Neurons are Cells After All…

As Tim Cook has said:

Anything can change, because the smart phone revolution is still in the early stages

Even as technology increasingly becomes an extension of man’s central nervous system, the only battery we remember to charge these days belongs to our cellphones!

Advertisements

Abundance!!

 

passion skill job

Kelly Formula (Part 1)

Most of my smart winners have been small capital bets while the losers have seen larger bets of capital. One of the reasons for this could be that catching a high beta asset at its low point (about to breakout; Lower Bollinger  touch; RSI < 20%, whatever) has given a higher alpha to me since the price move (due to high beta) has been swift and definitive. The low beta stocks, giving the illusion of secure staidness, have killed often by bleeding a part of my portfolio to death. Since these are the blue chip, low beta shares, I have been tempted to invest a larger proportion of my capital as compared to the mid/small cap sprightly upstarts being lured by the illusion of safety. As I look into my trading log, it is the multiple smaller bets that have really been multi baggers for me. So, I spent a good part of the long weekend reading on optimal bet sizes!

It is a common belief that a high amount of beginning traders do better with paper trading than once actually live. The reason being the realization of losing ‘real’ money. Knowing how to deal with profits doesn’t make one a successful trader insomuch as being able to handle losses. Every trade can have five potential outcomes – big profits, small profits, par, small losses and big losses. Taking the big losses out of the picture will logically give any trader less chances of frowning. Big losses can come due to decision paralysis (and waiting in the hope that the price will recover) or large capital outlay. The topic of this post is to ponder on understanding what should be the right size of trades. While its natural to think that the size of trades depends on a) your personal strengths and weaknesses, b) amount of capital, b) trust in an investment thesis, c) degree of diversification etc but like all good things there is a formula for this too! Namely, the Kelly Formula. This was designed by John Kelly who worked for AT&T and devised this for use in long distance communication – signal loss; signal to noise ratio and all those nice things. Since there are two basic components to the formula: win probability and win/loss ratio, this found application in the world of trading as well. The Kelly Formula states that for any given stock, you should invest the probability of winning times the payoff minus the probability of losing divided by the payoff. It is represented by the equation:

 where, P = payoff (i.e. odds offered by the betting syndicate); W = probability of winning and L = probability of losing

 

So while the debate on chance vs. skill in investing continues, here are some of my investigations and thoughts on whatever I have read and managed to understand so far. At the core of all this line of enquiry is a desire to figure out a personal method to ensure that I statistically end up putting more money on my winning ideas as compared to my duds. If I get to do that I am pretty sure, I would  be all :)s.

So back to the ruminations on the Kelly Formula: If you are worth 1 crore (say) then it’s clearly stupid to be risk averse for small amounts like INR 5,000 (say). You should regard “gaining 5,000” and “losing 5,000” as equal-but-opposite faces of the same coin. But it’s very rational to be risk averse for 50 lakhs. Whatever I have seen and heard from people with modest wealth around me suggests that people in fact behave in a predictably irrational manner when faced with these little choices of chance that their investment process throws at them. Penny wise and pound foolish…

But what if we step out of the frame a bit and focus on the really long term? Personally, if my health (mental as well as physical) stays with me, I very well have a further three decades of investing ahead of me. However, let’s take a hypothetical case – say you inherit a sum of money at a young age and to invest it for a really long time. [if wishes were true…] You always have a choice between a safe investment (treasuries) and a myriad of risky bets whose probabilities of outcomes is also known to you. [this is possible in a casino, but never on Dalal Street, but please lets go ahead with the logic]. Also, lets visualize that you are breaking up your investment time horizon into finite smaller periods (could be years, months or even weeks or days) at the end of which you take stock (i.e. P&L) and start again with whatever you’ve got at that time (i.e. rebalance your portfolio at periodic rests). The Kelly Formula gives you an explicit rule regarding how to divide your investments to maximize long-term growth rate.

There is this book called Fortune’s Formula by William Poundstone which I need to read to understand this better. Part two of this post will come after I have read that book and some more.

NIFTY Volatility Index (Part 2)

