The following blog post, unless otherwise noted, was written by a member of Gamasutra’s community.
The thoughts and opinions expressed are those of the writer and not Gamasutra or its parent company.

 

[The author makes strong assertions about links to narcissism and interactive media that are speculative and have not been proven scientifically. The following is an opinion paper and not to be construed as a scientific paper. The opinions here are solely those of the author and not those of Gamasutra or their agents, or of Wargaming as he is not an employee of that company currently.]

The Origins and Problems with Pay to Win

While Nexon may not have invented F2P, they are mostly credited with it due to the success of their 2001 Maplestory game which launched very successfully with it. At first F2P was thought as a way to lower the barrier to entry in Asia where (back then) consumer budgets were not as large for gaming as they were in the West or Japan. But once it became clear how well advantage could be sold, Nexon doubled down on what we now call Pay to Win monetization. This established the state of the art in F2P since that time.

But there are very serious limitations to this model. Let's say I have five consumers, who I will call A, B, C, D, and E. If I sell advantage to player C, and that player then beats players A, B, D, and E, then C feels a bit better and the other players feel a lot worse. Soon even player C feels bored because the challenge level has dropped and the outcome of competition becomes assured. This adaptation takes longer in children. Now if player D decides to buy advantage also, then players A, B, C, and E are suppressed.

As this process continues the net affect on the consumer product perception goes down with every transaction. Players become unhappy and perceive the game as unfair and not a meritocracy reflective of their skills or efforts. The game itself may be a lot of fun but the business model being used to extract money from the participants ruins the game rapidly over time. “Toxicity” builds up rapidly also. Because of this negative effect on the gameplay and product perception, I consider P2W a Darksided model.

The Narcissism Epidemic

A much more subtle but no less important dynamic is that when we charge a player to make them better than the other players, we are feeding narcissistic urges. The USA is currently experiencing an epidemic of narcissism. One cannot be diagnosed with Narcissistic Personality Disorder (NPD) until they are 18, so you would expect older people to have had more opportunities to be diagnosed with it. So why is it that only 3% of those over 65 have ever had NPD but for those in their 20s they are already being diagnosed at a 10% rate? How many of them will have received an NPD diagnosis by the time they are 40?

We live in a society where “being liked” is commercialized on a massive scale. Even a former leader at Facebook is concerned that FB may have done real harm to our society. Now being liked is the primary metric we use to see how valuable we are in society. Taken to the extreme, which is becoming easier every day, we end up with NPD. Many without an official diagnosis are likely well on their way to one. I'm told that one of the worst things you can do to a narcissist is let them know you don't like them. This can result in an emotional response. So once a person is on that treadmill, they will go to extreme lengths to stay on it. It becomes a “like addiction”.

P2W ties into this nicely, and I do not think it is coincidental that the conversion rates (~5%) in P2W monetized games mirror the NPD rates in the target demographics for these games. If you can keep a player on this treadmill forever, they will have a very difficult time getting off. Making an infinitely scaling narcissism engine is actually a very complex technological feat. If someone could build one, they could extract not only thousands, but even millions of dollars per user once a person with NPD is locked in. A normal “healthy” person can get off, but NPD is a disorder for a reason. Well it turns out that the infinitely scaling narcissism engine isn't theoretical, it's been built.

Okay, so narcissists are really getting screwed, who cares they have it coming right? Well wait a second. Most of these were healthy babies some time before. Technology is absolutely creating NPD even if we want to argue about the exact mechanisms. Left unchecked, technology will get increasingly better at this. If we know NPD individuals are vulnerable to exploitation and unlikely to be protected due to their unpopularity, then we can grow our next slave class. The technology used in first generation infinitely scaling narcissism engines is still primitive. A significantly more advanced version could keep even healthy people on that treadmill for a while and there is the risk of creating a system-wide positive feedback loop for NPD that could radically alter the species.

Okay so here is where I lose half the game developers who rush through Door Number 1 to start developing their own infinitely scaling narcissism engine based games, and this is why I've been extremely reluctant to talk about this publicly. But keeping silent is perhaps a worse option because without knowledge, consumers cannot defend themselves and regulators cannot begin the long process of regulating these technologies. Make no mistake, if nothing is done, the NPD epidemic will pose a greater threat to our society than loot boxes.

For the other half of you, I am going to show you Door Number 2. I've been reluctant to do this because even Door Number 2 can be weaponized. I'm trusting that those with that propensity are going to take the low hanging fruit that Door Number 1 represents. For the rest of you, and for the industry at large, I want us to have a future worth living.

 

Lightsided F2P Models

 

Some of these models have already been proven, and some are still theoretical. It is my belief that as they become more theoretical (further along in the article) they become more powerful. Getting even the second model implemented required an incredible leap of faith for Wargaming in 2014. I'm not sure I ever attempted to explain why it would work, they just trusted me that it would work. But the early bird gets the worm and Wargaming's World of Warships seems to be a huge success despite being a very complex design. Even knowing how it was built, it would take years to reproduce it so I feel Wargaming will be justly rewarded for their being first to take that risk.

