Does Black-Scholes even beat the market nowadays? In case you're allowed to tell this, which of the publicly available options pricing models do you find the most accurate, computational complexity aside?

If you want to get really good, really the only 4 things that are necessary and sufficient to get you there are:

learn by doing
talk with other good traders
read boring looking books
read research/papers

There are some good resources that you can use to learn and motivate you to think, such as:

seeking alpha
chat with traders podcast
wall street journal
investopedia

Finally, I think there are certain resources with no real 'value' but can still be +ev due to immersing you in whatever world:

BFI
stock twits
twitter
cnbc

Anybody offering you something online to teach you to win at trading is lying, especially if they want any money.

Black-Scholes makes a ton of assumptions, to the point where the world that gets you to that formula is not close to the world we live in. The only thing that BSM does not assume (and the reason they won a nobel prize) is the growth rate of the stock.

One of the biggest assumption is a constant volatility. That means that if I buy an option, the expected realized volatility is the same throughout the option period, regardless of the current stock price and of time.

In reality, during market crashes volatility increases. Prior to 1987, when most people traded directly off of BSM, you could make money just consistently buying out of the money puts, since BSM greatly undervalued how much their value increased during downswings.

curious what you mean by this

Both these responses are good. I am not looking to close individual positions, basically the entire time I've been trading I've ever seen my 'gross risk' increase. Everything is viewed from the portfolio perspective. So while it would be uneconomical to aggressive close individual positions, I'm looking to develop my entire portfolio in a way that makes sense for the risk adjusted returns I am aiming to achieve.

Joe's example could be valid for a single situation. But then what happens if the next 'motivated market player' decides they want to sell the 15s instead of buy. Am I willing to keep trading as long as I think it is +ev? Do I avoid the trade altogether and lose out on $. This gets into what I believe mr. baseball was talking about earlier. You can keep adding exposure to some risk factor but if you're not intelligent about it eventually you get run over.

I've only traded retail options once or twice, so not totally sure what the cost and margin structure is, but I don't see why not.

I do think it's pretty foolish not to be employed if you want to be a successful trader, assuming that profit is your goal. The barriers to entry/success are higher than poker, but working for a top firm means you've already had vast sums spent to set you up for success.

Could you clarify the questions, what does beat the market mean?

Black Scholes, and other option pricing models, give you a price given inputs. Are you asking if black scholes model with the correct inputs is a better representation of reality as compared to the stock price distribution implied by the option prices in the market?

Have you ever been pinned to a strike on expiration? It happened to me once and it was very uncomfortable even though I got out of it alright on the Sunday night open. I had conversions on and I thought I was safe as the strike was pretty distant with 30 minutes to go and then we traded right to the strike on the close. Expiring bond options would close an hour and a half before the actual close so once the move started happening it was too late to do anything. After that I would bail out of any synthetics that were even remote possibilities well before option expiration time.

Sorry for the sloppy wording. I used 'beating the market' in the poker sense, better expressed by the idiom 'generating alpha' in financial parlance.

Let me try reformulating the question: is it even possible to achieve a higher expected (longterm) log-return than the one of the benchmark 'market portfolio' (consisting of all stocks weighted according to their market caps) by trading options only while using the BSM as the only quantitative valuator of options (but possibly also using any qualitative source of guidance in addition, e.g. drawing lines on charts or reading news feeds opened manually)?

From your previous post about pre-1987 options trading, it's clear that either qualitative sources or another model have to be used. I was just worried that, after you advised the BSM as 'a good place to start', some reader would start using it as the only valuator, without realizing why 'the volatility smile' is justified and has to be taken into account.

I know that, in the Heston model, the 'fair' prices for European options are analytically expressible, while the valuation of American options admits a variational formulation that facilitates the computation, however, I'm afraid that Heston-type models might not be precise enough either. Also, I haven't yet studied how well they scale to the multivariate case (the consideration of correlations within a set of many assets).

Personally, I'm not going to trade for large amounts unless and until I fully automate the decision-making process, because I don't want to fall victim to 'trembling hands' (the human psychology factor), therefore I'd spend years designing a cutting-edge quantitative model (possibly incl. expert evaluation of news as one of the inputs) that would generate the biggest Sharpe ratio in backtesting on vast historical data (which is much easier to obtain for crypto than for options, though, considering that the sample size doesn't need to be that big for crypto: it seems that a calendar quarter is about as eventful for the crypto market as a year is for the stock one). I know that it wouldn't guarantee alpha even in the near future, but I think that would be the best shot at its generation.

* To clarify, I was talking about 'the best shot' specifically for a resident of an underdeveloped country with a bad CV (typical of online poker regs) who thus finds it hard to get a decent wage by US standards and/or a market maker discount on transaction fees.

I mean, at least in my personal case, where a net worth of $40K would be enough for a basic (semi-hobo) retirement, poker is still clearly the easiest way of obtaining it. The grind on the track of earning an MSc in quant finance at home and then a PhD somewhere like Canada is too grueling to be a legit hobby. Programming a 'dream machine' to be run on a quantum computer, to exploit the financial markets, where bots are legal and even endorsed to an extent, is a legit hobby because it's self-paced. YMMV.

