the US one, the site www.wowtoken.info immediately started recording price data. The data so far is both completely different, but at the same time the same. Here is the EU chart:

Notice that it's the same kind of waveform, a sloped line with a discreet slope. You can easily see the points at which the slope changes. It hasn't started forming a triangle wave yet because it hasn't changed directions enough to stabilize. But what we can determine from what we know is that the slope of the price change rate (What I will call the drift) is based on how long they set the intervals to. The rate of change is always 1% per interval, and it appears that they can set the interval is 15 minute increments. It has been as low as 15 minutes (When the US graph dropped very fast near the beginning.) and as long as 2 hours, which is where the EU graph is at now.

Right off, you can see that the EU prices are higher. And the simple reason for that is they started higher. In both cases, at launch, the slope of the drift is set to "up" driving the price even higher.

Think about that for a minute. At launch, buying a Game Time token was impossible. None of them existed. By any concept of “supply and demand” the trend price had to be downward at that point. As you had to have a supply before demand could step in and raise the price. But in both cases, US and EU, the price instantly started drifting upward at a 1% per hour rate. Then, at some point, the drift direction changes and the price drifts down at the same rate.

In the US graph, whoever is controlling the drift rate decided it was too slow, and cranked the interval down to 15 minutes, which set the drift to 4% per hour, it dropped like a rock until they set it back to 1 hour, where it has remained. In the EU graph, whoever is controlling the drift rate decided to set it up to 2 hours per 1% change, slowing the decent of the price. Why? I have no idea. 1% per hour seems like the perfect rate given the apparent rules of the drift. I would have designed the system to use “% per hour” where the % could be any value, but that's just me.

As a result, the EU price had just continued to drop at .5% per hour. It will likely do this until it reaches whatever point the controllers want to stabilize the price at and then start the triangle wave like we see on the US graph. It will probably be several iterations of an asymmetric triangle before it finally stabilizes at... Oh I don't know, 25K to 30K gold per token?

So. What determines the point at which the drift changes direction? Of course I can't know for certain, I can only speculate. but I think they monitor the time it takes for the Game Time tokens to sell and then try to keep that reasonably low, like 12 hours or less. Essentially focusing only on supply. Remember that the point is not to sell Game Time tokens, but to sell RMT tokens to combat illicit gold sellers. That is the WoW token's reason for existing, and is the logical reason for adjusting the price.

Once that timing pattern is stable, then there is little reason for them to not partially delink the RMT tokens from the Game Time tokens by giving gold for the RMT token immediately and just adding a Game Time token to the AH for eventual sale. They already know it's GOING to sell, and probably within 12 hours... so while this would be “directly buying the token by creating gold” they know the gold will be remove when the RMT token actually does sell. It's a futures contract. Now you can see why the sale price of the RMT token on the AH was fixed at listing, rather than getting what you got at sale.

This scheme works as long as the true time to sell the RMT token is stable. If it's open ended (I.E. gets longer and longer) inflation results. If it's closed (I.E. the demand for Game Time tokens exceeds the supply and the true time to sell is zero no matter WHAT you set the sell price on the Game Time token to.) then deflation results as the Game Time tokens cost Blizzard real money in subscriptions and become a gold sink in game.

The inflation side can be prevented by simply holding the price steady and only selling the Game Time tokens as they become available. No problem! The RMT tokens sell immediately, which is good for the fight against illicit gold sellers, and the Game Time token buyers have all month to buy one, so the delay for them is less of an issue. It might happen that not all of them get one, but that's life, and can be blamed on “hoarders.” The Deflation side is win win. Illicit gold sellers get slammed and a gold sink to get rid of what is clearly a lot of excess gold is created.

This is looking like a really clever design on Blizzard's part. Assuming they can avoid the moral hazards of directly selling gold.

Classes, races, and attributes

I worked at a front line online game company for neary a decade as a programmer. One day, we rented an entire showing of "The Lord of the Rings" when it first came out. That was great fun, just our crew in the entire theater. Now, I had never read any of Tolkien's books, but I was aware they were written in the 40's.

As we were leaving the theater. I turned to one of our line producers that was walking next to me. I asked him "The entire online game industry is just completely cribbing off Tolkien, aren't they?" He just smiled.

A lot of things are the way they are just because "That's how RPGs work."

It doesn't have to be that way, of course. My problem with attributes, classes, and races is that they, by their very existence, force your character to be pigeonholed into a single path of ability and usefulness. If you want to branch out and try something different, you have to create an entire new character instead of being able to reuse the components of the one you already have that the new path shares with the old one.

