Bigz/adl #209
Bigz/adl #209
Conversation
| let quote_entry = (-market.amm.quote_entry_amount_long) | ||
| .safe_add(market.amm.quote_entry_amount_short)? | ||
| .safe_add( | ||
| loss_to_socialize |
There was a problem hiding this comment.
based on our convo, sounds like this doesnt make sense
| { | ||
| -(mean_entry_basis.safe_sub(user_entry_basis.cast()?)?) | ||
| } else { | ||
| -1 |
There was a problem hiding this comment.
what does -1 represent here?
| }; | ||
|
|
||
| let max_user_payment = if deleverage_user_stats.base_asset_amount == 0 { | ||
| deleverage_user_stats.quote_asset_amount |
There was a problem hiding this comment.
does it matter if the deleverage_user has negative pnl here?
| loss_to_socialize | ||
| .abs() | ||
| .safe_div(market.number_of_users.max(1).cast()?)? | ||
| .max(1) |
There was a problem hiding this comment.
if loss_to_socialize is less than market.number_of_users. and if number_of_users is 0 dont want to divide by 0
| let base_amount = | ||
| (-market.amm.base_asset_amount_short).safe_add(market.amm.base_asset_amount_long)?; | ||
|
|
||
| let cost_basis_above_mean = if deleverage_user_stats.base_asset_amount != 0 { |
There was a problem hiding this comment.
nit: is above right terminology now? because above mean is good or bad depending on long versus shorts
| .amm | ||
| .order_tick_size | ||
| .cast::<i128>()? | ||
| .safe_mul(MAX_FUNDING_SOCIAL_LOSS_MULT)? |
There was a problem hiding this comment.
how'd you decide this? if bitcoin tick size $.01, then a user would only lose $1 per base?
There was a problem hiding this comment.
feel like ticks is the unit of "how much amount per base does the market care about". picking the constant across all markets is tbd, but .01 * 100 / 30 => 3% (which i'd argue might be too high). so a $30 asset would ideally have a perp tick of around .001
|
|
||
| let mean_long_pnl_per_base = mean_long_entry_basis | ||
| .safe_add(oracle_price.cast()?)? | ||
| .safe_div(PRICE_TO_QUOTE_PRECISION_RATIO.cast()?)?; |
There was a problem hiding this comment.
this made me think this ends up in quote precision, got kinda confused. might be better to pick a better constant
There was a problem hiding this comment.
this is in quote precision though, "pnl" per base
No description provided.