model: Adapt to last version and use real internal AMM

[bingen]

Results I’m getting now:

Essentially:

• The system works nice for the protocol, but most users have losses. So params need to be tweaked.
• When the market price approaches the floor (redemption price) too much, the system “halts” (around iteration 616). When that happens, bonding is not profitable anymore.
• Before the “halt”, bonding and LP’ing into the internal AMM are very profitable.
• Profitability of bTKN (buying and hodling) seems too bad
• The simulation relies too much in the fair price belief: users arbitrage constantly to converge to it. This may be too optimistic. Maybe we should then use a more pessimistic formula for the fair price.
• It would be interesting to show internal AMM size.

Also there may be bugs in the new changes.

[bingen]

Final state of the system:

Chicken Bonds state
 - Pending LQTY:               36,198.69
 - Reserve LQTY:               649,956.49
 - Permanent (DEX) LQTY value: 291,380.52
 - bLQTY supply:               506,621.27
 - Backing ratio:              1.28

Fair price:      2.30
Accrual param:   0.01
Rebond Time:     1,440.00
Chicken in Time: 1,440.00

AMM pool:

 - LQTY amount: 145,690.26
 - ETH amount: 145,690.26
 - Price LQTY/ETH: 1.00
 - Price ETH/LQTY: 1.00
 - LP tokens total supply 145,690.26

bTKN AMM pool:

 - LQTY amount: 27,098.96
 - bLQTY amount: 21,001.67
 - Price LQTY/bLQTY: 0.77
 - Price bLQTY/LQTY: 1.29
 - LP tokens total supply 27,506.32
 - LQTY Fees: 1,431.23
 - bLQTY Fees: 745.77
 - AMM APR: 0.000%

[danielattilasimon]

I've yet to check out the code, but here are some quick thoughts.

Would be interesting to see the permanent bucket, too. Or is that the combination of AMM TOKEN & AMM ETH? Is the yield of that rolled back into reserve? (How much is that)?

What do the results look like with a smaller premium (might have to reduce chicken-in fee too)? 100% seems a bit high given redemption price only grows ~30% over 4 years.

For tuning the parameters of LUSD Chicken Bonds, we can probably be more optimistic about the pending & reserve yield.

[bingen]

Would be interesting to see the permanent bucket, too. Or is that the combination of AMM TOKEN & AMM ETH?

Exactly, it’s the column in red and yellow in the first chart.

Is the yield of that rolled back into reserve? (How much is that)?

It’s defined by AMM_YIELD constant (in the run above, 2%) and you can see it rolled back into reserve in distribute_yield function:

        # AMM generated yield
        generated_yield = self.get_yield_amount(chicken.amm.get_value_in_token_A(), self.amm_yield)

        chicken.token.mint(chicken.reserve_account, generated_yield)