Here is the progression/timeline, and I will explain each in the following sections:

Pay for Progression: Status: Implemented first in World of Tanks (2011). Optimized in World of Tanks Blitz (2014). Conversion rates and LTV several times higher than P2W.

Pay for 1st Class: Status: Implemented first in World of Warships (2015).It appears that this is yet another boost to conversion rates and LTV on top of Pay for Progression, and allows for asymmetrical eSports for the first time.

Happy Pinata Economy: Status: Implemented in unannounced title under development. No Data.

No Narcissism Design: Status: Implemented in unannounced title under development. No Data.

 

Pay for Progression

 

When players do something cool and unanticipated, they get a hit of dopamine. This is the same chemical they get from smoking cigarettes, though in that case it is the nicotine (usually freebased with urea) that forces the dopamine release. In games we have to “trick” the brain into releasing it naturally. Just as adding urea increased the potency of tobacco 100 fold, by understanding how dopamine is naturally released we can increase dopamine delivery potency in our games. The consumer doesn't have to know what's going on, they just know that playing game A makes them feel better than playing game B. The longer they play game A, the more they are willing to spend. If game A becomes their game of choice then their willingness to spend can rise significantly over time.

In Pay for Progression all rewards in the game have to be earned, but we give them a bit of a boost on those earnings if they are subscribed. Subscription can be a traditional long term (monthly) subscription or we can use shorter duration “micro-subscriptions”. Thus the normal pattern of delivery of dopamine is not interfered with and may even be enhanced. For this to work the player must still have a very significant chance of failure because once winning becomes assured, the dopamine hit is substantially reduced. This is a big part of the reason why P2W causes reduced spending over time and significantly reduced LTV.

Developers sometimes see how effective this is and apply a “more is better” approach by adding the technique to their P2W games. This doesn't work. Going into a lab with a “more is better” approach will just get you blown up as you saw with EA's SW BF2 where they tried to stack retail and P2W models on top of each other.

I first developed the model in 2009 as part of my 35 page Sustainable Virtual Economies and Business Models paper which I wrote in part as a proof of concept to ensure I would be able to develop my virtual economic theories in a PhD program. When I observed the model in use in World of Tanks two years later, I of course became very excited. I began showing Wargaming some of my research in private, and by 2013 I was hired to begin deploying my models inside their games. When I met with Sergey Burkatovsky at the end of 2013, the brilliant lead designer of World of Tanks, I told him I could optimize his design. He had every right to say “how dare you!”, but to his credit he thought about it and gave me the green light, giving me total design authority over World of Tanks Blitz. Just two years earlier I had said something similar to Dr. Edward Castronova and his fiery response telling me I needed psychological counseling for not respecting his authority ended my PhD run.

Over 10 weeks I optimized the original WoT design to better make the rewards match the effort/risks involved to more closely match the natural way dopamine is delivered. I also changed the way tank add-on equipment was being handled to make these features extremely accessible at the early stages (previously this was end-game only content) in order to build habituation so that players would build comfort with that feature throughout their lifespan. I did my best to remove pay to win elements everywhere, but was not allowed to remove premium ammo, a P2W element, so it remained.

The results caused Blitz to outperform the already high performing WoT base game, though both games are limited in their market performance by being gender-male and not gender-neutral in demographic target. By high performing, I mean conversion rates ~30% with second year spending exceeding first year spending. Under a P2W model, players rarely go past 6 months and convert at 2 to 5% (again, mirroring the demographic NPD rate).

As a note of caution, I am told that after I left the company there was an attempt to “optimize” the design by restoring P2W elements including increasing the value of premium ammo. The result was catastrophic, necessitating a rollback. Don't mix your meat. If you are going to go Dark Side, go all out (as in Door #1). If you are going to go Light Side, go all out (as described in this paper). Just randomly adding mechanics that you don't understand will cause a disaster.

 

Pay for 1st Class

This is a model I invented in 2011 while trying to answer the question of how do we successfully apply differential pricing in real space? Repackaging cereal into fancier boxes wasn't going to cut it with gamers who are too sophisticated for that. But airlines have nailed this for decades, where cheap customers get crammed into the back of the plane like sardines while the 1st class customers cruise in comfort. It's not considered pay to win by coach fliers because all passengers have the same goal: to arrive at their destination alive and preferably on time. The coach fliers realize their seat is being subsidized by the 1st class fliers and that appreciation makes it all work.

When I landed in the Saint Petersburg Lesta studio in 2014 to begin building the economy and business model for World of Warships, the team was planning on having all ships be equal, sort of like in World of Tanks. Thus they were going to balance destroyers to be just as powerful as battleships. I was like “No, no, no, you guys are sitting on a gold mine design and you don't know it!”. I explained that we needed to make battleships and carriers way more powerful than destroyers like they should be. They responded by saying “But then everyone will play battleships and carriers!”

 

Do you see where this is going?

I began to build a model where the economic rewards for battles with capital ships were too small to break even over time, and with destroyers the rewards would exceed the cost of repairs after each battle (on average). This forced players to play small ships more to earn enough coin to play the bigger ships. Still F2P and it made the ship distribution heavier on small ships than on large ships, just like it would be realistically. Also, it raised the prestige value of cruising into battle with a larger ship. Balance was maintained between both sides with the matchmaker so that all battles were fair. Players could buy coin to allow them to play the big ships more if they were sufficiently motivated.

