Lots of good stuff. IMHO though, there are two fundamental issue at stake, the one following from the other:
1. The auction process was presumably designed to avoid overt manipulation. To manipulate the price requires that the manipulator have an accurate take on precisely when the stocks being used for the manipulation leave the auction. As I understand it, that is not at the end of the auction period but at any time during the auction when certain technical conditions have been met. He can not rely on the SETS price alone because that is determined purely by transaction flow through the order book, whether or not the stock is still 'in the auction'. ie if he bid up the price, then sold it down before the technical conditions were met, he would achieve nothing. He has to bid up the price whilst the stock is in the auction and sell it down ONLY when it leaves. To do that he has to KNOW that the stock has left the auction. The ability of a trading party to ascertain precisely when it has left (or at least estimate with a high degree of confidence) is the achilles heel of the process.
2. The technical spec allows for a 3% discrepancy, which rather suggests that its designers recognised the theoretical potential of manipulation. However, regardless of the spec, a settlement like the one we witnessed in June clearly undermines confidence in the product.