Comment on page

# Price Impact by Size

Premia pools disincentivize highly disruptive trades through size price impact.

After every transaction, the pool price level updates from

$C_t \rarr C_{t+1}$

, depending on the size and the direction of the trade. This results in either an increased or decreased price for the next buyer/underwriter. There are no obvious reasons to disincentivize larger blocks of provided liquidity (on the LP side), however, whale-buying behavior needs to be accounted for.Suppose a whale is waiting on the sidelines for the C-level to fall below their perceived market equilibrium, just to scoop up 50% of the pool's liquidity. Such a trade would cause a significant pool price level disruption; this disruption needs to be accounted for in the price charged to the whale. If the starting value is

$C_t$

, and the ending value (post whale trade) becomes $C_{t+1}$

, s.t. $C_{t+1} > C_{t}$

, what price impact penalty should be imposed on the whale?In discrete form, the whale would end up paying

$BS(V_i)* \frac{(C_t+C_{t+1})}{2}$

, however the differential form is slightly more accurate:**Note:**

$x_t=\frac{(S_{t+1}-S_t)}{max(S_{t+1};S_t)}$

or more intuitively - the normalized step size, relative to the free capital in the pool.Putting it all together, using

$C^*_t=C_t \text{ adjusted for slippage}$

and $\alpha=1.0$

as a potential future trade-specific steepness modifier, we get a final pricing function:$P_{t}(V_i;C_t)=BS(V_i)*C^*_t \\ s.t.\hspace{0.25cm}C^*_t=C_t*\int^0_{x_t}e^{-x} \alpha_x*(\frac{1}{0-x_t})$

**This ensures large traders have no advantage over smaller traders.**

Last modified 4mo ago