Options are more challenging mathematically than underlyings, hence my interest in the topic.

Yeah happens all the time and its super annoying. Once the pin starts looking possible it becomes almost impossible to trade out of since nobody wants to buy that strike.

It does happen less over time as I've gotten better at positioning myself to avoid those pins, but its unavoidable. There's a lot more weeklies and strikes than back in the day so even just by chance pins will happen all the time

Yeah with so many equity expirations I suppose it could happen a lot. When I started we only had quarterly expirations then they brought in monthlies. Never had to deal with weekly or almost daily like in SPYs. I was never much of an expiration week trader and spent most of my time longer out but would still have flat synthetic positions going into expiration day. Just tried to make sure I didn't have anything too close to the strike it looked like it would go out at.

Can someone explain what they are talking about? I understand how you can hedge can flatten long options going into expiration. Like if I'm long $250 calls, I can just sell $250 LOC, and be flat if the close is 249.97 or 250.03. But if I'm short $250 puts, there's no easy way for me to sell if the close is below $250 but do nothing if the close is over $250

Isn't it then a +EV strategy for a retail beginner (for the lack of less risky sources of profit) to make the market for ATM options on the expiration day to bail desperate professional MMs out of the pin risk? Or is all the EV of the strategy eaten by the transaction fee and/or the slippage?

My general line of thought is that, if there were no fee, it would usually be +EV for newbs to do any dirty job that pros refuse to do solely because of the high risk.

At our shop, we delta hedge expiring options. However the delta hedge position size is sensitive to spot and (dDelta/dSpot..aka gamma) gets infinitely large if the spot sits at the strike going into expiration, so the rebalancing frequency gets tricky. As you rebalance more frequent, your premium vs realize vol risk is smaller but your transaction cost goes up. Since not practical to continuously delta hedge, you have to decide how much transaction costs you want to pay to preserve the remaining premium.

When it happened to me I had about 75 synthetics on. Lets say the strike price was 100. So for example I was long 75 futures and short 75 100 calls and long 75 100 puts. The price closed at exactly 100. So the question is how many (if any) puts should you exercise and how many of your short calls will be called. It was a long time ago but I think I ended up exposed to about 35 futures contracts. So at any other price other than exactly 100 it's a complete wash as on half of the synthetic washes the futures.

For TBond futures this was very rare and about half of the pit was stuck and everyone was in on the Sunday night session either sweating it out or trying to take advantage of those stuck.

There's 2 factors at play. 1 is that your own trading activities can move the market against you as gamma increases. The other is that with a short option you can't know how many will be exercised, so you may find yourself stuck with a large delta position over the weekend.

And the problem is, as you and I do it simultaneously, we bring down realized vol at the end of the day drastically. It's a prisoners dilemma, as if neither of us hedged the realized vol would be higher, but if only one of us hedges the hedger will outperform

My first week of trading I found a put-call parity arb for $50 EV in under a week. I was mocking EMH on my way to the bank until I realized there was a very good chance of a pin and it was too illiquid to trade out. Luckily the stock moved and I collected my cheeseburgers (probably <3 cheeseburgers out here in SF)

****, going to respond to your posts but theres a lot to unpack so going to wait til I have some time.

@Jcrew, regarding the comparison of scalping and swing trading: though I guess it's unprofitable to scalp for large portions of the portfolio because too much is paid in fees, I don't mind intermittent portfolio rebalancing whenever it increases the expected utility of the portfolio, with the fee taken into account. Such situations are rather rare (because the fee is rather large) but perhaps significantly more frequent than once a week, and it may be the case that it's not optimal to hold most of the positions to maturity.

Therefore, though I agree that a prediction model has to be implemented to determine whether to close an ITM position before it becomes illiquid or to hold it to maturity (or sometimes to execute a deep ITM American-style option before the maturity), I don't find models based on delta hedging useless, as short-term price fluctuations can also be exploited at least sometimes.

Ideally, I'd like to create a universal model that would predict both short- and longterm price movements accurately enough within the available computational resources. Perhaps I'm wishing the impossible.

Sorry, 'slippage' was an incorrect wording. What I meant was that a limit order may be filled just because the market has moved against the maker (e.g. the order is to buy a call, and the mid-market price has fallen fast, and the maker has failed to adjust to the market by lowering the bid) and not because a motivated taker has shown up. Because my education was mathematical and not financial, I don't know the correct term for this effect.

If you can successfully find a way to identify stocks that look like they are about to pin but are actually going to move, you can make money this way.

I doubt that blindly buying ATM options 30 minutes to expiry is +EV.

If you can build a bot to win at options trading 40k/year is incredibly underselling yourself. If you can't you're better off sticking with poker.

I had a pretty bad CV outside of poker. Mediocre grades in a BA, a bit of back office work experience (probably a negative). This isn't I-banking where the goal is to consistently find as many mildly clever people as possible. The bar here is much higher and the rewards for finding a good trader are high, so resume matters less.

