For those people saying it was a test, let me see the proof. I want proof that this was a test.
I think this is a great situation where basic probabilistic thinking is useful.
Let's start in the case of zero information as a baseline. We know nothing besides that an event has happened and could have one of two causes. You should naturally assign a base probability of 50% to either outcome. In this case of course we don't have zero information. We have quite a lot. It's up to you to measure out the importance of each bit of information and determine whether it makes on cause or the other more likely. So for instance:
Market experiment more likely:
Market pricing is going to be a system with numerous levels of redundancy to ensure its security and stability. There's a reason "oopsie" style pricing mistakes are all but unheard of.
-EA has a history of unethical actions and lying to consumers for the sake of attempts at increasing profits in unusual ways. Simcity/sims are recent examples of this.
-The precision of the accident. The price wasn't set to $59.99 or other more normal prices. It was set to figures that EA might believe consumers would find reasonable for a demo.
-The lack of randomness in the assignments. Both games targeted were mainstream sports titles from EA.
-Companies this gen have been far more aggressive than usual in relation to creating new monetization methods even when it increases dissatisfaction in consumers. EA is at the forefront of this brave new world of pay to win and offering out free-to-play games so abusive of players that they've faced legal/governmental challenges in labeling them "free." Charging money for something that's always been free would not be far from inconceivable for EA if they felt they could "get away with it".
Pricing mistake more likely:
-This was restricted to the XBone which has a less broad audience than the PS4 and would potentially devalue any data collected.
-Trying to charge money for demos is somewhat unprecedented and would be a now low even for a company like EA.
I am giving EA's statement zero weighting since they have a lengthy and consistent record of lying to consumers, including extremely recently. As such I think they deserve 0 benefit of the doubt. For a more reputable company the fact they simply stated it was a pricing mistake would certainly be not only a weighting on the pricing mistake side of the bar, but a very substantial one.
In a nutshell very few people know exactly what happened, so it's up to you to lay the odds as you see fit. People who are 100% certain this was a market experiment (lacking inside knowledge) are idiots. People who are 100% certain this was a pricing mistake (lacking inside knowledge) are idiots. The truly reasonable position is to simply think of the scenario probabilistically. All that's left is to debate and decide on exactly which action/event/datum that should be considered and what their individual weighting should be. I think in general you're going to be hard pressed to commingle the variables such that a pricing mistake is more likely, let alone substantially more.