This is a fallacy. People use this term "priced out" as if it meant some sort of process, but it means nothing more than that the investor thought the startup's stock was too expensive. And it is very stupid to let valuation decide which startups you invest in, because the variation in outcomes between startups is orders of magnitude greater than the variation in valuations. I.e. there is no value investing in startups.
What we have here is a case of anecdotal evidence. A founder happened to get some investors who hadn't invested in other startups because they felt the valuations were too high, and those investors turned out to be really helpful. But there are other investors who are willing to invest at high valuations who are helpful, and investors who seek out low valuations who aren't.
When a startup is crazy successful, it doesn't matter what price you paid to get in. But in terms of portfolio, price does matter. The higher the average price you pay, the fewer deals you can afford to be in, and the lower your chances of getting into that one runaway success. Sure, its far more important to choose the right companies. But it is rational for YC, and others, to be price sensitive.
That's only true if you have to get a certain amount of stock, which is the case for VCs doing series A rounds, but not for angels or VCs doing seed rounds of the type Chris wrote about. How many deals an angel or seed investing VC can afford to be in is a function of the average amount invested, not the average valuation.
From a founder point of view, I think fishing for the investors that missed the bar by a few centimeters is a clever strategy.
I would personally favor an involved investor over a rich investor any time of the day, because an investment isn't just a spreadsheet you agree upon.
May I recall that:
startup => company
company !=> startup (a company isn't necessarily a startup)
This is not investing, this is gambling. When I invest, I have a hedging strategy and caps. If you offer me a gold mine for $ 15 and I have only $ 10 to invest, sorry, I'll pass.
Of course this only works if you are capable of choosing investors well.
(I guess you could consider the case where you got one at a huge discount through luck -- getting a $100k product for $10k. Then, I'd basically treat it as a $100k asset for the purposes of how much to help -- the 10x gain on the $10k already happened the moment you made the deal, even if it's unrealized. I'd be happy to buy $100k negotiable assets for $10k, but in general, bargains aren't -- there is a reason you're getting a discount, ranging from the founders being noobs who will equally be likely to get taken advantage of by others later, or the deal has more risk than you thought, etc. There may be some cases where assets are systemically mispriced -- I think Dave McClure thinks non-US startups are one, especially from LatAm and SEA.)
If I were making an investment into a public company (or a really late stage private company, like investing in Facebook the year before the IPO), I wouldn't think I'd have as much to contribute, true, but I'd assume I could help a $2.5mm valuation startup about as much as a $10mm valuation startup.
(Actually, my #1 metric on helping would be "is it fun for me to help" -- I'd probably end up spending all my time helping portfolio companies with security or infrastructure issues, and would avoid helping with design, HR-fiasco, or fundraising issues. And, if I could get Apple, Tesla, etc. to value my advice, I'd almost work for them for free ($50k/yr?) just to make them 10x more secure than they are now. It's just easier to get someone to value your advice when he's paying you $300/hr.)
I don't dispute that some people would behave like you're saying, but I would not, and most of the professional investors (super angels/VCs) probably wouldn't. I think this might be the difference between "professional" investors and angels.
You can argue that they're irrational/stupid, or that they're not the investors you want (some of them actually are very good, you'd certainly recognize them by name), but it doesn't change the fact that they exist.
I can't comment on relative usefulness of valuation sensitive and valuation insensitive investors.
It seems like that data wouldn't be anecdotal evidence, but cover a fairly large sample size.
And it is very stupid to let valuation decide which [house] you invest in
And it is very stupid to let valuation decide which [career] you invest in
And it is very stupid to let valuation decide which [spouse] you invest in
These are equally non-sensical. just shows your valuation methodology lacks any/appropriate resolution. so all you are saying (at best) is you (or one similarly situated, etc) doesn't know how to "value a startup". Even if this is true, it does not follow that "the market" knows more than the sum of its participants. And therefore it does not follow that "the market" knows how to value starups either. The market works only if someone at least knows what they are doing. You are essentially promoting the idea of index investing, which is one thing in public markets (with public info) but another altogether in opaque and thinly traded private asset classes (b/c: paradox of efficient markets, etc)
Whether or not this observation changes anything is uncertain. but self-awareness is usualy highly valuable when one's at the resolution limit of their reasoning.
