Slashdot Mirror
Using Math To Tune a Video Game's Economy
An anonymous reader writes: When the shipping deadline was approaching for The Witcher 3, designer Matthew Steinke knew there was a big part of the game still missing: its economy. A game's economy is one of the things that can make or break immersion — you want collection and rewards to feel progressive and meaningful. Making items too expensive gives the game a grindy feel, while making them too cheap makes progression trivial. At the Game Developers Conference underway in Germany, Steinke explained his solution.
"Steinke created a formula that calculated attributes like how much damage, defense, or healing that each item provided, and he placed them into an overall combat rating could be used to rank other items in the system. ... Steinke set about blending the sub-categories into nine generalized categories, allowing him to determine the final weighting for damage and the range of prices for each item. To test if it all worked, he used polynomial least squares (a form of mathematical statistics) to chart each category's price progression. The resultant curve (pictured below) showed the rate at which spending was increasing as the quality of each item approached the category's ceiling value."
"Steinke created a formula that calculated attributes like how much damage, defense, or healing that each item provided, and he placed them into an overall combat rating could be used to rank other items in the system. ... Steinke set about blending the sub-categories into nine generalized categories, allowing him to determine the final weighting for damage and the range of prices for each item. To test if it all worked, he used polynomial least squares (a form of mathematical statistics) to chart each category's price progression. The resultant curve (pictured below) showed the rate at which spending was increasing as the quality of each item approached the category's ceiling value."
As you say, you want to feel like your economic choices have consequence, and not just for your pocketbook. If I buy all the fish in town, then not only should the price of fish go up, but the fishing industry should spend its money in predictable locations and boost other industries.
As my subject line alludes, the real place this crops up again and again is in space games. If I'm buying and selling the complete available product of a planet in some particular industry, that should have significant effects. Or, if I blow up a bunch of cargo ships carrying spaceship parts, then that should have significant and immediately noticeable effects. Instead, none of this is true at all in basically any game but EVE.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
The economy is one of the very few real weak points of Witcher 3. The article doesn't explain in detail what he did, but whatever it was: Don't do that, unless you have an otherwise master-class game to outweigh your mistake.
Many big games hire actual economists to design and maintain a workable economy. They tweak things much like the Fed tweaks interest rates, in an attempt to keep the game interesting, and to deal with the occasional bugs that cause rampant inflation.
John
Yep, it annoys me a bit when at the beginning of a game, it costs a gold coin to stay a night at the inn at the beginning of the game, but it costs 100 gold coins at the end of the game in the "end game" areas. Granted, these are not games that are attempting any sort of "serious" world-building like Witcher 3. It still seems lazy, though - they should use other means for removing money from the player.
I've also thought it somewhat lazy for developers to have a single coin type (typically "gold coins"). I've always thought that designers who resorted to that sort of currency never really understood how incredibly valuable a single gold coin actually is. Common goods are typically highly over-valued, and weapons and armor actually tend to be undervalued. It's ridiculous for a fish to cost 5 gold coins while a sword, even a basic starter sword, might only cost 100. Granted, I think most gamers don't really care about such things, but just once I'd like to see a videogame at least attempt to make a realistic stab at this.
MMOs (particularly "theme park" types) have a particular problem with their economies, in that goods are essentially created from nothing in limitless supplies, and the final products from crafting last an infinitely long time. As such, they need to figure out how to create viable viable "gold sinks", or ways of removing money from the economy in ways that doesn't irritate the players too much. I've always thought that, no matter the solution they came up with, it always ended up feeling rather artificial (or simply inadequate), and it often ended in a massively inflationary economy.
I'm still working my way through the first two Witcher games. As soon as I'm ready to play the third, I'll have to take some note of the general economy and see if it ended up as satisfying to play as one hopes.
Irony: Agile development has too much intertia to be abandoned now.
As a professional videogame programmer, I can assure you that it's fairly newsworthy when a game designer uses math to solve problems like these.
Irony: Agile development has too much intertia to be abandoned now.
I'm exaggerating a bit, as I was going for a humorous remark. So, killing the joke dead and answering your question seriously, yes, I've heard of Eve Online, and know quite a bit about it, but haven't played it. In terms of economies and designer math skills, MMOs are a slight exception. They take their in-game economies very seriously, and some are known to hire actual economists to help guide their virtual trade and monetary policies.