 In my previous post I had talked about the high negative correlation between the VIX and its underlying, the NIFTY. The chart on the right shows the actual movements of the NIFTY compared with the corresponding VIX moves. The period covering Mar’09 to Sep’10 marked a very rapid increase in the market (NIFTY rose from 3000 levels to 5,500). Since the VIX indicates volatility, the period at the start of this phase saw the most rapid increase in the NIFTY (in percentage terms) and that is why the VIX readings were above 40% during this time. The period from Sep’10 till date however have seen the market mostly move in a sideways direction. It started off around 5,500 in Sep’10 and is now at the 5,300 levels. Accordingly, if you look at the VIX, it has also remained mostly flat. Now, the big thought in my mind is this: we hear mostly bullish noises these days. Golden Cross; correction within an overall bullish trajectory etc. So, if the negative correlation were to hold and the NIFTY were to move up from now, it is logical to expect the fear (i.e. VIX) to come down from its current levels (~24%) t0 something like a 15% (?) or so. Since the start of this chart (chosen since that is the date range for which NSE publishes its VIX values) marked the markets coming out of a rather exceptional period in its history (global banking sector meltdown), VIX values > 40% may be really rare to come by again in the near future.  I have nevertheless plotted a dotted red line on the VIX portion of the chart which shows the median level of the VIX range. In the event that the VIX ‘reversion to the mean’ becomes a reality in the months to come then I guess we should brace ourselves for another secular fall in the NIFTY. If you take the economic events near the starting period  of this chart to be exceptional, and ignore the corresponding VIX values as outliers, then the new median will be quite close to where the VIX is today (22% – 24%) in which case there is definitely a case for it to fall down to late teens/earlier twenties. As you can guess I am rooting for a rise in the NIFTY and trying to twist my data interpretation to fit that notion!! 🙂

 The correlation factor in the above chart comes in at a high -0.83. Why such a high number? The reason could be because increased volatility (i.e. high VIX) signifies more risk. To keep their portfolios in line with their risk preferences, market participants must deleverage. Since long positions outweight short positions in the market as a whole, deleveraging entails a lot of selling and less buying (since the longs have to be shortened). The relative increase in selling causes downward pressure on stocks. The volume rise in the NIFTY puts really drives the VIX up. The NIFTY VIX is a weighted sum of puts (strikes < forward) and calls (strikes > forward) on the NIFTY. The weights are proportional to 1 / [(strike)^2]. As the NIFTY goes down, all the out of money puts become more valuable and those start having the highest weights (since the weights are the reciprocals of the strikes).

The unit of measurement of the VIX is in percentage terms. Its value essentially signifies the % expected movement of the NIFTY over the next 30 calendar days. So, for example, since the value of NIFTY VIX was 24.33% (on 29Mar’12), what the market is essentially saying is that it expects the NIFTY to move (either up or down) at a 24.33% annualized rate over the next 30 days. This implies that the market is pricing in a 24.33% /SQRT(12) = 7.02% movement at current levels over the next 30 calendar days. Which btw, is huge! In terms of confidence level, this means that the near term options on the NIFTY are being quoted with the assumption of a 68% likelihood (1 sigma) that the magnitude of NIFTY’s 30 calendar day return will be less than 7.02% (either up or down)

Taleb here points out how even seasoned market participants wrongly derive mean deviation (the 7.02% deviation as at yesterday’s close) from measures of standard deviation. My uninformed take on this is that if all are consistently making this error then that becomes the norm. Consider for instance, that a standard market data provider like Reuters or Bloomberg inadvertently distributes erroneous reference data – since both sides of the trade are using that same value, it becomes the de-facto standard and no one is out of balance.

VIX as a trading indicator: Since the VIX is a measure of dispersion and has a reversal to the mean property, all the standard technical indicators like RSI, Bollinger Bands etc can be used as identifiers for trade entry points. Since we do not have derivatives on the VIX in India yet, these technicals can give entry points into the underlying (i.e. NIFTY) by inverting the logic (buy for VIX would be a sell for NIFTY) due to the high -ive correlation between NIFTY and VIX. There have been studies which point out that the 2 period RSI on the VIX gives a very good trading signal. If the RSI (2) of the VIX >90, then buy the underlying and if RSI (2) < 30, then sell the underlying. Underlying = NIFTY in our case. I do not have any ready charting software at hand, so I will have to painstakingly generate the RSI values in a spreadsheet if I have to test and see how this really could have played out on historical data. Maybe fodder for a future post… 

NIFTY Volatility Index (Part 1)

The volatility index measures the market’s expectation of near term volatility. A low value of VIX accordingly implies that the market believes that prices will fluctuate very less from its current levels. This would typically be associated with low volumes – see chart on right, it shows a correlation of +0.72 between NIFTY volumes and VIX. Now, correlation is obviously not causality, and accordingly a lack of trading volumes does not by itself cause depressed VIX values. On the other hand, VIX is negatively correlated to the market, a high value (of the VIX) indicating loads of nervousness (option writers demanding high premiums due to the uncertainty in prevailing underlying prices) while a low VIX value signifying low levels of nervousness and higher confidence in the sustenance of the market levels (options writers will not find takers if they charge more for the options they write, so the premiums will be lower in a market with low volatility). A similar conjecture can be made by looking at the put/call ratio as well – low indicating range-bound or slowly rising market in the near term.

One point is important to note re the inverse correlation with the market. The table on the right shows average correlations for different time periods of observation. I downloaded VIX values from 2Mar09 till date and correlated them with corresponding NIFTY values in windows of 5, 20, 40, 60, 90… trading sessions. The table on the right shows the average correlations I got. Assumming 5 x 52 ~ 250 trading sessions in a year, this points to an average correlation of -0.63 to -0.64 over a year. In the second part of my post on VIX, I’ll try to study the actual movements of VIX over the past years and any trading opportunities that may have come about in the past.

