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." Today, we'll show you whether those bigwigs actually know what they're talking about. To help, we've enlisted Motley Fool CAPS to track the long-term performance of Wall Street's best and worst.

Applause for AMD
Advanced Micro Devices (NYSE: AMD) wowed the Street yesterday, reporting $1.574 billion in revenues (within a whisker of the expected number) and $0.09 per share in adjusted profits (beating by the proverbial penny). Earnings, at $828 million over the past 12 months, are 75% higher than what AMD reported for its 2010 fiscal year. As you might expect, the "crowd went wild," and AMD shares are up 15% as of this writing. That's not the end of the good news, though. In fact, according to ace stock picker Canaccord Genuity, it might be just the beginning.

Responding to the earnings report, Canaccord announced this morning that it's upgrading AMD shares to "buy" on a belief that the company has finally turned the corner. "More benign PC headlines," better guidance for Q3, and a strong suspicion things may get even better than AMD is promising all have Canaccord feeling optimistic.

Sure, hardly a day goes by that The Wall Street Journal doesn't run an article on how Apple's (Nasdaq: AAPL) iPad is taking over the computer industry, or how Qualcomm (Nasdaq: QCOM) and ARM Holdings (Nasdaq: ARMH) are dominating the race to supply mobile devices with computer chips. But on Wednesday, PC-centric producer Intel (Nasdaq: INTC) announced better-than-expected revenues in its PC chips unit, and predicted revenue growth going forward will be in the 8% to 10% range, rather than the "low double-digit" forecast it had given previously.

Never underestimate the power of low expectations
On the one hand, Intel reassured investors that things aren't as bad as they've seemed of late. On the other, the company essentially lowered the bar for outperformance going forward, making it easier for companies like Intel, NVIDIA (Nasdaq: NVDA) and ... AMD to post additional "earnings beats" in the quarters to come.

But is Canaccord right?

Let's go to the tape
On the one hand, I have to admit that I'm still skeptical of AMD's profits. Sure, the headline numbers look good. But if you read past the headlines, you'll quickly see that the objections I've raised against AMD in the past still hold true: GAAP profit or no GAAP profit, AMD is still burning cash at a prodigious rate. Over the past 12 months, negative free cash flow amounted to $505 million, at the same time as the company claimed to be earning the aforementioned $828 million "profit."

I also can't help but note that over the course of the five years we've been tracking Canaccord's performance, this analyst has proven anything but infallible on its semiconductor recommendations. While this is the first actual "buy" rating we've seen Canaccord assign to AMD, it's recommended rivals Intel and NVIDIA in the past -- and been wrong on both. Indeed, to date, Canaccord has published 26 affirmative buy/sell recommendations in five years within the chip industry. Only 48% of them have managed to beat the market.

Caveat emptor
With those caveats in mind, let's now look at what it is exactly that has Canaccord thinking AMD will do better in the quarters to come. Among other things, the analyst points out that: "Management guided revenue up 10% Q/Q." On the one hand, this matches Intel's outlook. On the other, Canaccord thinks AMD's estimate "looks beatable."

Why? Because "[notebook] unit growth of 5% implied in Q3 guidance looks conservative in light of notebook ... builds tracking toward 7%+ and with AMD gaining share. In our view, 12% sequential revenue growth is a more likely scenario." Canaccord also thinks "ASPs should move meaningfully higher." Last but not least, the analyst notes that across the computer industry, AMD's market share sits at "unsustainably low levels. According to Mercury Research, in Q1 AMD's server share was ~ 6.8% compared to peak share of ~25%, notebook share was ~13.2% versus peak share of ~18%, and desktop share was ~25.5% (typically ranges between 25-30%)."

Foolish takeaway
In essence, Canaccord is playing a "reversion to the mean" theory on AMD. The company's doing worse than it has done historically, and the analyst thinks AMD's due for a bounce. And the analyst may be right about that.

However, for the time being, AMD is still playing second fiddle to Intel in the PC industry. Completing the analogy, it's doing so while its free cash flow "burns." Unless and until that changes -- until AMD begins generating honest-to-goodness free cash flow to back up its claimed "net earnings" -- I'm personally not interested in owning it.

My advice: If you own AMD shares, you're 15% richer today for it. Congratulations. Now take your winnings off the table and wait for the actual business to improve before buying in again.