FORTUNE — Nearly three years ago I wrote a story saying the Federal Reserve should make bank “stress tests” an annual event. Now that it seems they have, I am not sure how wise my advice was.

On Tuesday, the Fed said that the nation’s largest banks would be able to weather a severe economic downturn. Even the four banks that “failed” the test – including most notably Citigroup (C) – weren’t told to raise more capital. Instead, the Fed said those banks should rethink plans to use their money to pay out dividends or buy back shares. So not quite an F. More like a C-. In the end, the test was less about survival and more about whether the nation’s banks would be able to continue to lend in another financial crisis – one with a 13% unemployment rate, a 50% drop in stock prices, and an additional 21% drop in the housing market – something they were clearly not able to do last time around. And that in the end was the problem.

Virtually no one believes the unemployment rate is headed to 13%, or for that matter any of the Fed’s other fake economic maladies, which also included a one-year 8% drop in GDP. Stress tests are supposed to test extremes. But the downturn the Fed tested for is of the 2008 and 2009 variety, and not the problems people actually worry about these days. For instance, in its extreme economic scenario, the Fed predicted interest rate on 10-year Treasurys would fall to 1.64%. But what would happen if interest rates skyrocket to 10%. A plunge in bond prices could cause huge losses in the banks portfolios and trading operations. What’s more, high interest rates are sure to cut demand for loans and corporations ability to issue bonds, drying up business for the banks. The Fed didn’t included that scenario in its tests.

Independent bank analyst Christopher Whalen said the stress test also included a rosy outlook for the legal liabilities the banks still face from the mortgage crisis. For instance, the Fed predicted that losses in its loan portfolio could cost Bank of America (BAC) as much as $60 billion over the next two years, which includes any money the bank may have to pay out to investors who are suing the bank to recover losses on mortgage loans the bank has already sold. Whalen says those costs alone could end up being as high as $80 billion for Bank of America.

Another hole that one analyst pointed to in the Fed’s stress test was second mortgages and home equity lines of credit. The Fed predicted that the banks could lose as much as $56 billion from those loans, or about 13% of their portfolio. Mostly because those loans have defaulted at lower rates during the current housing downturn than people expected. But given that nearly a quarter of all home owners are already underwater on their mortgages, the analyst thought the Fed’s predictions on losses on those loans should be even higher, particularly given that they predicted home prices would fall farther. Lastly, while the Fed did say it factored in problems in Europe into its test, unlike say mortgages, it didn’t break out how much the banks could lose overseas. Debt troubles in Europe remains one of analysts’ biggest concerns for the U.S. banks, and the U.S. economy in general.

And that may be the problem with running stress tests in general. No matter how “extreme” the conditions of the test are, they will always be backward looking, and as a result give us, bank executives, and most importantly regulators a false sense that all’s clear. So perhaps it’s not much of a surprise that most of the banks passed the Fed’s tests. The risks the Fed tested for are exactly the things banks have been most focused on barricading themselves against for the past few years. But it’s the things that banks, and more importantly regulators, least expect – like falling house prices in 2005 – that cause the biggest losses, and the biggest problems.