-- we were supplied numbers by representatives from doimp bank, they are in our report -- from deutchbank. the numbers are deeply worrysome. and the question of the ad casey of capital reserve, looking at potential losses into the future, it really does stay, the stress test would be stronger and more meaningful if they reached across a longer period of time and dealt with this issue. the third problem of -- that we are concerned about, is the inability to replicate the tests. i'm afraid i'm going to expose once again my academic background, i do a lot of impeerk cal work, and one of the most important things you want to do in any model is establish that it is robust. and what we mean by "robust" is just, if something changed slightly, you altered the time period a little bit, the g.d.p. crags or growth shifted just slightly, you have relatively similar results. different results but within the same beark. -- ballpark. if you do, you have a lot of confidence that the test is measuring something real, that is that it is robust. the problem is, you can't rerun those tests unless you have a