@JohnRoxton A tweet from EY reminded me that I never updated this comment. From chatting to energy traders, this issue doesn't really seem to apply in their world, for two main reasons.
TL;DR: Comparisons to commodity markets are not useful or particularly relevant here. This was not a productive avenue of exploration.
1) Taking a position in a 'real' market is a much bigger deal than betting on Manifold (obviously - but the scale may not be obvious). Suppose that you want to bet that the price of TTF in 2030 will 'go up' vs the current price. I can't actually find any evidence that anyone has done this outside of some Gazprom-run LTCs, but the last public price for a 2028 trade is ~€30/MWh. The smallest trade you can normally make in that market is 5MW, so you're looking at 5 x 24 x 365 x 30 = €1.3m notional volume. Because this position is so far in the future the collateral requirements for this will be massive - that position will likely have massive swings in profitability until it closes (nb I'm using 'collateral' here as a shorthand for whatever risk mechanism is used to allow the trade to take place, which will vary depending on where and with whom you're trading). This massively increases the opportunity cost of such a trade and makes it very unlikely for it to be the best EV of your cash even if you're 100% accurate!
Which, according to my logic in the post above, ought to make the pricing even worse in these futures markets than it is in Manifold because it's harder for people to correct it, which brings us on to:
2) The pricing in these markets isn't as wrong.... because it just doesn't exist. As I said above, there are no public trades available for 2030 TTF. Some firms will have done private deals but those will be directly linked to securing physical supply and have motives that aren't 100% profit-related (eg they've been made by or at the demand of govts who are just spending other people's money and don't care much about value). Internally, firms will have their own predictions of the value but they won't be sharing those and there's no profit motive to contribute that knowledge to a market for reason (1).