This is in one of the other HPFP threads, but there is an approximate known percentage for 2009 and 2010 models. It's based on the numbers that VWoA submitted to NHTSA as part of their first info request. They were supposed to document all HPFP failures. Based on that first submission, slightly more than 1 in 200 2009 models failed, while about 1 in 1000 2010 models failed, as of the time of that submission. There is of course some question as to whether VWoA's numbers were complete, and NHTSA's Oct 7 request is pushing them for way more information. Also, most of us here believe the rate will increase with miles on the fleet, so 1 in 200 might become 1 in 40; 1 in 1000 might become 1 in 200.
Measured a different way, another member of this forum (I can't find the post right now but read it earlier today) estimated that 2.5 percent of the CR driving members on this forum have had the failure. However, it is noted that people often join this board only after trouble, so you could say the members of this board are failure biased.
When you ask that question, I think you have to set a specific mileage boundary. I think most here (and I would especially rely on the mechanics' opinions, not mine) would agree that any fuel pump will fail eventually. Prior to these CRs, the expectation on the life of a fuel pump was, depending on who you ask, somewhere between 200K and 300K miles. Some of the people here with experience in larger motors have said they see pumps go even longer. And, importantly, the older pumps did not take the entire system with them every time they failed... So if you asked "would ALL of these pumps fail by 300K miles," then I think the answer would be a very large percentage, at least for early 2009 models. Later 2009 models will behave more like 2010 models, because there are four versions of the same part number pump. Only early 2009 models got the first version. Late 2009 and early 2010 models got the second version. Late 2010 and early 2011 got the third version. Any pump made after a certain date (exact is unknown) in November 2010 will be the fourth (and perhaps final) version of this pump, so mid to late 2011s and all 2012s should have the fourth pump. Whatever VW has done to these pumps will almost certainly reduce the failure rate relative to a specific mileage boundary, but WITHOUT solving the core design defect. In other words, you can feel better about your odds with a fourth version pump than with an original version pump, but if you intend to drive it for 300K you will still see it fail. But long before then there may be an after market or VW issued supplemental filter to keep the pump from destroying the entire system.
Also, just a tip for shopping... before buying a TDI be sure you know how to get a cell phone picture of the pump manufacture date, so that you know you have the latest pump version for the year model that you are buying. Personally I would not buy anything other than a 2012 at this point, which I just did, but that is another story for another post. The pump is the part directly above the alternator (the alternator being easily identified by its copper wire coils). The sticker on the pump is on the under side on the right, so you wedge your cell phone camera in between some hoses. The easiest angle from which to get the shot is from the right side of the pump looking left toward it. The fit from straight below is too tight, at least when using an iPhone.
Also, remember that the pump may be made well before the car, so do NOT go by the manufacture date of the car. I have noticed on Jetta, both wagon and sedan, that the pump tends to be made 3 months before the car. This is almost certainly a result of the parts being shipped from Germany to Mexico. In the Golfs I have observed that the pump and car often have the same month, though the pump obviously has to be at least a few days older than the car. The August build Golf that I just bought had an August 8 made pump on it.