 There is an abstraction issue relating to "black swans". Namely, is the valuation of a portfolio of startups easier/more tractible than the valuation of individual startups? If so, one value the pool X and tha assign X/n(a) to any individual startup, where N is number and (a) is proportionate variable. This is one way to say or reconcile the investment in startups with the statement that investing in any one startup is not an analytically tractable problem. But notice it just pushes the Valuation question out to one level: you still need to make a bet on the valuation of the pool. eg, the following are contrasted idea:
Ex A: And it is very stupid to let [company] valuation decide which startups you invest in
Ex B: And it is very stupid to let [pool] valuation decide which startups you invest in
Hypothesis: A rational person/investor needs to believe they have a grasp of one of the two formulations, otherwise they are like monkey throwing darts at a list of stocks.
I don't know whether it's true or not that variations in outcome are orders of magnitude greater than variations in valuation. But (a) it does not sound like an overtly crazy insight, and (b) if that's true, then Graham's argument is mathematically obvious: get in to Youtube or Heroku at 1/10th the stake you'd get in Carsabi and you're still way ahead.
For the inevitable long reply you're going to provide to make sense, it must directly engage Graham's core point. Maybe he's wrong on the numbers, and seed-stage valuations range as widely as outcomes (you're probably going to find out you're wrong about this). Or maybe there's some signaling you can uncover that shows lower-valuation companies tend to outperform those that receive outsize seed valuations. But it will be something like that, not, respectfully, a series of allusions to careers and spouses and draft horses and whatever else it was you said.
That assumes that variations in outcome are correlated with variations in valuation. If they are not (which may be the point of mentioning “black swans”), then an investor is better off investing at lower valuations.
-- Read the footnote, and read his essay on Black Swans.
If you think this is "babbling", you're just not paying attention.
 And understand a black swan is a non-stochasic event. > http://www.paulgraham.com/swan.html
This is only true if the math is hard. If the math were trivial, people would just do the math. The math is hard. Therefore math is of no use/irrelevant/stupid. Does the last sentence follow? Not obviously.
PG wrote a nice essay on why the math is hard.
There is some argument that the complexity can be simplified at higher levels of abstraction. There is another argument that the goal of a VC is to make the math work at a higher level of abstraction. There is empirical evidence that both arguments are pretty good.
So, its not unreasonable to (1) agree with PG that the maths are hard; but (2) disagree we should dis-regard maths altogether...so we don't throw the baby out with the bathwater.
[Addendum: YC invests at a fixed valuation in its companies, wich are selected via a competitive admission process. This is a logical approach if you believe the math is hard on a granular basis, but is more tractavle at higher level of abstraction. The co-hort is thus the unit of investment, not the company.]
This point is so obvious that I find it hard to believe you wrote the above comment in good faith.
If "what to do" was "very easy", everybody would do "it". But that doesn't happen. Investors are not all created equal, and investor returns show massive variance (ie, they are not observed to be uniform.) So something is very wrong with your approach/level of abstraction.
You might want to also read this comment from paulsutter that PG responded to, if you think this is a personal attack on you.
 ref: http://news.ycombinator.com/item?id=4859922
That is yet another very counter-intuitive thing about the startup world. In basically every other finance vertical, terms matter greatly. Not so much with VCs. That's why a lot of finance guys make poor VCs, they fundamentally misunderstand the game.
You're almost right. Early stage investing is very much power law driven, later stage (more VC than incubator territory) comes in several flavors, some more relying on outliers than others.
Plenty of VCs won't even look at companies that are not yet capable of generating substantial revenues and profits, they are there as growth accelerators or to diversify founder (and early stage investor) risk.
There are as many risk profiles as there are VCs, you can't just lump them all on one pile.
CB Insights says that 50% of seed companies will go away. And the reward that angels are given to take this risk is a 20% discount to the next priced round.
However, one of the benefits of writing smaller checks or running smaller funds is that you're less dependent on monster exits. For investors who are going smaller, valuations and terms do matter a lot more than at typical venture funds (that are very dependent on monster exits).