I forget where, but I remember reading a while back that a twin engine size of 800 is just too big to have the same reliability as a triple or a smaller twin. It has to do with the difficulty of cooling those big pistons down, with the breathing characteristics of a huge twin making things worse. There's also a ton of weight swinging around in the bottom end , which makes crank reliability problematic. These are reasons why Yamaha never built a twin larger than a 600.
The other issue at work is emissions regulations. Somebody from Polaris stated in a magazine interview last year (again, I forget which magazine) that it costs ten times the amount to design an engine as it did in the late 90's. This is because they have to go through a huge hassle and expense to get the EPA certification. The engine has to be designed around it. Then, whenever they want to make even a relatively minor change to it, they have to go through the whole process again.
There have been threads on here about why the 800 CFI has been so troublesome, and Polaris can't "get it right". Apparently Polaris, unlike other manufacturers, didn't use their emissions exemption on the 800 CFI engine, and this forced them to immediately meet the new regulations on the 800 dragon. Running extremely lean in the midrange was the only way they could meet the regs, and was a major reason that that engine was a blowup special. If I'm not mistaken, the loose pistons had something to do with meeting emissions as well. It's also the reason they are so stingy with oil injector settings on the pro.
Like Jaynelson said, these motors are putting out a lot of hp per cc. An Indy 500 or XLT wasn't a high performance engine. These are. We're getting close to the MX bike level of power per cc. As such, top end replacement intervals are coming down as performance goes up. On a 250 two stroke MX bike, anything more than about 20 hrs. is running on borrowed time for a piston. Suzuki used to spec a new piston on the RM125 every two hours. Yamaha used to include a top end kit with a new YZ just so the first one was free. Nobody blew up a 250 after 40 hours on a piston and blamed the bike or the manufacturer. If you were considering buying a year old bike and the guy hadn't done a top end yet, you knew that was the first order of business after you brought it home. We're getting to the same level of performance, and parts replacement intervals, with sleds. It might be worth asking "Do I really want a high performance engine?" If you do, it might be wise to forget about the warranty and get used to working on the thing yourself, doing top ends every 40 hrs. or so, and bottom ends every 100. If not, maybe the current crop of 800s is not the best choice.