If you want, you can send me your resume and I can let you know if you have a shot at pursuing trading for a firm.

You've taken the time to read up on these things but there's a lot where you're way off base. There's a lot of content so I'll try to just hit a few points.

Alpha is generally relative to some benchmark (the market). It's not really appropriate for something like trading which has a vastly different risk/return profile, and which has a lot of time and human capital investment. We have a specialized skill that requires a lot of focus and can only scale so much, and if it was all for 10%/yr with the market's variance it would be a huge waste.

Depends on your inputs, but the strict answer to your question is yes. BSM, particularly if you are using local rather than constant volatility, is the most flexible option pricing model. You can assign any price to any option, so if you assign the right prices you will win.

I think, however, that you are getting ahead of yourself. Modeling volatility is nice and important, but you seem to be under the impression that if you have the best model you will have a higher ev. In reality, no models are actually reflections of reality, and various models have different pros and cons. Some are more flexible, some can fit historical data more accurately, some are more stationary. They are simply tools, and you will not win simply by having the 'best' model.

Calling back to the 1987 comment. I've sometimes thought that if you were a market maker trading flat pre 1987, and you had the foresight to use a vol skew, you'd probably go bust since you wouldn't be able to find both sides of the market. maybe mr baseball can comment.

Since I don't know what your strategy is, I can't tell you what the best model is.

Something that might help you is to realize that BSM says absolutely nothing about the EV of an option. This seems counterintuitive, surely the option price implied by BSM is also the EV according to BSM? In fact, look at the derivation and think what that means about the use of volatility models in trading.

BSM in its original form is not used by any practitioners. I brought it up because its a good place to start, and is the language used by practitioners. Anyone who wants to trade options needs to learn about it at some point.

See previous posts. Maybe think a bit about when/why a heston model might outperform other popular models. What do they miss? What do you gain by precision, and how important is it?

I thought that bitcoin options were still incredibly illiquid and had tremendous counterparty risk. I do expect it to be a lot easier for a sophisticated quant to build a crypto bot than an options bot.

Nice thread.

1) Biggest losing trade/day?

2) Do you think you'd be able to generate alpha trading cash equities? Do options flows provide you with meaningful informational content to predict underlying returns?

3) Earlier you stated: If you aren't able to hedge options with options you're going to have a tough time.
This seems a bit unusual. Outside of ETFs and maybe the top 20-50 names, options are pretty illiquid. And you aren't the only market maker trying to capture the spread. You cannot expect to see enough two-sidedness to naturally hedge your exposures (even at a portfolio level, can you?) quickly enough, so I'd think you'd often be forced into hedging, at least partially, in the cash market, or have some pretty strong risk tolerance/big balance sheet to reload and average down in +EV spots and ride things out until you can get a natural hedge.

I know too well that I have no shot. To quote a wise man:

That my CV is 'bad' was an understatement; the personality flaws are too obvious, especially in person, can't be hidden at an interview.

Besides, my brain is wired for calm analysis but not for fast high-quality decision-making, that's why 1) I'm incapable of day-trading manually at a profit and 2) have to maintain a low poker table count, but entertaining European and Canadian customers at twice fewer tables is yet a more robust way of earning the bread than getting a 'real' job dependent on the declining health of the local economy, for as long as online poker is alive, which will hopefully be long enough for me to earn the remaining $25K within the lifetime to be put into an index fund and secure a minimum retirement (likely with a short expected lifespan because of the lack of motivation to invest into the health).

Having a good model and a good classifier of opponents allows to win at SnGs despite the lack of prior info about the opponents' entire strategies - mersenneary's HUSnG ebook was the one that prompted me to start thinking the Bayesian way - that's why I was expecting a data-driven exploitative approach to work in the financial markets too, because, in a game with so many players, the presence of parties who're not doing their best to make a profit (hedgers and retail newbies) and, on the other hand, hefty fees, would make everyone deviate from the equilibrium even if it were known.

Thanks for your time and answers, sorry for bothering you - I have a bad habit of logging onto 2+2 and exhausting my keyboard with rants when I have too many emotions to deal with. May your trades go well!

Thanks. Won't put any numbers, but have had some fun losses. The biggest ones are all random takeovers.

A couple months back I took a vacation day. Happened to sell a ****ton of ATM calls in APRN (stock at 4). Stock ran up a little bit over the next few days and I sold a bunch of stock. Over the next few weeks the stock just consistently drops 5-10% every day without fail. I don't hedge and before I know it suddenly the stock is trading under $2. Whoops. If I had just sold the calls without stock I would have made 100%!

They definitely do have some, no clue if I'd be +ev trading only stock, never tried and hopefully never will.

You're not going to hedge options only with other options, you're going to need to trade stock as well. But trading stock has transaction costs, and only covers 1 risk factor while ignoring many others and introducing new risks. I don't remember my exact context but it shouldn't be the only tool