Saturday, September 19, 2009

Why is the bid-ask spread in Jita so wide?

I'm picky about game economics and I've yet to see a game come even close to right, but am happy to see that Eve has done a pretty good job. Most of the stuff in the game is built from raw materials that players extract from the environment. That's pretty cool. Manufacturing is not a second hobby that every character picks up (and therefore there aren't 500 million experts at manufacturing X), it's requires the player to spend time training manufacturing skills rather than fighting skills and it requires their time with no benefit besides whatever profit they can make. While there are definitely inefficient ways to make money, it's less clear if there is a best way to make prices bring the profit from different professions into equilibrium over time. And the market rocks. Well done, CCP. You can actually place buy and sell orders. There are relatively few items that the NPCs buy and sell, most of the market liquidity is produced by players. And best of all, almost everything in the game can be traded (even game time!) so nearly every action can be motivated by profit.

More coolness: People and things don't instantly teleport across the game universe. When you sell an item to someone, they have to pick it up where they bought it (or courier it). Space matters. And, mirroring real life, hubs have formed naturally over time, meeting a desire for concentrated market liquidity. Why do I go to Jita to buy some shiny new toys or sell my loot? Because I can sell or buy it quickly with relatively little markup. Why would I want take the opposite role and make a market in Jita on some item? Because there are hundreds of players buying and selling there meaning I can turn over stock pretty quickly.

All good so far, but one thing I expected that hasn't happened: the bid-ask spread in Jita has not narrowed so much that it's not worthwhile for the normal player to trade. I would expect this b/c I think the spread would be a function of:
  • The relevant transaction costs (could have an interesting side discussion about 'relevant')
  • The risk of adverse price movement (how likely is it the prices fall and you're stuck holding stock)
  • The volume of trade (in ISK, not units)

Basically, if the spread after transaction costs is profitable, the risk isn't enormous and I can do a good enough volume to make it worth my while I am going to trade trade. I don't think I'm terribly unrepresentative. And if there are lots of people competing to get the buy and sell orders, I would expect the spread to shrink since they're forced to outbid each other. It should keep shrinking until the spread is so narrow that it's not very profitable for most players to to trade the market. In an extreme case, it could shrink so far that the transaction costs would be larger than the spread for all but the most skilled toons. The extreme case does happen. Tritanium right now has a buy of 2.94 and sell of 2.97. I calculated my gross profit (sell minus buy) then subtracted out transaction costs and it was negative. So I can't profitably trade this. But that is an extreme case. It is more likely that the spread on an item would be wide enough to make money, but not enough money to justify the risk or be worth my time.

So, looking at recent Jita prices, I could place the highest buy order for a Small Salvage Tackle I at 900k and once I buy one I could place the lowest sell order for 1500K. Taking into account my transaction costs of around 30k (0.85% broker's fee x 2 + 0.7% sales tax) for the round trip trade, I can still make 570k per rig. That's a pretty good % return off of one trade. If I could move enough of these I could double my money. But I wouldn't trade this because I wouldn't be able to move enough of them to make it worth my while. The volume on this market is too low, only about 200 units a day. Meaning if I managed to capture every single trade over 24 hours (which would require constant attention for 24 hours and a lot of luck), I would make as much as 100 million. That's a little less than 5 million per hour, which is far less then I can make running missions (I like this break-even measure for considering if I should spend my time trading). And it would be incredibly tedious. For players poorer than me (of which there will be some, but I doubt a lot since I'm new to the game) this may be a good trade, but for most players it isn't worth their time either, so I'd expect them to ignore it. This means fewer competitors so I'd expect wider spreads.

So let's look at another example. I could trade improved implants. They frequently have a spread of 15M in Jita. Even with a volume only 50 per day, assuming you could snag every trade, that's 750 million per day, around 30 million an hour. That's pretty good money. I have to have at least 150 million to start with, but that isn't a high bar for most players. And I'm not speculating here, I've put my money where my mouth and I have made pretty good money trading improved implants . So if I would play, I would expect others to play. If others played, based on my earlier logic, players would be outbidding each other so I would expect spreads to shrink. But this is the weird part...they haven't. Spreads are volatile, but I haven't seen any trend toward them converging. And it's not from a lack of competition. If I place a buy order it will be outbid within 5 minutes, almost anytime day or night. Often I'll see multiple outbids within a few minutes (I place an order, when I check back in 5 minutes 5 buy orders are higher than mine). When that happens it implies 5 people are watching the market (or fewer people moving multiple orders). Side note: For a while I was curious if there were macro trading programs or something doing this, rather than actual people, but I can see no evidence of them searching around on google (please tell me if I'm missing this...would be curious how they work (though I wouldn't personally use them since it's a bannable offense (and yes, I tend to nest parenthesis in written English, it just seems so obviously reasonable))). So it seems like at any given point a few players are spending their time watching this market, hoping to capture that sweet spread. But as the market gets busier with competition, I don't see the spread collapse, I just have a harder time keeping my bid at the top of the pile. And, interestingly, THAT is what discourages me from playing this market. If there are 5 people outbidding each other then any given bid is only going to get 1/5th of the sales (assuming equal distribution of price changes, yada yada). So rather than the spread shrinking until it's unprofitable to trade, the spread stays reasonably constant but instead it takes more of my time to make that 15M trade. That is definitely not how I've seen real life work. When I buy something off Amazon I may go for the lowest price, but I don't think there are people behind the scenes checking in on their competitor's prices every 5 minutes and lowering theirs by $.01 (please please disprove me here b/c that would be awesome). This seems like some sort've artificial weirdness in the game, but I'm not sure what's causing it.

Who cares? Why have a I written a few pages to describe a phenomenon that is far removed from the pewpew and will likely interest the standard gaming populace about as much as a play-by-play of a 4 hour high-sec solo mining op? Because it hits all of our wallets and causes people to profit through mindless boredom (therefore encouraging this mindless boredom). If I can make a 15M spread on a high volume low risk item with total transaction costs of around 3M, you're essentially paying me an extra 12M to mindlessly outbid my competition by 0.01 ISK. That doesn't provide a service for you and it's not fun for me. I don't like that. So I'd like to understand this phenomenon and if possible fix what's causing it. But why does it happen? What is the game mechanic at the root of it? Or is it even a game mechanic that's causing it? Maybe I'm wrong and this happen IRL too (in which case RL examples could help me understand what's different between the Eve market and Ebay or Amazon). Regardless, I want to get to the bottom of this. I have some ideas of my own about causes but would like to start a discussion rather than continuing my monologue.

Oh, and feedback on my write-up would be appreciated too. Does it flow well? Do I make sense? Are my ideas logically connected and ordered in a way that makes it clear what I'm trying to communicate? Is this an interesting topic?


  1. I'd have to say it's because there isn't a significant "menu cost" (see in EVE. You can change your prices every 5 minutes for the entire day and it'll only cost you 100 isk per change, which is much less than you make by holding the highest sale price there for any appreciable amount of time. I see two solutions possible:
    A) Add a significant cost to the lowering of prices so it costs enough to make people think (do I want to spend that much to hold the market for a few minutes, significantly undercut the competition when initially listing, or just wait for mine to sell tomorrow? Once you can find that point then the EVE market will look like a real market and as a side-effect you'll see a bit more ISK sinking into EVE's black hole. I think it should be around 1% of the total order cost to change an order price.
    B) Add a price-change time limit so that people can only update a price once every 12 or 24 hours or each day. So if you set a price, you can change it immediately but after the first time, you have to wait 12 hrs each time you want to do it again.

  2. Hi,
    Good article and got nice insight.


    Forex bid and ask