Acceptably dumb proof. The method of mean-reversion.

My editorial on You Tube

It is 5:43 a.m. (yes, forty-three minutes after five o’clock in the morning, and I am completely sane), and I am starting another day of fascinating life. I know I could say: another day of this horrible epidemic, or another day of that limiting lockdown. I know I could, yet I am not. I say: fascinating life. This is how I feel. This whole situation, i.e. the pandemic and the resulting lockdown, it all makes my blood flow faster. There is a danger, out there, and there are us, who can face this danger. Us, not just me. There is the collective ‘us’ who adapts, organizes, and collectively says: ‘There is no f**king way we surrender’. This is the beauty of life.

Would I say the same to someone who has just lost their job, due to the lockdown, and has a family to take care of? In spite of all the apparent ridicule of such a claim in such a situation, yes, I would say the same, and you know why? Because there is no viable alternative. Should I say to this person: ‘Yeah, they’re completely right, those people who say you are f**ked. There is nothing you can do, just sink into despair and complain occasionally’.

I am drawing a bottom line under my yesterday’s quick trade, in the Polish stock market. You can read about the details in A day of trade. Learning short positions. I am progressively wrapping my mind around the day of yesterday. Conclusions start floating on the surface of my mind. When I go into a quick, daily trade on short positions, the best moment for making decisions seems to be around 11 – 12 o’clock CET, in the middle of the day. Deciding early in the morning, e.g. starting to trade with a morning sell-out, is not really a good idea. Deciding by the end of the day is tricky, too: the end of the trading day frequently pushes me to selling or buying just out of sheer rush, under the hot breath of time rather than the cool breath of reason.

Recently, a student of mine asked me what I think about short trade. I answered that it is interesting, but it generally sucks for me. It is true that never before have I done any short trade successfully. I remember feeling the pump of adrenaline, peculiar to gambling, yet the financial results were never good. I said it generally sucks for me, and then I tried again, yesterday. This is something I discovered lately: facing my fears and apprehensions can be a fascinating experience. At my age, 52, fears and apprehensions come out of accumulated learning, and the big thing about it is that we accumulate learning in order to stop learning. Facing the things which I am wary or afraid of means questioning my acquired knowledge and habits. It is like digging into one of those cellars, full of objects from the past: I discover new kinds of beauty.

And so, I did it again: I tried short trade, and I meant confronting my acquired wariness. I can see that trading on short, daily positions is a useful skill, and I can develop that skill, to a reasonable level, quite quickly. It is most of all about being aware what I am doing, i.e. cognitively stepping back from action, for a moment, and correct it slightly, so as to make it coherent and purposeful. The key is to own my own story. When I have both cards in my hand: that little gambling nerve, and the intellectual discipline in self-questioning the gambling reflexes, I can thrive on that mix. I love it, actually. My action leads me to forming new ideas about myself. I have just realized that I thrive, as an investor, on two types of action. I can be like a gardener, for one, watching my long-term positions grow and bring fruit in due time. For two, I can be like a hunter, going out for an informed, wise kill.

Wise kill means predating, i.e. violently harvesting from the ecosystem, not killing for the sake of it. There comes an important question I ask myself: how to practice short trade, every now and then, and stay sort of constructive in my investment? I have already learnt, after the day of yesterday, that short trade is a powerful method of quickly adapting my investment to just as quick a change in external conditions.  On the other hand, I want to join, in an informed way, a big stream of investment in positive technological change. Can I reconcile these two: short-term, wisely predatory strategies of adaptation with long-term, positive orientation?

There comes an afterthought, which has just popped up in my mind: wise hunters wait for their prey, instead of running after it. My experience of short trade tells me that it is wise to have a strategy prepared for those days of short positions. To me, short trade is adaptation, and, logically, I should do it in the presence of quickly changing conditions. Just as logically, I should tool myself with some kind of early detection mechanism for violent outbursts in the stock market, when a local speculative bubble is about to swell, or about to implode. Detection in place, I should have strategies for riding a mounting wave, as well as surfing down a collapsing one.

My point is that I can stay constructive in my episodes of short trade when I stay strategic, informed, and prepared. Blueprints seldom work perfectly in real life, yet they provide robust structure. I can become destructive in my days of short trade if I go chaotic, and to the extent of chaos in my actions.

