I mean if someone is putting thousands of BTC in (which some are) then how hard is to figure out that rather than put it all in at X% (where X% is the lowest rate they'd prefer to risk at rather than not bear the risk) they should put all except 100 BTC in at X% and the last 100 BTX at 100% or 1000% or whatever - to guarantee getting the whole pot when there's insufficient capital offered for the rate to be capped.
Let's see, how would that work? Guy A with 5000 BTC, guy B with 5000 BTC, guy C with 5000 BTC. Bonds structure:
4900 BTC @ 0%
4900 BTC @ 0%
4900 BTC @ 0%
100 BTC @ 5%
100 BTC @ 10%
100 BTC @ 15%
Now f(BTC) = % does something like this: (0 , 14700] -> 0%; (14701, 14800] -> 5%; (14801, 14900] -> 10%; (14901, inf) -> 15%. That make any sense to you?
That's why I was surprised at absolutely no cap on rate - as when one individual can submit different bids it's very easy to game (yet noone seems to have bothered doing it).
I'm not so sure it can be gamed at all, but if you don't feel like explaining it theoretically you can always do a demonstration I guess.
Lets take your example where there's 3 guys A,B and C all putting in 5k.
At the moment (i.e. not doing what I propose) their bids would have been:
A = 5000 @ X
B = 5000 @ Y
C = 5000 @ Z
Where X,Y and Z represent whatever values they otherwise believe are best for them to bid. I'd NEVER suggest bidding at 0% - as the investment is NOT risk-free. What X,Y and Z are (or should be) is a different discussion - but in general it should be the lowest rate at which you'd prefer your capital to be risked rather than unused.
If A was to do this on his own then instead of bidding 5000 @ X he'd make 2 bids:
4900 @ X
100 @ 500% (or any rate large enough to ensure grabbing all profits)
If the demand for capital doesn't exceed available bond capital then only 4900 of his capital would be used and his profit (or loss) would be 98% of what he'd have got had he just bid all 5000 @ X. So the WORST outcome of this scenario is losing 2% of profits.
If demand for capital exceeds available bond capital then the bonds get ALL the capital. For this to be profitable for A, the gains when insufficient capital is offered need to exceed the 2% of profits he loses when sufficient capital is offered.
But now consider if A, B and C collude to a very limited extent - by combining to make the 100 bid. Then the bids look like:
A = 4966 @ X
B = 4966 @ Y
C = 4966 @ Z
(A+B+C) 100 @ 400%
Each of them now only has to throw in 33.3333 BTC for the high bid - ensuring they all get the lot if insufficient capital is offered. That only costs each of them under 1% of profits when sufficient capital is offered in return for guaranteeing a scoop when when there's a shortfall. Does the benefit from this more than cover that? Well take a look at the last 2 months and you tell me? Remember - I suggested this BEFORE this month. How would the real equivalents of A,B and C have done this month had they done something like this? How many months of giving up under 1% (or 2% if done solo) of profits does that cover?