Part of the problem with saying you can be dead wrong 50% of the time and still come out ahead is that you're assuming that you're dead right the other 50% of the time. The reality is that the "dead wrong" number is probably fairly accurate at around 50% for most people, but the "dead right" side is probably much lower. A big chunk of that other 50% is going to be made up of guys that you were kind of right on and guys that did OK but you weren't completely off on.
Yeah, there's definitely some truth to that. There is a huge spectrum of potential career outcomes for any highly-touted rookie RB. Outright useless bust (Irons, Shelton, Pead). Mediocre spot starter (Felix). Mediocre 2-3 year starter (Greene, Mendenhall, Ridley). Outright star (McCoy, Peterson, Rice, MJD). Even if you successfully identify Seastrunk as a quality talent, he may end up being more Spiller/Mathews for FF purposes than McCoy/Peterson. However, I think there's another variable your post overlooks. And that's the opportunity to cash out for significant value.
Even if Seastrunk ends up being more of a tease than a reliable every week FF starter, his draft background as an (I'm guessing) 1st round pick will provide a pretty big cushion to his perceived value. If he shows any flashes of talent when given the opportunity to start, almost everyone will rate him as a top 10 dynasty RB. Look at Richardson, Martin, and Wilson last year. Even Miller got a lot of hype and he was a 4th round NFL pick. Same story this year with Bell, Lacy, and Bernard. Are all of these guys long term rocks? No, but at one point many of them were valued on that level. Ditto with Mendenhall, Spiller, Mathews, Stewart, McFadden, etc. There's not a perennial star in that group, but each one of those players was valued accordingly at one point.
It isn't reasonable to expect owners to always separate the frauds from the legit talents and act accordingly. I was able to get away from Mark Ingram before his value imploded, but I got stuck with Mendenhall, Mathews, and Kevin Jones in some leagues because I bought the hype rather than cashing out for top value when I had a chance. So I'm not going to suggest that it's realistic to think owners are always going to be able to take advantage of the escape hatch, but it's a legitimate variable. If you draft rookie X, he shows moderate flashes as a rookie but you decide that he's not legit long term, you have the opportunity to get out for a decent return before his value collapses. I mean, I think LeVeon Bell is a plodding tub of goo, but that perception isn't unanimous right now. If you own him, you have the opportunity to act right now. Even if he's a fraud long term, there's a sell window.
I think the reason why this argument is never resolved either way is because the viability of the general principle is going to vary wildly depending on the specific case. For example, if you traded Adrian Peterson for Beanie Wells and a 2nd round rookie pick, you look like an idiot in hindsight. If you traded Larry Fitzgerald for Dez Bryant and a 2nd round rookie pick, you look pretty good. So inevitably what happens in these discussions is the risk-averse side constantly harps on the dangers of buying junky players and the upside folks emphasize the potential to reap big profits by taking advantage of a frightened wait-and-see market. Neither side is totally right or wrong. It's about spotting the right opportunities and acting accordingly when they present themselves.
That doesn't resolve the argument either because there's a lot of disagreement in the ability of owners to make good spots. If you don't think anyone can identify talent much better than random chance alone, the idea of trading established commodities for what you believe to be lottery tickets obviously won't carry much appeal. If you think that it's possible to eliminate some of the bust risk through subjective analysis, you're probably going to be more intrigued by the idea of trading a proven star for an untested player who you feel has a very good chance to hit. I fall more in the second group and so I'm more open to this kind of idea, although more cautious than I was in previous years. Players like Demaryius, Marshall, and Calvin who fill up the box score every week are super rare, so I don't think you can look at very many prospects and reasonably expect them to hit that level. Sometimes a player like that will come along. What's probably more common is to see a prospect who has a likely range between pretty good---great. I feel that way about Tyler Eifert. I don't know if he's just going to be pretty good (i.e. Todd Heap and Greg Olsen) or great (i.e. Jason Witten and Tony Gonzalez), but I'm pretty sure that he's not going to be a bust. With that being the case, I feel comfortable assigning a high value to him.
That's kind of how I see Bernard and probably Seastrunk. I don't know that they're going to be perennial rocks like Frank Gore and LeSean McCoy. However, I'm pretty sure that they're at least "tease" level talents like Ryan Mathews and CJ Spiller. If that's the floor and stardom is the ceiling then I think you might be getting the right odds to take one of those guys over a much older, much more established player. If you switch out Bernard for someone that I'm a lot less confident in (like Bell or Ball), I'm more on board with your take. Those guys probably won't be great long term players and have a pretty high chance of busting outright, so I'm not interested in giving up a proven talent for them.