Yes all asset classes and in practically every market… this has been written about extensively for the past 50+ years in financial journals. You may have heard of Gene Fama’s work in the 60s and 70s writing about efficient markets, which was around when option theory was being formalized culminating in Black-Scholes.
TA question often asked... "Do trendlines work"?
- Thread starter Scataphagos
- Start date
Yes of course. Will check this out. But I'm so curious how this is possible? Like if I discover an anomaly on histiric price data assuming minimal bias in the analysis, then isn't that the proof that the market doesn't price in your insite? We're probably not going to solve this here, but it raises a lot of questions to me. I'm always skeptical of Fama. I certainly respect the accomplishments, but he does propose some pretty poor statistical analysis, which today has become financial cannon nonetheless. Thanks for the mind expansion.
But see I feel this is only true for long term horizons, long only, institutional players. TA should be able to succeed in any market especially over shorter frequency data going long/short. And FA doesn't really have that quality, unless there's something I don't know about (which is actually quite a lot). You can of course short with FA, but you can't scale down to 15min data.
Well of course he’s backwards now, he wrote that stuff 50 years ago… it might behoove you to conduct a literature review before arriving at some sort of conclusion.
Sigh. There’s a lot for you to learn about market structure, market participants, and how they optimize decisions given constraints and objectives.
The “fundamentals” for a market maker employing an HFT model (and they’re all employing an HFT model), is to solve for a match within the b/a spread.
For example, if you’re a citadel securities trader, your system is likely aimed at crossing internal TD Ameritrade orders before exchange-based market makers can change their b/a spread. This is a tech stack and logistics problem. Is citadel securities trying to profit from the direction of the stock? No. Do they have an impact on liquidity anyway? Yes.
“Fundamentals” vs “technicals” is largely a retail gimmick debate where both sides are largely uneducated. Stick with the terms I suggested (discretionary vs. systematic). Citadel securities uses a systematic strategy. They’re analyzing volumes and tick data across venues internally and vs. benchmark of price (displayed liquidity and b/a). Because they’re not betting on the direction of price, they’re not doing a valuation analysis or reading a 10-k lol. Their work is equally as hard though but it mainly involves improving their tech stack, updating their model parameters, working out bugs, and managing their spread to benchmark prices.
For mid-frequency strategies (intraday), which tend to be dominated by stats arb-esque styles, they’re looking to benefit from the drift in a basket of stocks vs. say an ETF or mutual fund. If a large wealth management firm rebalances their portfolio by selling IWM and adding to SPY, there can be moments where the underlying basket of stocks within those ETFs are different from the prices implied by the ETF price. These fall under “systematic” strategies because your fundamentals are primarily technical data (volumes and prices vs. benchmark or index). So these guys aren’t reading 10-Ks either.
Reading 10-Ks or conducting “fundamental equity research” is mainly the purview of teams trading stocks (or bonds, fx, etc.) at a mid to longer frequency (intraday to days to weeks to years). You can also run systematic strategies at that interval, and those mainly revolve around factors (beta, momentum, value, etc.) or liquidity provisioning (off-menu, on-menu).
Last edited:
They still publish the FF model daily on their website and all those regressions fail diagnostics and always have from the beginning. Like I said, I'm not an asshole, I respect the guy and his accomplishments, but you gotta call out what's right and Famma definitely at times makes assumptions that serve the model over the data and I was trained to point that out and improve on it if possible. I mean at the time, there weren't snot nosed f**kers sitting around on their laptops at home who could actually look into this stuff, so it was the best to go on, but if a model does not fit the data, you gotta recognize that in an age when data and computing give us better options.Well of course he’s backwards now, he wrote that stuff 50 years ago… it might behoove you to conduct a literature review before arriving at some sort of conclusion.
Now the efficiency of the asymmetrical return thing, I have no opinion, just pointing out that I enter skeptically.
Yeah no one really uses the FF model at the institutional level, but it’s a good starting point. The underlying premise is that the price of a stock reflects known information about a stock. The distribution of performance is a part of this, and is also incorporated into price. The primary mechanism for this is through the options market, where if someone is betting on a stock at X strike, the dealer (who sells the option) will hedge their delta dynamically, impacting price. So as the probability of a stock going from $50 to $70 increases, the price does too. But this is observed in all markets afaik, so it doesn’t need an options market to necessarily work.
Not sure what the sigh is about cause I actually 100% agree with everything you said and I thought my comment reflected that, so maybe I need to learn about communicating my ideas better? I was just addressing how when people point out the kind of stats like only 5% of stocks go up and its impossible to predict, etc etc (Random Walk Down Wallstreet type stuff), it just totally discounts all the different ways to profit from the market. And I use TA just to describe any mathy quantitative algo driven style that doesn't require taking an opinion on fundamentals. So that would include the HFT guys. Anyways, we agree. But I definitely will pick up the new vocabulary. Like it much better, thanks.
Cheers.
For what it's worth, most professionals view the market as semi-strong form efficient. It's just a framework though -- it allows one to focus their time and effort in finding anomalies to market efficiencies. There are probably thousands of documented anomalies (many studies are stale or don't replicate).
See this makes a lot of sense to me in options, like you couldn't profit from the asymmetrical return in options cause the price of the option would make it efficient, but I'm not quite there yet with the underlying. If the anomaly is showing up historically. I'm gonna check out the original and I"ll come back with all my complaints.
