Markets plunge amid coronavirus fears

TheEconomist said:

It's not obvious how fast trading algorithms interacting with one another will influence prices over time. Nonlinear dynamic systems can behave in shocking ways, so you're certainly not alone harboring that fear.

Click to expand...

Just to be clear, which I’m fairly certain you understand, I don’t want to mislead anyone into thinking algorithms are flawed from a mathematical standpoint, but that they are only as good, fair, useful as the recipe from the human who creates them.

TheEconomist said:

With that being said, most of my background is at a much higher level of aggregation and abstraction. I'm not very familiar with market microstructure to be entirely honest. How much do you know about the nuts and bolts of these algorithms?

Click to expand...

It’s not that I have any specific knowledge into any proprietary HFT algorithms, just that I have knowledge of what components are required to create these algorithms. I was raised in the investment world by my father, but was educated and practiced my earlier years as a pure mathematician focusing primarily on combinatorics. My interests and what I have pursued since then I rather not share on such a vile and currish forum.

TheEconomist said:

What kind of research did you do to determine it was a matter of algorithms running amok?

Click to expand...

Just from reading academic reports and evaluating whose explanation seemed more plausible as well as implementing some of my own thoughts into the matter. For whatever reasons or info they may be reacting too, algorithmic traders seemed to have set their commands(unknown to each other), especially on sell and stopsell orders at levels where they created a massive grouping near each other, which in turn forced out the algorithmic marker makers from the market, thus increasing order flow toxicity to a level where liquidity dries up, creating a sort of no liquidity zone where price free falls. There have been many of these instances, although not at such a level. The SEC fantasize their regulations may limit it, but we haven’t seen a bear market like 1929-1932 with computerized trading and don’t really know how trading algorithms react with each other(as you noted) which may effect the speeds of price action when markets enter extreme turmoil.

I think I read a report by Clarivate Analytics(but could be mistaken on who wrote it) whose explanation I determined to be plausible. And I followed the Commodity Futures Trading Commission and the Securities and Exchange Commission’s report of that day and didn’t buy their single order explanation. Since that day, a Flow Toxicity metric has been created to give realtime estimates for which liquidity is provided, which is a better response than SEC regulations, IMO, but it remains to be seen if and how many HFT’s integrate it into their algorithms.

Anyway, none of this affects swing or long term traders, so it’s not important from my standpoint.

Buck naked said:

Just to be clear, which I’m fairly certain you understand, I don’t want to mislead anyone into thinking algorithms are flawed from a mathematical standpoint, but that they are only as good, fair, useful as the recipe from the human who creates them.

Click to expand...

In essence, some tools are designed to solve the wrong problem.

Buck naked said:

It’s not that I have any specific knowledge into any proprietary HFT algorithms, just that I have knowledge of what components are required to create these algorithms. I was raised in the investment world by my father, but was educated and practiced my earlier years as a pure mathematician focusing primarily on combinatorics. My interests and what I have pursued since then I rather not share on such a vile and currish forum.

Click to expand...

I would not have guessed that you're a mathematician. Apparently, we were both under some misapprehension.

Buck naked said:

Just from reading academic reports and evaluating whose explanation seemed more plausible as well as implementing some of my own thoughts into the matter. For whatever reasons or info they may be reacting too, algorithmic traders seemed to have set their commands(unknown to each other), especially on sell and stopsell orders at levels where they created a massive grouping near each other, which in turn forced out the algorithmic marker makers from the market, thus increasing order flow toxicity to a level where liquidity dries up, creating a sort of no liquidity zone where price free falls. There have been many of these instances, although not at such a level. The SEC fantasize their regulations may limit it, but we haven’t seen a bear market like 1929-1932 with computerized trading and don’t really know how trading algorithms react with each other(as you noted) which may effect the speeds of price action when markets enter extreme turmoil.

I think I read a report by Clarivate Analytics(but could be mistaken on who wrote it) whose explanation I determined to be plausible. And I followed the Commodity Futures Trading Commission and the Securities and Exchange Commission’s report of that day and didn’t buy their single order explanation. Since that day, a Flow Toxicity metric has been created to give realtime estimates for which liquidity is provided, which is a better response than SEC regulations, IMO, but it remains to be seen if and how many HFT’s integrate it into their algorithms.

Click to expand...

I read similar things from various sources, but I didn't know about the flow toxicity metric. That might be useful for something later on.

Buck naked said:

Anyway, none of this affects swing or long term traders, so it’s not important from my standpoint.

Click to expand...

Yeah, what happens over nano seconds isn't your cup of tea if you're thinking in terms of quarters and years.

TheEconomist said:

When I say that returns are almost pure noise in those markets, I mean something specific that isn't inconsistent with the point you are trying to make.

