Can someone confirm the correctness of the following formula and interpretation/conclusion regarding "Microprice"? I think there must be a bug in it b/c it's so...hmm... buggy & unintuitive, IMO. Not? Was it maybe written by another Covid victim?... Why not use the obvious & logical formula MP = (Bid * BidSize + Ask * AskSize) / (BidSize + AskSize) and instead use the IMO buggy & illogical variant MP = (Bid * AskSize + Ask * BidSize) / (BidSize + AskSize) ? Source: https://www.quantstart.com/articles/high-frequency-trading-ii-limit-order-book/
No, this is correct. You can do a sanity check yourself: the theoretical price (microprice) should be higher when there's stronger buying interest (larger bid size), and lower when there's stronger selling interest (larger ask size). The ask price is always larger than the bid price, so you want the product of the {bid size with ask price} and {ask size with bid price} to achieve this direction.
So, then let's analyze this situation: Bid=50, BidSize=10000 Ask=75, AskSize=100 (ie. the greedy MM a$$ ) Using "your and their" formula one gets: b=50 ; bs=10000 ; a=75 ; as=100 ; mp = (b * as + a * bs) / (bs + as) ; mp 74.75 Ie. the single AskSize=100 controls the price. Using the logical formula one gets: b=50 ; bs=10000 ; a=75 ; as=100 ; mp = (b * bs + a * as) / (bs + as) ; mp 50.24 Which makes more sense? Normally such big orders usually get filled...
The OP was not really asking a question, but more showing his so called expertise in trading options. There is no way to explain him why the formula is correct and why it's used.
Are you saying, you as a buyer are willing to pay $74.75 whereas the majority thinks it's worth only $50.24 ?