Ramble: Auction vs. Consignment


I recently started playing City of Heroes again to check out Going Rogue. One of the first things I noticed on coming back was that various changes to how money is earned, the merging of the markets between heroes and villains, and the growing number of max level characters with nothing better to do than farm currency meant that my assumptions about how much it should cost to “gear up” a character were out of date. I started reading some market strategy guides out of self defense, and found them very interesting… primarily in that my assumption about how the player to player market worked was fundamentally incorrect.

But first, keep in mind that the dominant player market system in MMOs is World of Warcraft, which has an auction system. Unless it’s changed a lot from when I played a few years ago, it works a lot like eBay. A player lists an item along with the minimum he or she will accept for it, whether he or she is willing to let it go for a “buy it now” price, and how long the listing will be available. Players see the current winning amount, and know that, if they want the item, they will have to beat this amount and hope no one else comes along offering more money before the listing expires. Impatient players can spend the “buy it now” price and get the item immediately, but for potentially far more than it could be had if they were willing to wait for the auction to expire. Players have to rely on 3rd party sites or their own research to track what the current average sale price for the item is, or, if it’s a common item, they can simply bid on the one that’s cheapest at the moment.

Most MMOs that I’m familiar with use a similar auction system, though often less polished or complete than WoW’s.

City of Heroes is different, in that it uses a consignment system. Players that want to sell an item simply list it and the minimum price they are willing to accept for it. Players that want to buy an item select it from a list of all available items and enter how much they’re willing to pay for it. For slow-moving items, there may be only one person buying or selling at any given time. If you’ve listed an item, as soon as a buyer requests it at a price equal or greater than your asking price, it’s sold and you receive the amount spent. If you’re a buyer, as soon as someone lists your requested item for equal or less than you’re offering, you get it for that price. There’s even a list of the last five sales to show both parties what the going rate is for an item (this is far more accurate for rare items than for common ones: if an item is selling in huge amounts, the last five merely displays a constantly shifting snapshot of current demand, and can easily skew the perceived value of an item by a single person paying too much for several of them).

What I hadn’t understood about the consignment house until reading the market threads was the method the system used to match buyer and seller when there are a lot of bids and/or a lot of sellers. When something only has 1 buying or selling, it’s very easy to (barring listing fees for the seller) figure out through trial and error whether the seller is willing to sell for the price the buyer is willing to pay, even though the transaction is completely anonymous. Things get more complicated when things are selling briskly: you may have to bid 1,000 to immediately get something that has a thousand for sale and usually goes for 100. Meanwhile, you might be able to bid 20,000 for something currently selling for 100,000, and get one in a few minutes. From a seller’s perspective, you’ll sometimes list something for what appears to be the going rate, and fail to sell it for hours or days even though the sales are still turning over at roughly the rate you listed, sometimes lots more, and sometimes you’ll list something for a pittance and receive way more than you expected to get.

I had assumed that there was something complicated going on with first in/first out based on time of listing being compared to amount offered. The common wisdom on the forums is that it’s much simpler than that: bids are sorted from highest bid to lowest, sales are sorted from lowest list price to highest, and the two are paired off until the highest bid no longer is enough to get the lowest list. If there are three people bidding 5,000, 4,500, and 4,000, the item listed at 1,000 will get 5,000, the item listed at 3,000 will get 4,500, and the item listed at 5,000 will get nothing (because the remaining 4,000 bid is insufficient, even though it would have matched perfectly to the original high bid). This goes a long way to explaining how there can be certain items that will have hundreds of the same item both listed and bidding: high bids don’t peel off the high list prices until all the cheaper items are sold. Meanwhile, theoretically you can list an item at 1 and get the maximum amount currently on offer (though you may have no real idea how much that is).

I’m genuinely curious why the market was designed this way. If I’d set it up, as mentioned previously it probably would have had something to do with priority: earliest listed item goes to the first person to bid at least that amount. Another way to do it would be to try to match highest to highest: If there are items listed for 5,000, 4,000, and 3,000, a bid of 4,500 takes the 4,000, then a bid of 4,000 takes the 3,000, and, finally, someone will have to bid over 5,000 to take the 5,000. Either method would seem to make it less likely that the situations of lots for sale and lots bidding would happen, but I wonder if the current method doesn’t have advantages. My first thought would be that the load on the server/database is less severe by doing a simple sort of two tables and matching the top rows; anything even slightly more complex might add up over the presumably millions of transactions hitting the system each day. Another benefit is adding in a risk vs. reward scenario: a low-listed item might sell very quickly for the going rate, but just as easily might sell for a pittance if the demand suddenly drops.

To sum up: City of Heroes has a very unusual player to player market system, and I’d be interested in seeing more games attempt a consignment system rather than an auction system, possibly with different rules for matching players to see what variations do to the overall model.