If you run backtests for a wide array of forecasting models for stock market returns, you will generally conclude two things: (1) it is almost impossible to beat a constant for point forecast which maps into a random walk model of prices; (2) when you do find some superior method, it generally has a laughably small forecasting power. It might explain less than 4-5% of variance outside out-of-sample, unless you cheat (for example, by using the final values for macroeconomic data as opposed to using real time vintages). Then you turn around and try to build a trading strategy to capitalize on this. Chances are, as soon as you take into account the fact that trading isn't free, that good trading opportunities might be rare, that it might force you to sit on cash, etc., you will find that it's just better to buy and hold simple portfolios that expose you to well known risk factors. It's also the case, by the way, that if you look at active fund managers, you almost always find that they aren't worth their fees...

That is specifically what I mean when I say that stock market returns is almost pure noise. It means the efficient market hypothesis is like 99% true. Now, does it mean that what you say is wrong? Well, maybe you ran into something useful. In practice, the dynamic behind the results I describe doesn't magically come into existence: you need people trying to capitalize on new information or new insights to force prices to kill off the opportunities they offer. I am aware that you have something odd in mind that isn't necessarily easy to grasp and it seems to rely on your own experience. If I had to bet on unexploited or less exploited signals, it would be something you'd have trouble approximating with obvious mathematical models or ideas that people have long forgotten and do not teach either in business school or in popular trading books. In other words, maybe you're right, but I have no clue how to check that. I'm a scientist. If I cannot test it systematically, I have no reason to fall on either side of the discussion.

Click to expand...

I guess I’d agree with the above... but.... explain hedge funds.

Are they all just insider info like Billions?

Threegoofs said:

I guess I’d agree with the above... but.... explain hedge funds. Are they all just insider info like Billions?

Click to expand...

In the academic literature, active investment management strategies have been extensively investigated and the results tend to be very disppointing. Let me be a little more specific.

The simplest way we think of returns in financial markets is as a compensation for exposing yourself to risk. For example, you can think of the cross-section of returns as reflecting a weighted sum of exposures, or risk factors if you prefer. Some of the earliest papers on the subject such the French and Fama 1993 paper or the Carhart 1997 paper proxied their factors using cleverly designed portfolios with each being ways to expose yourself more to the market portfolio, to small caps, to momentum, etc. The idea is that you only get systematically rewarded for that so, on average, if the model was exactly right there wouldn't be anything on top of these things. If you ran a regression, the intercept should be statistically indifferent from zero. We call this intercept the alpha of a trading strategy.

The first thing you do when someone gets excited about an active strategy is... drum roll... check if the alpha is big. If the alpha is big, it can mean one of two things: (1) you do have abnormally high returns; (2) you found a way to exposure yourself to sources of risk we didn't put in the model. So, my first point here is that you can explain away nearly all of the high performances of active managers as rewards for risk. In other words, it seems like the 'exceptional' returns aren't that exceptional. Another thing you can try to look into is if there is some persistence in manager performance. Usually, that's also very disappointing: typically, performance isn't that persistent.


Now, your problem here is that hedge funds manage a lot of capital, but so far I have only said things suggesting it's a huge scam. One possible reply to the first point I rose about small alphas comes from John Cochrane. He had a discussion with some fund manager and he told him he could replicate his return with a combination of market portfolio, momentum, carry trade, etc. The guy stopped him and told him: "Well, that's my alpha. What I add is that I know how to do that without loosing my shirt while Joe Average doesn't." In essence, if you want exposure to a large number of sources of risk, things might get complicated enough to justify having people doing this. It's possible that some people are worth their salt.

Another aspect of what confuses people about this is that they seldom appreciate how noisy is financial data. A track record means almost nothing. In fact, there are so many people trying to trade their ways into great wealth that, by sheer luck of the draw, you're almost bound to find at least a few people who line up a long sequence of favorable trades. I'm not sure it's all that easy to figure out you're not getting your money's worth in this environment. I mean, it's not like a chef where it takes a handful of dishes to know if they can do their job right and it's so odd versus what people experience everywhere else that there likely is some of that happening too. But at that point, asking "why do investment firms" of all stripes exist, involves quite a bit of speculation. Take all of this with a grain of salt, or as food for thought.

Wall Street tumbles as virus fears hit California - Reuters

Wall Street tumbles as virus fears hit California

Reuters) - U.S. stock indexes fell sharply on Thursday as the swift spread of the coronavirus in the United States led California to declare an emergency, while airline stocks were hammered by crippled travel demand.

The S&P 500, which fell almost 12% last week, its worst since the 2008 financial crisis, had recovered some poise as Joe Biden’s surge in the Democratic primaries distracted traders from the widening impact of the virus.
==================================================
Great news for the Dems: a raging plague & a plunging market will affect Trump's reelection hopes.

JacksinPA said:

Great news for the Dems: a raging plague & a plunging market will affect Trump's reelection hopes.

Click to expand...

I don't think for a nanosecond "a raging plague & a plunging market", is "great news" for anyone, it's terrible news for everyone. I think Trump is highly incompetent and makes decisions completely based on his ego and his own interests, but do not believe in any form that he's bigger than "a raging plague & a plunging market".

I also think whomever is the next President will be a one term President as the market will be disastrous during that period of time as well and as I'm forecasting in another thread.