Sergey Kryvorotchenko said:
-snip-
Your Reward depends on your losses. If your comrades lose most of Units, they will receive a higher Reward.
My experience with this game (and casinos) is that larger losses lead to larger losses, not better pay offs.
The game mechanics on the locations differ slightly from the game mechanics on slot machines, but it is similar in concept:
You "place a bet" and hope to get rewarded. From time to time you "win" and that keeps you going back to play some more.
When you place small bets, you tend to win small when you win.
You can move up to "bigger machines" and experience larger wins, but also experience larger losses.
Where the analogy breaks down is that every pull on a slot machine is "an even pull" meaning that previous bets and results have no impact on the current pull. In this game, there does appear to be some tracking, and the Field Manual even states that there is.
However, what the Field manual states is that a lack of payout will be carried over until a payout occurs. What exactly triggers the pay out, and how much the pay out is, is something not specified. Any statements about how and when the payout happens, and how much it is, is conjecture unless released by the developers. Some of the conventional wisdom on the topic does not hold up to testing. If the data doesn't support a theory, then the theory must be rejected.
For example:
I can get a pay out of X, and based upon various comments expect to have my next payout by the time I lose 3x troops. (Some use smaller numbers than 3x). In practice: If I send in 10x the troops I was last rewarded with, and lose close to 100%, I will frequently not realize my next payout. When taking very high loss percentages, I have seen a break from the correlation between cost and reward.