First, read about Newcomb's problem. Basically, you get a choice: take a $1,000 bonus, if you want, but if I think you would take that bonus, you'll find I haven't given you the keys to the Wonka Factory. Whatever you take is up to you, but my prediction is pretty okay – I do know you pretty well from the tour, after all. Here's my pitch: these structures, of prediction and social consequences, are very common.
In particular: I want corporations to only pollute a tiny amount, when it's really helpful. If I predict they won't do that, I will impose regulations. Those have a cost to good and bad actors alike – compliance ain't cheap – but it helps me build the system I want.
This has that same structure. Pollute more to maximize profits, sure. But if I think you would, I'll make the regulations, which harm you. In the case of the American coal industry, the regulations will basically put them out of business. So you can pick up the extra money, but being that type of person costs the larger payout. Of course, taking the extra money has strategic dominance – if you gobble all possible profits, that's good with little regulation or heavy regulation. It's more money.
But, of course, we'd prefer modern corporations use a different decision theory. We'd prefer they "one-box" on regulation – or, at least, I do. This reveals my own preference in Newcomb-like problems, but it extends past the precise limits of the prompt. I've always felt my best social asset, aside from my goof-sense, was how I was basically predictable. I keep my word, or try to. I'm punctual. I don't have sneaky motives, and I'm willing to jump into tough conversations pretty early on. I'm as polite as I can manage, and truly have gotten little negative feedback. But hard conversations are hard, and harder still when delayed or disguised.
Honesty and guilelessness are, in some ways, the ultimate game theoretic strategies. Maybe. I just behave this way because every time I did something else it made me sick to my stomach, or had my anger-prone older brother unleash violence on me in childhood. It isn't strategic for me, but again, that's probably a good thing. I come by my guilelessness honestly. It was hard won from all the times I hurt someone's feelings and felt regret for, sometimes, years and years.
Perhaps this is the true recommendation: don't "one-box on Newcomb-like problems", just feel shame and regret when people predict something about you that makes them not want the best for you. Be as good as you can be in their imagination – and to keep up the "ruse", of course, it'll mean you have to be as good as possible all the time.