MMO Theory: Incentivizing Crafting


(Originally Posted June 2009)

Right now on City of Heroes, generic crafted level 50 accuracy-improving enhancements are selling for 500k each in the game’s currency at the player auction house. This isn’t too bad, considering that level 40 versions of the same enhancement are selling for 400k.

Except that level 40 enhancements cost 200-300k to craft (depending on how much you paid for the components), while level 50 enhancements cost 1 million to craft before you even add the components. Anyone off the street can make a tidy profit from crafting and selling the level 40s, while anyone trying to sell the higher level versions is going to bleed currency like a sieve.

The problem comes from how crafting is regularly incentivized in MMOs.

Back when I was playing World of Warcraft (and I have no reason to believe it’s any different now), the advice to players wanting to make money as quickly as possible was to take two gathering professions (instead of a gathering profession and a craft). The reason for this was very straightforward: for almost every crafted item at every level, the components of the item would sell for more on the auction house than the actual item. Nearly every crafted item was sold at a loss.

The reason for this is very simple: the crafters weren’t making items for sale, they were making them for skillups. WoW’s crafting follows a very hierarchical progression; a crafter can’t make items appropriate for a higher level until increasing his or her skill by making many items of lower level. The vast majority of crafted items posted to the auction house were, therefore, priced mostly incidentally: their value to the crafter was that making them had increased his or her skill by a point, not what they’d sell for.

Remember those people with two gathering professions? Once they reached the level cap and had plenty of money, they might decide that they wanted to dabble in the crafting system. They’d drop one of their gathering skills, start a new craft at 0, and then proceed to buy lots of components (driving up the price), craft recipes to skill up as quickly as possible, and sell the crafted items as an afterthought (glutting the market with cheap items and driving down the price). The cycle repeated itself.

Why level a crafting skill at all? Each contained a few decent crafted items that were either usable only by the crafter, or were hard to get. Thus, amidst the dross of crafted goods, there were a few rare items that were worth more than their components, and worth all the trouble of skilling up crafting to get.

City of Heroes’ situation is a bit more complicated, but similar. Instead of offering skillups, CoH offers crafting badges. Once you’ve made a certain number of enhancements of a given type and level, further enhancements of that kind don’t require a new recipe and the actual crafting is done at half-cost. For the Level 40 Accuracy enhancement mentioned above, that’s a saving of around 150k (3/8 of the sale price becomes pure profit). In order to sell enhancements at this profit, though, a crafter has to make a lot of enhancements at a minimal profit, or even a loss, first.

And even when the crafter gets the badge that makes crafting the enhancement profitable, he or she is still competing with random-drop non-generic recipes that cost the same to make no matter who crafts them. The level 50 enhancements are so much less profitable largely because level 50 is the max level, and characters there would rather buy specific set-based enhancements than the generics.

But, ultimately, the low price of crafted enhancements is the same problem as WoW. Accuracy enhancements are one of a few exceptions that are in high demand; most crafted enhancements sell for far less than the cost of their components, much less the cost of crafting. I spent the last week as part of the problem: I was trying to get all the crafting badges* on one of my characters, so the actual enhancements were just the trash left over from getting the badge. Anyone on when I was could have gotten enhancements that cost up to 200,000 to craft for as little as 100. And I still had to delete a lot of the lowest level ones because the market was so glutted that it wasn’t worth tying up an auction slot for days on something I was basically giving away.

This long and rambling explanation brings me to my point: MMOs will never actually be virtual worlds for the purposes of crafters as long as learning to craft is part of the grind.

In the real world, crafted goods very rarely cost less than the materials used to make them. There are simply too few people making any given thing relative to the population as a whole, and even fewer of them that can afford to take a loss on their work. But in an MMO, fully everyone in the population can be a crafter; and as long as a number goes up on the character sheet and something useful eventually comes out, very few are going to actually care that crafting is an expense rather than an income. The game is to make the crafting skill number go up, not to make money with the crafting.

But I think there is a significant minority of players that go into these games with the fantasy of becoming a crafter in the traditional sense: buying components wisely, putting hard work and love into a creation, and selling it for a profit. For these people, requiring them to make 10 of Widget A and 10 of Widget B before they can make Widget C (the one they really want to make) is not a feature. Requiring them to go gather their own materials if they want to actually make a profit on their crafting (and still knowing they’d have made more of a profit just selling the components) is not a feature.

And you’d likely have enough of these people that your economy would be perfectly healthy if they were only competing against one another on price, instead of against the unrealistic prices set by high-level characters grinding up a tradeskill. The rest of your players probably wouldn’t care, as long as they still had things to do that interested them.

For this to happen, it requires designing part of the game around players that have little interest (or even capability) in beating up walking sacks of EXP and leaving one more number that goes ding off of the majority’s character sheets. It’s a lot harder than just letting everyone craft, forcing them to grind it up, and thereby adding an additional time sink to the game.

But it might result in a much healthier economy and much happier crafters.

*And I did, too!