I think really that the answer is to take risks at selective, optimal times. For example, selling 20 kilos of cocaine to make a ton of money is probably not worth it, even though the reward is huge. Not leaving your house because you don't want to get hit by a random car is probably not worth playing it safe.
Think of it this way. Say a random risk has a 1 in 100 chance of backfiring. That means that if you take 100 such risks over the period of your life, you may never get screwed, but most likely is that you will get screwed at least once, maybe more. If you take the chance 1000 times, you are almost certainly going to lose a few times. It doesn't matter if your risk is only 1/100, because you hopefully will live a long life full of risks. You will have to take some of them, and you can avoid most. The key is to pick the ones with the best risk/reward ratio, and be selective over when to take chances, but commit when you do, and don't look back. Know at that point that the risk is worthwhile and that you have to take some sometimes. At that point, just do it.
I hate it when stupid friends take huge risks (like the example above) and justify it by saying "it's 1 in 10,000 that I get caught". Sure it is, the first time, but not when you do the same dumb crap every day. It's 1/10k that you will get caught THIS time, but very likely you will get caught at some point in your life.