Historically, booms and busts are what the market does, go back over the last 150 years, since the end of the American Civil War - you'll see a boom, following by a bust - the Federal Reserve and the changes enacted in the aftermath of the depression has dampened it a fair bit, but it still happens - and could happen just like it did before.
Our financial system is very very fragile, and like a chain stitch, one loose thread in the wrong place, and the whole thing could unwind - thats the problem. So Remember in the end, a wave of bank failures are what made the depression last a decade, not just the recession itself.
However, it just doesn't matter for anybody with an investment horizon over 10 years. The gravity of the markets is upward. Timing the broad market consistently is statistically impossible.
The subprime mortgage crisis was statistically a once in a lifetime drop, only comparable to 1929. It's occurance was due to a fundamental flaw in the system. It wasn't part of the standard "pattern of the market." Everybody sitting on the sidelines waiting for the next flaw to rear its head will likely decompose before that happens.
The size of your future returns directly correlates to the amount of risk you are taking. If you are on the sidelines it's a statistical certainty you'll get destroyed by inflation.
A linear y-axis is terrible for viewing historical market data because you can't compare percent changes across historical periods. Because the price of the market has grown exponentially, any movement in 1929 for example looks like a minor blip in comparison to anything happening in present day (due to the exponential price appreciation).
On a log scale, you can see that movement in the proper context compared to 2000 and 2007.
Another way to think about it is that you don't really care if a stock moves $x. You care if it moves x%. Logarithmic scales let you visualize this more accurately.
On the other hand, retail investors are probably the least likely to dump money in after a crash, which is the best time (buy low, sell high). Instead they are selling and generally trying to protect capital. Then they stay out until we do it again.
Then, one day years later, after they've lost out on ridiculous gains waiting for the "big drop," they learn that when looking at a historical market chart to switch from linear to logarithmic scale.
Suddenly the crazy recent capitulations just look part of a steady climb with minuscule blips along the way. Go ahead and try it on Google Finance with the S&P. You'll be amazed.
edit Changed 'exponential' to avoid confusion