At The Motley Fool, we poke plenty of fun at Wall Street analysts and their endless cycle of upgrades, downgrades, and "initiating coverage at neutral." So you might think we'd be the last people to give virtual ink to such "news." And we would be, if that were all we were doing.

But in "This Just In," we don't simply tell you what the analysts said. We'll also show you whether they know what they're talking about. To help, we've enlisted Motley Fool CAPS, our tool for rating stocks and analysts alike. With CAPS, we'll be tracking the long-term performance of Wall Street's best and brightest -- and its worst and sorriest, too.

And speaking of the best ...
In the Twilight Zone that is Wall Street, bad news became good news last week. No sooner had graphics card specialist and Motley Fool Stock Advisor recommendation NVIDIA (Nasdaq: NVDA) reported an earnings miss, coupled with guidance about 2% below Wall Street's sales projections for the current quarter, than two analysts jumped out and ... upgraded the stock.

Yes, Fools, hard on the heels of pre-earnings upgraders Wedbush Morgan and AmTech Research, Stifel Nicolaus and Canaccord Adams threw in their lots with NVIDIA as well. And perhaps even more interesting than the unanimous endorsements, was the alacrity with which Stifel and Canaccord spun on a dime and switched from "sell" to "buy" in a heartbeat.

Why the change of heart?
Canaccord brushed off the news of NVIDIA's disappointing guidance, predicting that this is the last estimate cut NVIDIA will likely need to make. Canaccord sees some stability returning to the PC market, and, presumably, an upswing shortly thereafter. Stifel echoed this view, saying "further downside risk is limited" and agreeing that it sees "an improving outlook for the company's earnings."

So, as Wall Street comes to a consensus on NVIDIA's prospects, should you listen? We discussed Wedbush's and AmTech's records last week. This week, let's take a look at Stifel Adams and Canaccord Nicolaus.

Um, you got those names mixed up
Did I? Well, you have to admit that it's an easy mistake. When you look at their records on CAPS, the two analysts are -- for better or worse -- basically indistinguishable. Both have unimpressive records for accuracy, hovering just above 50%. Both have considerably better records for scoring big on the guesses they do get right: Each ranks in the 87th percentile of investors as scored by CAPS. Reviewing a few of their respective semiconductor picks, we find:

Company

Canaccord Said:

CAPS Says

(5 Max):

Canaccord's Pick Beating (Lagging) S&P by:

Xilinx  (Nasdaq: XLNX)

Outperform

****

36 points

Cree (Nasdaq: CREE)

Underperform

****

5 points

Credence Systems  (Nasdaq: CMOS)

Outperform

**

(75 points)

Which demonstrates both Canaccord's ability to call more semiconductor plays right than wrong, and the severe damage a single mistake can do to the analyst's record when it is wrong.

Company

Stifel Said:

CAPS Says (5 Max):

Stifel's Pick Beating (Lagging) S&P by:

PMC-Sierra (Nasdaq: PMCS)

Outperform

***

27 points

Broadcom  (Nasdaq: BRCM)

Outperform

***

(15 points)

AMD (NYSE: AMD)

Outperform

**

(42 points)

Likewise with Canaccord's doppelganger. Some guesses it's semi-right on, others, semi-wrong. Overall, Stifel winds up with a rather mixed record in this industry. However, in its defense, I should point out that Stifel's last call on NVIDIA was one of the good 'uns. Shortly after we got CAPS up and running, Stifel placed a "sell" bet on NVIDIA in November 2006. When it closed it in late March last year -- with an upgrade from "sell" to "hold" -- it had racked up just shy of 20 points of market outperformance. It reiterated that sell rating last August and racked up another 18 points.

Foolish takeaway
Still, if the analysts' less-than-perfect records give you less than 100% confidence in their buy recommendations on NVIDIA, I understand your hesitation. In that case, I'd suggest you use valuation as your touchstone on this stock, because that's your best bet, long term.

Right now, NVIDIA is selling for 17 times trailing earnings, which looks to be an entirely fair price given consensus expectations of 17% annual growth for the stock going forward. And as I've said before, at last report, the company was generating about 35% greater free cash flow than its GAAP figures let on.(NVIDIA hasn't updated us on its cash flow picture yet.) So the way I see it, what we have here is a stock that's at worst fairly priced, and at best downright cheap. Call me a Fool, but I'll take that bet.

Do the friendly Fools at Motley Fool Stock Advisor agree? Take a free trial and find out.

Fool contributor Rich Smith does not own shares of any company named above. You can find him on CAPS, publicly pontificating under the handle TMFDitty, where he's ranked No. 862 out of more than 100,000 players. The Fool has a disclosure policy.