Now, one concept that I do really like is the trinity, the concept of tank, healer, and damage dealer working as a team. Of course, for simple encounters, that team could be all one person... but as the encounters scale up, players need to be more specialized in their roles. But there should be nothing stopping them from putting the role they've been doing "on hold" and training to do one of the other roles. Now, if you WANT to have a completely separate character doing that other role for “role playing reasons”, that is your right. But you should not be forced to do it that way.

This sums my feelings up in one sentence: "Forcing a player to choose a path before they even know what they are doing is bad."

Pretty much every advantage I can think of in forcing a player to choose a race and class, and by extension their associated attributes can be countered by one question: "You can't think of a way to do that without forcing a player to choose a class and race?"

My two favorite arguments for classes and races are:

"But if you let everyone do anything, they will all just use the most powerful skill." To which I say "And who's fault is that, exactly? The guy who chose the dagger over the sharpened stick? Or the guy who designed the choice?"

"Choices must have consequences!" To which I ask "Why? It's nearly impossible to choose perfectly. Let them fix their mistakes."

Auto incrementing buy / sell orders

World of Warcraft's Auction House only allows you to put in sell orders, and then only with a "bid" price and a "buyout" price.

Eve Online's "Auction House" (Simply referred to as the "Market") is different, you can have sell orders AND buy orders, but there is no bidding. The price is the price for the duration of the listing unless the owner of the order specifically adjusts it.

And in both cases, adjust those orders they do! Both games have armies of "traders" that sit in the Auction House, Station dropping the price on the items they are selling by 1 copper / .01 isk in an effort to make THEIR item the "next one to sell."

Neither market has an inherent mechanic to find the "Optimal price" where the law of supply and the law of demand meet to find the price of the good at the quantity it is available at. And by "inherent" I mean the function brokers perform in the real world, connecting buyers to sellers. In game, you interact with each other more or less directly. You could argue that "traders" perform the function of brokers, but they do not, they are either speculators or profiteers. Eve has better tools, with a graph that shows price fluctuations over time, and a constantly adjusting "average region price." These are nice features.

Here's a possible solution that would work with the current systems to streamline that entire process: Auto incrementing buy / sell orders. Now, to be fair, this is just a thought exercise. I'm not convinced the AH / Market systems can really, fully work to produce true price equilibrium in all cases. The basic problem is the laws of supply and demand must be applied symmetrically. As a market gets smaller, the ability to accurately assess the demand decreases in relation to the ability to accurately assess the supply.

The system would track "current price" as well as allow buy and sell orders as Eve does. But no "bids" like WoW does. It's not an "Auction House" any more, it's a brokered trading house.

When you enter a buy or sell order, you are, of course given the option of just immediately accepting the best available price. Or, you put in an order. The default is to have the order be 10% above or below current (above for sell orders, below for buy orders.) with a drift rate of 1% per hour and a duration of 12 hours. In a market with high volume and perfect price equilibrium, it would take 10 hours for your item to reach the spot price and sell. If the item does not sell, it's delisted at 2% below current price. You can, of course, adjust from the defaults on an item by item basis, or change the base default to fit your strategy if you do a lot of orders.

The market interface, by default, would have one line per item type as opposed to one line for every lot of items for sale. That line would tell you the current average price, the current buy / sell price, and the number of items for sale. You should be able to sort the list by any of those criteria, or an item name search. A “show details” button gets you to a screen that shows the current price on individual lots, averages charts and sales volumes.

You might be thinking “Well, I'll just game that by setting MY default start point to 2% over the current price.” But if you try that, you realize why this works: buy orders are also drifting upwards to meet you. If there are enough buy and sell orders drifting towards each other, the most logical course of action is to just accept the best available price. The true purpose of the constantly adjusting prices is for speculators to indirectly influence the current price. The aggregate of their actions sets the “actual” price for people just playing the game to easily buy and sell their goods for whatever the actual price is.

Let's “what if” some outlier scenarios and see what would probably happen:

Case 1: Speculator buys out a low volume market and relists the items at a higher price. People who want to buy see this because all the items are above average. They put in a “buy order” at the default, 10% under drifting up. People who obtain new instances of the item and want to sell them see what's going on too, they set THEIR item to 20, 30, even 50% (somewhere between current and what the monopolist is trying.) These prices will merge towards center and either chase the monopolist out, or be bought by him in turn. But he can't burn money forever. The system's “perfect information” immediately informed buyers and sellers of this manipulation attempt and gave them an easy tool to deal with it.

Case 2: Way too many of an item are being listed to sell for a price that is much higher than the equilibrium, or “true market” price. By allowing the market to automatically trend towards equilibrium instead of being propped up by mass greed, the price in constantly pressured towards equilibrium. If that price is so low the secondary market is attractive, the excess inventory is removed that way. Also, less items are crafted / farmed if the ability to convert them into currency is reduced.