This topic has been on my mind for a LONG time (ever since I was on the Endyn TOO forum). After looking at the chamber and the piston, I came up with a piston design that allowed good, tight squish where needed, directed the charge toward the exhaust valve, and still kept a reasonable static compression ratio. Although, if done very well, the static compression ratio isn't really that important.
Basically instead of the piston having a symmetrical deep dish, the dish was a ramp shape (deeper side on the exhaust side of the chamber) and the edges closely followed the contour of the chamber. Obviously this would have to be a custom piston, but it is possible to do. Depending on cam choice and how close the dish ends up being to the valves, this design might need valve reliefs and most certainly would become an interference engine. This was all based on information from Larry Widmer both from the forum and from his archived articles. The only thing different is that the dish on my proposed pistons wouldn't be as tight on the intake side so as to reduce the static compression ratio a bit since we are going to be artificially making the density of the charge higher than it normally would be. This is used by Larry on pistons that are designed for FI applications. This is a VERY good read and I think it should be mandatory for anyone looking to explore this type of design: http://www.theoldone.com/articles/The_Soft_Head_1999/
Todd, by "Cosworth" engine, you are talking about the Masi 16V engine, correct (because the head was cast by Cosworth)? If/when I ever need to build a new bottom end for my engine, I am going to employ the strategies of the "Soft Head" design. I believe 100% that the Masi head would be even more incredible if the entire system was used. As a note: the Venolia pistons that are in my Masi engine do mimic the stock piston design, but with a more shallow dish which gives a static compression ratio of 8.5:1 IIRC (could be 8.2). The area outside the dish are at the top of the bore and match the head's chamber shape to provide squish.
Warren Stramer is running his pistons down in the hole by a bit (similar to the NSRT 2.4 Turbo) and when asked about it he responded that he has noticed a current trend for high performance FI engines to not be using squish much, if at all. While I can't argue with his results, knowledge, or skills, I can't agree with that thinking simply because it doesn't make much sense if you look at the "problem" from an engineering physics point of view. Maybe I'm all "starry eyed", and have fallen prey to an elaborate gimmick, but I've seen the results from using this train of thought in practice, and I've had other very experienced engine people agree with it as they'd also messed with portions of the idea and seen results themselves. Also, given that new, high efficiency engines are employing this type of design (which Larry and his company have been contracted by auto makers to develop products for them)...I can't help but to believe this is the way to go.