How Does One Measure An Economy?

Posted by Markco at 8:00 AM

This post has been submitted by Xsinthis. Please leave a comment letting us know if you'd like to see Xsinthis write again here at JMTC.

This has been a question that has been plaguing me for a while now, how does one put a server’s worth into tangible terms. Is it even possible? Is it within a player’s ability to calculate? Could even the fine folk at Blizzard solve this? In the real world we have many ways to quantify a countries economic status. GDP, stock markets, average income, quality of life etc, no such thing in World of Warcraft. That information either has no warcraft parallel (no true stock market warcraft) or isn’t readily attainable (GDP, average income).

So How Do We Go About It Then?

While working on my Economy Quantifier Program these past weeks, several ideas have been going through my head while working on the data The Undermine Journal so generously provides.  The sum of the purchase price of an auction house is one of the first that came to my mind, but after a few days of processing the data I realized how prone it is to being manipulated. Some days I’d see a spike of over 20% in the AH worth for a 48 hour period. Obviously someone put up some rare item(s) for large sums of cash, but that event screwed up the stats for the two days because they obviously didn’t sell since the spike lasted exactly 48 hours, and since they didn’t sell it had no real effect on or from the economy, probably just some idiot overpricing their items. Good stat to track either way (just in case) so I started recording the averages, left it in, and moved on.

The number of listing on the AH is another stat I started recording, but even from the get go I realized it was in no way a proper measurement for an economy, since those listings could be listed for coppers of thousands of gold; still an interesting stat to track though, so I left it in.

After some work I managed to calculate how many listings were posted per hour. Not how many were sold, and not the change in the listing numbers, but how many auctions were created per hour. This turned out to be an excellent stat – but not for what I intended.  This number provided an excellent measurement on how busy the server is, and so far is one of the most important stats I’ve been able to pin down, but again it has no meaning on the actual economy of the server.

One last figure I’ve thought of is the average price per item. Very good stat to measure the economy, but two major problem with this, along with every other method I’ve gone over so far: First, this doesn’t distinguish between items that sell and items that don’t, it’s  the average price items are posted at, not sold, so it is polluted. Now while it should be a very minor pollution, anything other than 100% is pointless for this exercise. Second, all these are based from the auction house solely, and while it is a major source of economic interactions, it completely misses out on face to face interactions, and even C.o.D. interactions. I can’t even begin to fathom how much gold changes hand this way.

In a Perfect World, How Would We Measure It Then?

If I could request any data for Blizzard to make available, within reason, it would be what auctions actually sold.  This would lead to an explosion of analytical information in the World of Warcraft blogosphere, we would finally be able to measure not only how many auctions get sold, but what the average cost is, and even how much gold changes hand per day! We could even calculate how much gold bleeds out of the economy from AH cuts. This would still leave out face to face interactions, but it would be a huge step in the right direction and any other information that would include face-to-face would be too hard or simply impossible for Blizzard to provide.

It would also provide much more useful and relevant information to sites like The Undermine Journal and we’d be able to calculate the average sale cost of items, not just listing price, and more! Yes I think until Blizzard releases this information we can get closer and closer to fitting a server’s economy with a number that defines it.

What is the Point of Collecting These Stats?

Many people might be questioning the point of collecting all this data and that’s good, it is always good to question! I have many reasons why I want this data, and I’m sure there are some people out there that could come up with more.  Besides general curiosity and labelling purposes (my server’s got a small economy, my server has a big economy etc) the main use would be theorycrafting, trend tracking, trend predicting etc. We’d be able to predict short term trends (when does most business happen during the day,  long term trends (when does the most activity occur during the week)  and trends across seasons and patches (what effect do different holidays have? What patches in the past have had the most effect on the economy?). It could be a very powerful tool to predict future market trends once we can study past effect. Of course we can only start now tracking these trends, but the longer we do, the better we’ll be able to predict future ones, and imagine what we could do with our gold then!

Do you want to write for JMTC? Send in a post to Markco and if it makes it past the initial screening process then I will publish your submission on the blog. JMTC pays 50% commission to writers on the day that their post goes live.