At first the team was like “that's a crazy/impossible/stupid idea”, but after they saw it develop they decided it might work. In the end, they fell for it and World of Warships was born with the ability to play big ships that actually played realistically like big ships. The game is amazingly complex with the potential to get much more complex over time, but the Lightsided business model makes return on that further development worth it as players play crazy numbers of hours and spend freely over that time.

The potential of this model for eSports is important to note because this allows “fair” play even when individual members may differ significantly in power level. This allows future products to allow “coolness” levels that just are not possible with existing models.

This theme is common with the next two models that are intended to allow us to build much cooler games, by removing some of the most frustrating design constraints that exist now when we try to make games “fair” (and usually bland/predictable...).

 

The Happy Pinata

 

I came up with this model in 2011 in order to allow me to design the game I'm working on now, which is a massively multiplayer multi-platform (mobile first) spaceship eSports game that permits differences in scale even more extreme than in World of Warships. Progress is a bit slow since we are trying to keep development costs to a minimum and I'm doing 100% of the design work which is new for me. I usually just handle the meta game, not the gameplay.

The Happy Pinata is essentially a trickle down economic model. Conventional trickle down economic models, where you take the money from poor people and hand it to rich people are actually “trickle up” models. You can imagine how unpopular such a model would be in a game. But a real trickle down model that actually trickles down like it is supposed to can be great for everyone.

In this model the rich/motivated players can bring in a more powerful avatar (a starship in the case of my current game) and players are rewarded more for shooting that avatar. This makes the rich gal happy because she's laying waste to everyone in the coolest ride in the galaxy, or at least that battle. The paupers are happy because they are getting rich by taking it to the (wo)Man.

There are a lot of ways to implement such a model, but the key elements are that everyone in the economy must always benefit and it has to feel at all times like a meritocracy. The outcome of any contest should also be in doubt at the start so that dopamine release is not undermined. I manage this by using a rock paper scissors system where even the biggest ships are weaker to some other ships so that a balanced team approach is most effective.

 

No Narcissism Design

 

This is probably the most radical and unintuitive model in the set. The goal here really is to suppress narcissism/toxicity, and maximize oxytocin release. To do this the focus is entirely on the team, not the individual. Rewards are delivered to the team, not the player. I do allow the player to advanced based on their play, but all currencies other than XP go to the team. Those currencies are then spent by the team leader on things that benefit everyone on the team. In the case of my current game, I call those things “technologies”.

The better a player does, the bigger their reward in the form of praise from their fellow teammates. Even narcissists benefit under this system, but not disproportionately. The inspiration for this model was a combination of my studies of oxytocin dynamics and my experience coaching the winning USA women's Olympic track team in 1988. For more detail on why it works, or should work, read up on the linked article where I explain why the records set by our athletes that year are considered unbeatable.

This model is probably more complex than it sounds, and I'm going to keep the details mysterious for now until I get a chance to see how it works when the game goes live. The main thing is that in single player rewards, one person feels good and everyone else might feel bad. Here by having everyone get rewarded for a person's actions, a tremendous amount of positive peer pressure is generated and it could cause a positive feedback loop that will be very difficult to voluntarily disengage from. People become more much likely to give to other people under the effects of oxytocin and this could cause them to spend a lot more than they usually would, but on other people.

I am stacking all four of these models in my current design, trusting that I know what I'm doing enough that I don't create a mushroom cloud. I would recommend starting with one or two of these at first and seeing how they work before trying to go hog wild. Some will work better than others on your design, unless the game was actually designed from day 1 to use all four of these like mine was.

 

Cautions

 

If you make a game that consumers prefer to the real word, which is certainly possible, you can run into problems. While this situation is mostly theoretical for those that don't normally have a problem with gaming addiction, I think it's likely to be a big problem when enough Lightside effects are stacked. This is especially true when you start improving the oxytocin dynamics in the game because it seems that oxytocin is a much more powerful reward chemical than dopamine is.

My concern is that players could continue to play until they collapse, which is already possible under certain circumstances. As we move to models that are several times more engaging than current models, it's going to increasingly be an issue. As I explain in detail in my I'm Dying to Play article, players can get over their head and be dying without recognizing the symptoms in well under 24 hours. Because of the mechanisms involved, if you wait till they collapse to rush them to a hospital, the hospital's options are limited.

Long term game fatigue that falls short of causing a collapse can still destroy work, relationships, or cause disease. On a purely mercenary level, this is bad for LTV as an incapacitated player is not a spending player. I am operating from the assumption that this will be a problem in my current design because of how many Light Side Techs I am stacking. So I have built in an 8 hour per day play cap across the board on all accounts to allow the player to come down and seek food, liquids, sleep, hygiene and maybe even work if for some strange reason they still do that. I will adjust that number up and down as data comes in and I see how players respond. I believe such a cap will be well tolerated as long as it applies to everyone.

The Wall

No comments
You need to sign in to comment