I don't think behavior blocking alone is the holy grail of security really. It's just one component of several, that sum up to a great security concept. Each component has advantages, but also disadvantages, and the right balance effectively does the trick.
Surf protection is a good first layer of defense because it's blocking stuff long before it can get started. It's a safe bet (technically), but the disadvantage is that you can never have a complete list of hosts to block as they change every minute, and not all malware is spread via websites directly.
File guard usually does the biggest part of the protection job. It's also relatively safe because the scan engine gives clear results on files without many FPs, and after all, it's able to detect >90% of all threats reliably. Downside is that it requires frequent updates. It naturally misses zero day threats that no AV vendor knows about yet (setting side heuristic methods).
Behavior Blocker's job is to provide a safety net to pick up stuff that the other two layers missed. It's really difficult for malware writers to bypass it because the behavior of a program can hardly be manipulated in a way it gets invisible. But the downside is that it is a very blurred detection. Good software often behaves exactly like malware (from a technical view), so that needs to be reliably filtered out. If you turn off cloud lookups in everday-use, you'll probably see many false alerts, so I'd strongly suggest to always keep it on, unless you feel comfortable dealing with alert boxes for good programs.