I'm framing this post from the perspective of ebay being cheaper than BV or seller asking prices, in most cases. I know there is the occasional odd reversal of the trend, such as the Shutt auto or the Tavares YG mentioned above.
Sorry to hijack the thread but...
Respectfully, I think your first assertion's off base. Regarding eBay, not all completed sales actually even happened, unless you include non-paying customers, shill bids, fraudulent purchases, and other dirty tricks that I'm probably unaware of.
These represent an astoundingly small percentage of ebay auctions. They may get the most attention on a hobby message board such as this one. While I know using ebay feedback as a proxy for numbers on successful transactions may not be perfect, the fact that most sellers (and especially sellers of volume) are
conservatively above 99% in their ratings would seem to indicate that no more than about 1 in 100 sales are fraudulent or not completed.
Of course, throwing out the shill bids and non-paying customers, you'd probably skew the actual ebay sale numbers
lower.
Setting those non-sale sales aside, it's also a ridiculous premise (to me) that any single valid bona fide sale that actually happened is necessarily a reflection of fair market value for a number of obvious reasons: poor listings, geographic bias, asymmetry in technological expertise between and among new/veteran sellers and new/veteran buyers (that doesn't happen at an LCS or card show for example), outliers, etc.
This is a much more valid concern, but I think you're overstating it a bit. Again, we're dealing with a minority of ebay sales, and I'm totally happy to toss out the numbers on a 3-day auction listed by a single-digit feedback seller with a misspelling in the title. No one reasonable is going to try to use something like that as a real measuring stick for a card's worth.
If a properly-listed, successful transaction takes place on an auction that goes for 10% of book value, then you have what I stated before (and not necessarily anything more or anything less): one confirmed sale of the card that actually happened. Beckett's book values are not guarantees of even that much.
Ebay means a lot of different things to collectors, so it's important to remember what the ending price on a card really represents:
the lowest amount necessary for the winning bidder to exceed the bids of all other interested parties.
To me, the way some people use eBay completed sales to justify values can and has been every bit as divorced from reality as those who take advantage of Beckett's valuation shortcomings.
I'm happy to use ebay sales to justify the value in a card I'm looking to acquire, because ultimately, they play a large role in what I end up paying for the card. I love Hobby Insider and working out deals with the people here, and I'm certainly willing to conduct transactions that aren't tied to the bare minimum of what ebay would suggest. However, let's say I'm interested in a card whose value is in question. If it sells for $10-$15 on ebay, and its book value and/or message board asking price is $50, then what is the card's real value? Subjectively, the card's value
to me is the one I end up paying -- the one with my name as the winning bidder.