Paul,

I am not sure I would jump to the obvious answer in weighting the sales. Not enough info, yet. But if there is nothing else about these properties we need to know, I can prove you are right. You say, sale 1 because of expenses looks like the best comp and comes to $1.1 mil for subject.

Part of the point, I was trying to get to is the "circularilty" of certain calculations. We are dealing with different ratios, but they involve the same numbers, so yu can get to the bottom line. For example, how can you know these GIM's and expense ratios and not know the cap rate. Here's what I get.

Code:

```
Sal 1 Sal 2 Sal 3 Sal 4
9.9% 9.7% 9.7% 9.1%
```

Proof:

Sale 1, assume EGI = $1 (so that numbers are per "cent"), then NOI must be, $0.71, and the purchase price must be $7.16. Therefore, the cap rate must be .71/7.16 or 9.9%, etc.

This is what I was indicating before. The cap rates give values close together because there is a sane market, but the GIM's can be all over the place. Given the magnitude of variance of the GIM's versus the magnitude of variance of the overall rates, I would conclude that GIM variance makes it unreliable.

Based on your numbers of $1.1 mil at 7.16 EGIM, I get a subject EGI of 153.6k and for $1.6 mil at 10.39 EGIM, I get subject EGI of 153.9k. And using your 29% expense for subject, I get $109,000 for subject NOI. That's $1.1 mil to $1.2 mil (capped at 9.9% and capped at 9.1% from sale 1).

Unless I screwed up, that proves us both right. Your pointed to sale 1 as similar by expenses and its EGIM came to 1.1mil. I said that GIM and overal call would produce the same number, and it did, 1.1mil.

This is why I argue GIM is almost useless. If the expenses are the same, GIM and the overall cap rate give the same number. If the expenses are not the same from sale to sale to subject, GIM produces a lot of noise.