I know. We invented wheels, navigated with stars, shot arrows, flew kites, shot gun powder rockets way before we knew Newton's Laws. On the other hand, Newton's Laws allow us to try new things without actually building them. If I have a better option pricing model, I can simulate new trading strategy without actually trading with real money. Am I making any sense?
I didn’t really understand this sentence. Idk what a “better” pricing model iS?? Better in what manner? Also what type of model? Option pricing model is a broad term, that’s like saying BMW, but idk if you mean the 3,4,5,6,7 series? M class? 2 series? Also you can simulate trades via paper trading so not sure what you mean there either mr chef
I was hoping the power law could derive IV skew and IV term structure so I don't need historical option data to study strategies, paper trade without actual data like flying an airplane in a flight simulator.
Option price depend on many variables. Non directional trading looks like this to me: Find several patterns in the gazillions of variables like IV or delta, earmark a play/trade that works this specific variable as it typically behaves, and stack these patterns and plays, create strict quantitative (time or greek or whatever based) entry and exit rules that don't change and backtest as far back as you can. Take the loss when the pattern doesn't play out, and never look back. There are a lot of patterns in the options world, I like the patterns based on volatility most, IV is a variable I think it is easy to incorporate into a strategy. When I say only one side can lose at a time, I assume that you close both sides when one side gets tested, as you might when playing a binary event. Also, you can use ToS to backtest up to 2010, second by second, if you're willing to use their UI and pull all the info you want out manually. I did You have to be able to say, the future value of these trades is positive over 50+ instances, average gain is higher than average loss over the instances and know that after commissions and fees, assuming you get filled at NBBO every time, you still make more than the 15% a year you could make on QQQ and have a low enough standard deviation and good enough risk management to not blow up an account with a bad series of un-probable events. Sharpe ratio of 2+, but you have to have enough data to calculate that and have it be valid You're too hard on yourself I think, we shouldn't take ourselves so seriously. Pretty much all amateurs to someone.
You understand wrong. Buying order flow is primarily about getting priority access to uninformed players, or in poker terms, these guys are paying to play at the children's table. The idea is as follows - you get an order routed to you and you have a right but not an obligation to fill it. If you are fast and sophisticated enough, you can lean on that order to get advantage over other market participants. That's why limit orders are more valuable from PFOF perspective (inherent optionality) and complex order are even more valuable. Options are better from a PFOF perspective since you have a relatively wide spread and most retail traders are willing to post non-directed limit orders (as opposed to directed orders with a delta reference). As a result, the retail guy on the other side of the PFOF trade will get negatively selected more frequently, but I doubt anyone of them ever performed any sort of TCA on their orders/fills.
Interesting. What do you mean by negatively selected? I have done meager post trade TCA, but what can I do as a child trying to stand on books to see over the horizon of my abusive stepfather's poker table? The fill is the fill. You're saying I should peg to stock and leg into complex orders?
You're saying that there is more information in a limit order than a market order, because a limit order creates an order driven market, so an exchange when they receive a limit order now has more information on that side of the market, and they can use it to squeeze the other side and vice versa, working the spread and running a more arbitrage-ish operation
Negatively selected - i.e. you getting a desired outcome but the longer-term result will be "statistically bad". Let me offer you an analogy. Imagine that you go to a club and a girl starts hitting on you. It's unusual, but you think of yourself as a proper stud and take in stride. You buy her drinks, she takes you home. You have wild sex and wake up the whole neighborhood. Sweet! Next day you got back and get picked up by a different young lady. Again, drinks, snacks and you head to her place. However, instead of fornicating, you wake up in a ditch, with a headache and without your wallet. You write it off as bad luck and go to the same club the next day. This time around, yet another girl but the same sad outcome as the second day - ditch, headache and missing wallet. You start realizing that "money for nothin' get chicks for free" actually comes a cost, i.e. sometimes you do get to that coveted third base but at cost of potentially losing your wallet 2/3rds of the time. By going to that club and responding to these chicks (i.e. desired short term outcome), you are getting negatively selected (getting robbed as a longer term result). In case of PFOF, when you get filled (at NBBO or with improvement), you'd find that the price will frequently move against you right away. Whenever you miss a trade, you'll find that the price would have moved in your favor right after. Just like in the club, you are getting fucked, minus the condoms. Well, you should look at the price of your fill and compare it to the price some epsilon later. You'll find that you get picked off. Similarly, if you see yourself not getting filled at your limit even though it trades there or through elsewhere, see what the price would have been an epsilon later - that's opportunity cost. Do that many times and you'll find that despite it being free, you are indirectly paying for it.