Designers don't necessarily avoid math, but they often use it in ways that drives programmers insane. For instance, they'll use formulas in spreadsheets to calculate tables of figures for loot, armor, damage, etc, and then painstakingly enter all those figures by hand. Why? This allows them the freedom to individually tweak components as desired. Programmers look at this and shake their heads, because we're hardwired to avoid that sort of redundancy whenever possible. For whatever reasons, some designers seem to revel in it.
Many designers tend to be control freaks. They're deathly afraid of automatic, feedback driven, or chaotic systems (all of which could apply to an economy), and often for good reason, as some outlier condition could tend to cause autonomous systems to spiral out of control. This is why most MMO in-world "economies" are tightly controlled affairs, with lots of limiters to prevent things from going too far in any particular direction too quickly. Programmers are always suggesting ways to create automatic and chaotic environments driven by algorithms, but designers want carefully scripted and planned systems, because they want to control and tweak it.
From the description, it appears this designer only turned to math as a last resort, because there was simply no time to do the sort of hand-tweaking that designers typically do. They were impressed enough with the results (gosh, math actually works!) that they wrote about how they did it. And the programmers on the team are probably thinking "I told you so. In fact, I told you so years ago, and you only listened to me when your back was up against a wall."
Irony: Agile development has too much intertia to be abandoned now.
Most modern MMOs don't have a real economy. EVE Online is a rare exception (and very much a niche proposition). Of the "normal" MMOs, then depending on who you talk to, Final Fantasy XI (released 2002 in Japan, 2003 in the West) or Star Wars Galaxies (2004) was the last to allow any significant degree of player freedom to shape the economy. Everything since World of Warcraft has locked in-game financial transactions down so hard that actual trading between players is knocked right out to the margins.
What do I mean? Basically, that the items of highest value in most MMOs cannot be obtained through player-production and in-game trading. In a subscription MMO (eg. WoW, Final Fantasy XIV), the most powerful gear will normally be obtained by players through skill-based team challenges (high-end raiding or PvP). Once obtained by a player, it is almost always locked so as to be untradeable to other players. A second tier of good-but-not-best gear is usually available via a pseudo-currency (tokens, honour, whatever) that is rewarded in fixed proportions from performing repetitive activities (daily quests, dungeons etc), often with a cap on how much can be earned in a week, and which can't be traded to other players. Consumable items are still crafted and traded, but the trend in MMOs has been to downplay these over the last few years.
This has largely come about due to the prevalence of real-money trading; people using real-world money to buy in-game currency. When the companies running MMOs realised that trying to crack down on this trade was like trying to wrestle with fog, some of them shifted their strategy to downplaying the importance of the in-game economy.
The others, of course, went into the "freemium" pay-to-win model, whereby he who spends the most real-world money has the best stuff.
What's interesting in the subscription MMOs is how the free market has acted to route-around the ever-tighter economic controls. You might not be able to buy the best gear, but in both WoW and Final Fantasy XIV, you can hire teams of players as "mercenaries" who will run you through the high-end content and guarantee you the items that drop. Blizzard don't like this, Square-Enix turn a blind eye. Various other types of grey-market activity have also sprung up around the margins to allow players to turn either in-game or real-world wealth into an advantage.
Magic Online, the digital version of the famous trading card game, is currently undergoing a kind of "economic recession". Basically, trading for cards is facilitated by event tickets (roughly $1 in value) and booster packs (containing 15 cards, roughly $4 in nominal value) that act as a digital currency. Events are entered using event tickets as payment and they pay out as booster packs to the most winning players. This is done to avoid gambling laws.
What has happened is that number of players entering events has gone down while the amount of booster packs floating around has increased, so that the going price for booster packs has fallen to around 2 tickets (so $2 equivalent). This has made entering events unattractive for all but the top players, since the expected value of the prizes to win are now half of what they should be, while the entry fees remain the same. This further drove the number of players down, with many people selling their collections and leaving Magic Online.
Several months ago Wizards of the Coast set up an "economy strike force" that supposedly consisted of several people with "advanced math degrees" to solve this conundrum of a depressed economy. They finally announced their "solution" some weeks ago. It was to double the entry fees and basically cut the prizes to 75% of the old one.
The economy predictable tanked even harder, leading to more players selling out.