Click here to get a white paper explaining the calculation methodology. A summary is here. Since herd behaviour like greed and fear can be deduced from VIX values, it is logical to assume that greed will chase fear and vice versa. VIX therefore has a mean reversal property making it a very useful tool for short term traders. Given the high volatility in Indian stock markets, VIX was long overdue. The CBOE (Chicago Board Options Exchange) had launched its VIX as far back as 1993. It finally came to us in April 2008 when NSE launched it, expectedly basing it on the option prices of the stocks that make up the NIFTY. There is not much noise about the VIX in mainstream financial media in India yet. Is it because it is relatively new on the scene? Or maybe it might be a bit too nerdy for most people to understand. But so are options. How many of the people who trade options really understand the math behind option prices?

Loss

I person I knew, much younger to me, full of life and with a very cheerful countenance and always helping of others expired some hours ago. I am in no mood to write anything. Pushing forward what I was to write today by another 3 days.

I did book profit on two positions today, but the transactions were devoid of any emotion.

The genius of Jim Morrison keeps coming back to mind. When he had cleverly swapped words around in a poem to create what was not his swan song, but probably the last of his own that he heard before he ….

This is the end, my friend only the end

became

This is the end, my only friend, the end

Jim’s “The End”, probably his best song was recorded in a dark studio with just one lone lit candle serving as the focal point for the right atmosphere.

The Bond Bubble

The bubbling stories going around this week in the financial blogosphere have mostly centered on the heady climb of US treasuries. In fact the topic has been quite hot the past month but the din is getting louder now. Comparisions with the dot com bubble and the housing bubble have started doing the rounds. The yield on the 10 year US paper is currently around 1% now. Which means that if you freeze the frame today, it will take a hundred years for the interest component to add up and match the price you pay for such bonds today. The P/Es (inverse of yield) of the no-brick and no-mortar tech companies were also in the heady hundreds during 1999-00. I don’t have too much of a view since it’s all happening outside of our shores. The Small Investor writes about it here as also the links I’ve listed below: it’s important enough for us to pay attention since we have NOT decoupled ourselves from the west. It’s actually the FII money that’s driving up our local markets here. Hot money.

  • FT Alphaville on the conundrum that equity prices and bond price are now moving in step. i.e. UP!

Logically, I’d guess that the bond market is bigger, more liquid and less amenable to manipulation. So, if the bonds and the stocks are sending out conflicting messages, should one not trust the former?

However, are bond markets better predictors of the economy? I think not: since nominal GDP growth and interest rates are both driven by inflation. Correlation is NOT causality. It’s a mistake many make – if two lines A & B move in tandem, that does not necessarily mean that A and B have a causal relationship. There could be a third factor C which is driving both A & B. So, bond prices are ↑; equity markets are ↑; economic data (US) is ↔. Thats the confusion. 

  • A website called bond-bubble (what else!) has come up and the graph on it’s homepage is quite telling.

It shows the super steep rise of US public debt – almost a parabolic rise. To me this looks similar to the rise of the Chinese stock market. That looked parabolic as well ( y = 4 * A * x↑2) and it could not defy gravity. But can US debt come crashing down? Maybe – if the currency crashes.

  • That seems to be awesome news for the gold bugs! It makes the case that the bursting of the bond bubble will pave the way for a massive upsurge in gold prices. The article notes that the yellow metal shines brightest in three situations – “heightened economical/financial risk; outright inflation and/or deflation”. And therein makes the case for a coming Gold bubble! Marc Faber,  (who keeps telling people to buy gold) has been bearish on treasuries right through the start of April but no one seems to be listening.

 TULIP SOUTH SEA RAILROAD ROARING TWENTIES → POSEIDON → JAPAN → DOT COM → HOUSING → BOND → IS IT GOLD NOW? 

This is making people like me (the “half informed”) even more nervous now. Ignorance is bliss – part knowledge is most painful. Anyways, the local markets are frothing on all the money that’s coming in from the US. The Fed there is busy buying up treasuries and sloshing money in their system (to buy the bonds, the Fed has to release money by paying whoever is holding bonds). They’re doing it by working their printing presses overtime spooking inflation. But I guess the game with inflation is that if you whack it too much too fast, the thing just snaps and the party careens towards deflation. I wish I had paid more attention during my economics classes. But to me it sounds logical that what comes in, goes out. So, this money will go back from where it came (at least in the interim). And all will fall down.

Though there is some more ground for the NIFTY to cover. That’s what the “experts” here are saying. The market isn’t fully stoned yet. It’s just started rolling the weed, maybe a few drags….let’s stop hallucinating. In 2008 so many of our local “experts” were shouting out that India is decoupled and that the housing bubble will not effect us. Even politicians had joined the chorous. De-coupled my moon. We are as joined to the US hip as our big bro in the vicinity.

Maybe I’ll be able to call the top.

%d bloggers like this: