The economic rationale for hackathons is that they create value by injecting uncertainty into development. The trend towards “pitches” for hackathons, while instituted with good intent, destroys value.
In the 3X: Explore/Expand/Extract model, hackathons are exploration. The payoff of a hackathon is convex, a very small chance of a very large payoff (but an almost-certainty of over-eating and sleep deprivation).
Success is always a surprise (except to the wacko who proposed it). The bigger the payoff, the bigger the surprise.
The mistake of pitching for hackathons is that the highest expected value projects are exactly those most likely to be thrown out with the bathwater.
Those calling for pitches for hackathons are smart and trying to do good. I’m guessing here (feel free to chime in and I’ll correct), but it seems to me that the positive role for pitches is to protect some limited resource: hardware, expertise, time.
Rather than filtering, higher value strategies are to:
Increase the supply of the limited resource
Reduce the cost per hack of the limited resource
Let the hackers negotiate access to the limited resource instead of resolving contention early
As humans, we are incapable of making rational decisions about infrequent events. That’s why hacks tap into passion and intuition and a little primate dominance display. Hacking gives us access to decision wisdom otherwise unavailable. Embrace the chaos.
P.S. A value-creating hackathon pitch would be to present the hack idea and if everyone approves, to absolutely forbid the hack on the grounds that such a good idea is likely to happen anyway and we don’t need to waste hack time on it.
Looking for encouraging & inspirational ideas for your organization? Commission a custom talk about The Forest & The Desert, about Tidy First & Software Design, or about leading with purpose.

I am not sure what you mean by pitches. I am guess you mean that before the hackathon starts, each team pitches their ideas, and only those that are selected by judges are greenlit?
I agree that sounds kinda stifling for creativity. Instead, why not just let people do what they want, but then have them present at the end to a panel for judging or feedback.
Years ago when I was at Shopify, we did indeed use "pitches" for hackathon ideas. The difference, though, was that you were looking to recruit interested people to work on your idea, rather than trying to get the idea green-lit. Any and all ideas were welcomed even if you worked alone.