It seems the same as with any "hit" driven environment. Getting a top-40 song, a top blog, the next NYT bestseller, or top iPhone app takes a lot of work and a little luck. There are those who are suited to it, persevere and get that big break. Then there are a few who just don't get lucky and thousands of wannabes who fail to execute, be it on minor flaws or epic levels.

I think the best description of the process comes from an editor explaining why so many manuscripts fail: http://nielsenhayden.com/makinglight/archives/004641.html#00...

It's harsh, but has given me insight into how the startup world, including iPhone app startups, works.

Apparently, even being a drug dealer works the same way.

http://www.freakonomicsbook.com/thebook/ch3.html

I don't have statistics to back me up, but I'm fairly certain the odds of survival are considerably higher for iPhone developers.

Yes, but the pyramid structure of wealth (concentrated at the top) definitely applies.

With the right binoculars, you can see Steve Jobs standing way up there on top. I do get what you mean about that, but it seems smaller pyramids can be built next to the big ones, even if they do pay taxes (so to speak) to the bigger ones. Maybe the iPhone isn't the best environment for that? Or maybe the real money is in building apps for other companies who need their services extended to a new platform.

I strongly suspect your last sentence is the key!

The survival rate for both groups is exactly 0%.

The problem doesn't lie with the curve.

It's that even those on top aren't winning big. Why play the lottery when the top prize is modest, and the risk is high?

What you meant is Pareto Distribution, not Gaussian distribution. They are completely different.

Maybe you're right. Apparently Pareto originally used his distribution to describe wealth allocation & intuitively it seems fair to assume that 20% of the developers account for ~80% of the earnings. But are you sure?

Btw, if developer talent is normally distributed (talent, it seems, usually is), shouldn't developer earnings also be? I'm not a statistician but it's an interesting question..

The impression I learned in my statistical mechanics course 20 years ago is Gaussian distribution works well in the case similar to Random Walk. If the system has restrictions on microscopic behavior, then Gaussian distribution will not apply. I think the benefit of talent may not be in normal distribution. The distribution of probability depends on the case and conditions that we pay attention to.

For example, a talented chef can only serve similar amount dishes in a night as less talented chefs. But the dishes they made taste much better than those and charge more.

At the same time, a talented plumber or a talented doctors can not serve more customers or patients than a less talented plumber or doctors.

It seems when the benefit of the talent scales, the distribution of their influence is Pareto distribution. So it seems developers, musician, artists, physicists fall into this category.