A numerical strategy comes to my mind. I target a handful of companies I would like to sort of hang around with, equity wise. Let’s suppose they are Polish companies from the biotech – medical complex, plus some interesting IT ones. I check regularly their prices in the stock market, as well as the volumes traded. I assume that the market can be in two alternative states, from my point of view: either it allows me to be the placid gardener of my investment positions, or it forces me to become the alert and violent hunter. The ‘gardener’ state is when I don’t need to do anything quick, i.e. when I don’t need to adapt through daily short trade. I need to go for a day of short trade hunting when the market somehow goes off the rails. I need to define those rails.

Mathematically, I assume that whatever happens to those stock prices, happens inside a stochastic process, i.e. something slightly crazy, yet crazy in a generally predictable manner. Within that stochastic process, there is the calm and picturesque Gaussian process, where local values go hardly away from their moving average, like no more than one moving standard deviation away either way (i.e. plus or minus). Anything outside that disciplined Gaussian happening triggers the hunter in me and makes me go short trade. This is an approach similar to mean-reversion: the further something drifts away from the expected state, the more alarming it is.   

I assume that cognitively, the still waters of Gaussian process, from my subjective point of view, are set by the behaviour of prices over the last 30 days. I take the moving average price, and the moving standard deviation from that price, from the 30 preceding days. Below, I am exemplifying this logic with historical prices of the company whose shares I sold yesterday – and I regret having been too hasty – namely Biomed Lublin. The curve in the graph shows values calculated as:

Mean_reversed Price (day xi) = {[Price(day xi)] – [Mean(Price xi-30, …, Price xi)]} / Standard deviation (Price xi-30, …, Price xi).

On the graph, I marked with green dashed lines the corridor of ‘calm’ variance, within one moving standard deviation around the moving average ( -1 ≤ x ≤ 1). Inside that corridor, I assume I can just hold whatever stock of Biomed Lublin I have, or, conversely, I should abstain from buying it, unless I really want. The bubble marked with red dashed line shows an example of price wandering way out of that safe corridor. It is an example of alarm zone: it is price rocketing up, and a possibly good occasion for the short trade I planned, and did not complete finally, for yesterday: selling in the morning for a higher price, and buying back, for a lower one, by the end of the day, or next day. If the curve flares in the opposite direction, i.e. below the bottom green line, it is a signal to buy quickly, with an expectation to sell at a higher price.

The graph shows a time window between May 27th, 2019, and yesterday, April 8th, 2020, thus some 10 months with a small change. During that period, should I have been actively trading Biomed Lublin, I should be about half of the time on alert, and going into short trade. As you can see, this otherwise simple strategy of trading involves behavioural assumptions about myself: do I want to go hunting, in the grounds of short trade, as frequently as the graph suggests? It is reasonable not to narrow down the zone of calm, i.e. below one moving standard deviation away from the moving average. On the other hand, I can increase my zone of tolerance (calm) beyond one moving standard deviation.  

Summing up, I have two perspectives on trading a given stock, with this simple model. First of all, in the long view, I can observe how does the curve of mean-reversed closing price behaves generally. Is it rather wild, i.e. does it swing a lot out of the safety zone between -1 and 1, or, conversely, is it rather tame? The more swinging is the curve, the more the given stock is made for a series of short-term trading operations, like buy in and sell out, in a sequence. If, on the other hand, the mean-reversed price tends to stay a lot in the safety zone, that is the type of stock to hold for a long time rather than to prance around a lot. Secondly, I can observe the short-term tendency over the last few days, like the last week of trading, and make myself an idea as for the immediate stance to take.

I use this simple tool to study my own current portfolio of investment positions, plus the two stocks I sold yesterday but I sort of keep them in my crosshairs, as they are biotech, presently dear to my heart sort of generally. Biomed Lublin, to follow, is a wild one, especially those last weeks. Its mean-reversed price has been swinging a lot out of the – 1 ≤ x ≤ 1 zone. This is the type of stock to watch closely, and to be ready to go for a quick kill about it. As for the last days, you can see it gently returning from a ‘quick sell’ zone, and getting into the ‘hold’ one.

Mean-reversed price of Biomed Lublin

01.04.2020      3,196722673

02.04.2020      3,590790488

03.04.2020      4,173460856

06.04.2020      3,713915308

07.04.2020      1,870944561

08.04.2020      1,71190807

As regards 11Bit, it used to be a wild one, with a high potential for ‘sell’ recommendations. Yet, since the COVID-19 panic erupted in the stock market, and after the Polish stock market started to flirt a lot more with biotech, 11Bit has gone sort of tame. A few weeks ago, there had been a short window for buying, which I missed, unfortunately, like between February 27th and March 27th. The latest developments suggest holding.

Mean-reversed price of 11Bit

01.04.2020      -0,470302222

02.04.2020      -0,418676901

03.04.2020      -0,308375679

06.04.2020      0,518241443

07.04.2020      -0,230731307

08.04.2020      -0,036589487

Asseco Business Solutions is in a different situation. In the past, before the COVID crisis, it would stay a lot above the 1 barrier, thus offering a lot of incentives to sell and consume profits. Yet, over the last month or so, it has nosedived into the alarm zone below -1, just to climb into the -1 ≤ x ≤ 1 safety belt recently. Looks like it morphed from something to kill into something to farm patiently.  

Mean-reversed price of Asseco Business Solutions

01.04.2020      -0,498949327

02.04.2020      -0,454984411

03.04.2020      -0,467396289

06.04.2020      -0,106632042

07.04.2020      -0,062469251

08.04.2020      -0,006786983

Airway Medix is another wild type, with a lot of spikes out of the -1 ≤ x ≤ 1 zone. Still, since May 2019, there was more occasions to buy rather that to sell. Those last weeks, it seems to have really changed its drift, and has rocketed up above 1. I have to be vigilant about this one.

Mean-reversed price of Airway Medix

01.04.2020      2,362791403

02.04.2020      1,922976365

03.04.2020      3,70150467

06.04.2020      3,768162474

07.04.2020      1,973153986

08.04.2020      1,441837178

Biomaxima is a strange case, at least as compared to others. For months, like until the first days of 2020, it had been mostly in the safety zone, with occasional spikes down, below -1, thus with occasional incentives to buy. Since January 2020, it started to sort of punch the ceiling and to burst more and more frequently above 1. Right now, it seems to be in the ‘sell or hold’ zone, with a visible drift down. To watch and react quickly.

Mean-reversed price of Biomaxima

01.04.2020      3,81413095

02.04.2020      3,413908533

03.04.2020      3,001585581

06.04.2020      2,378442856

07.04.2020      1,631660778

08.04.2020      1,668652998

Bioton is a still different story. Over the last 10 months, it had remained like half in the calm zone between – 1 and 1, whilst spending most of the remaining time in the ‘buy’ (x < -1) belt. There was one spike up, in July 2019, when there was some incentive to sell. Yet, now, it is a different story. As it is the case of many Polish biotech companies, the last 2 months have dragged Bioton out of that grey lethargy, into the spotlight of the market. Right now, the mean-reversed price from the last week suggests selling (if I have profit on it) or to hold. Looks like I bought this one on a selling wave: a mistake I could have avoided, had I remembered and applied earlier that method of mean-reversion in price (which I read about regarding the market of electricity).  

Mean-reversed price for Bioton

01.04.2020      1,219809883

02.04.2020      1,50983756

03.04.2020      3,76644111

06.04.2020      3,986920426

07.04.2020      2,434789898

08.04.2020      1,888320575

Mercator Medical is another case where, although I have currently some profit, I should have rather bought earlier (August – September 2019, something like that). That had been a relatively long window of ‘buy’ recommendation. Right now, as I have been investing in it, it is rather the ‘sell or hold’ time.  

Mean-reversed price for Mercator Medical

01.04.2020      1,664368071

02.04.2020      1,605024371

03.04.2020      2,408595698

06.04.2020      3,673484581

07.04.2020      1,846496909

08.04.2020      2,130831881

That cursory, technical analysis of my investment portfolio, together with my immediate targets in the biotech sector, brings me a few interesting insights. First of all, and once again, it pays to do things, and to write about me doing things. The urge I felt to phrase out my feelings after the yesterday’s intense day of short trade pushed me to formalize an acceptably dumb-proof strategy, based on the method of mean-reversion, which I knew theoretically but never thought to apply in practice to my own investment business.

Leave a Reply