Applying hyperbolics to the treasury curve

I simply took the existing rates and their log. Then I subtract one rate from the other. The hyperbolic angle is largest at the short end. I always get this mixed up, but now that I got off my rear and did the work it makes more sense.  The largest hyperbolic angle is pi()/2, so we see that the first three rates are split evenly by some delta of the hyperbolic angle, they have the same liquidity. Then going past the knee of the curve, liquidity changes. But this is clear from looking at the curve, it is linear up to the knee. All this proves is that two period planning is likely the norm and that we change the market liquidity at the knee of the curve. It also trains me to orient myself better when trying to match hyperbolics to any aggregate system, its not clear how to orient the thing until one actually works through a data sample.

Rate	log	diff
0.200	1.349	-0.221
0.610	1.107	-0.242
1.560	0.903	-0.204
2.210	0.828	-0.076
2.980	0.763	-0.065

Getting oriented then.
Hyperbolic angle is large at the short end of the curve, rates lower and loan/deposit close to one.  Loan/deposit decreases at the long end of the curve. But these are the aggregate numbers for the lending market. The problem is when lenders and borrowers individually deviate from the two period model.  DC, for example. It acts like a member banks, but it segments the members banks into itself with all the loans and the real banks that have all the deposits. Then at the long end, DC appears again, with all the loans and most of the deposits held by wealth.  G fouls the whole mess up and causes liquidity crashes, the markets becomes unstable.