A deep dive into volatility trading – part 1

Volatility trading is not:

  • Price action trading (between support and resistance)
  • Trend trading
  • Fundamental analysis

Volatility is hidden behind all these techniques but it frightens most people because we talk about noise, probability, leptokurtic distributions, and many more abstruse words.

Volatility is very easy to understand:

P(t) = P(t-1) + e

Where P(T) is today’s price, P(t-1) is yesterday’s price and ‘e’ is the delta between the 2. ‘e’ is the raw volatility!

About price, we usually use median price of the day instead of closing price, to smooth a bit the volatility 😉

‘e’ can also be seen as the 1-day momentum. If ‘e’ is positive than momentum (speed) is up; if negative, momentum is down. This is the beginning of a trading strategy. Don’t throw away your other indicators just yet, because this not usable as such! (At the very least, consider the weekly 1-week momentum and take only signals in the same direction, but the signals can unfortunately be used only in intra-day type of trading)

Interestingly enough, should you program the formula above on a computer using random function with ‘e’ representing anything between 1% and 2% variation up or down, and run 10000 loops at least, and display the results, you will see a graph that is looking very much like the graph above. You will notice support and resistance lines, head and shoulder patterns, triangles, … because you have learned to recognize them but these are mere illusions! Because these patterns can be generated by a very simple program, the conclusion is that they have no predictive power! Or in other words, it just means that if a supposed target is reached, it is purely by chance.

So you are now left with ‘e’ and you can do whatever makes sense to you. You can average, look at standard deviations, look for hidden frequencies with Fourier transforms, …. Remember to keep it simple. When some big pocket start to sell, everyone sells; when price start going up, nothing happens until the hiking is obvious, so there must be some linear component. Which we will see in an other post!

Selected books for your perusal

Hi! With on-going containment, that may come back a few times in the next year or so, I thought I would recommend some good books to read. This blog will not teach you everything you need to know!

Let’s be clear, there are many good books out there but authors usually make their living out of selling books rather than trading, justification always being having more regular revenues… or just revenues? 😉

Anyway, most books do not cover more than basics, giving only very basic information about RSI, stochastic, … Hidden divergences are almost never covered, not talking changing default parameters!

So here are a few good books with in-depth reliable information.

First one will teach you the basics. Dr Elder is quite good about explaining how to do trading and provides very advanced trading strategy. His 3 steps screening strategy is brilliant.

If you are looking at creating your own indicators, Larry Williams shows his creativity. The indicators he developed may or may not work any more, but he finds the words to kick-start imagination.

Pitchforks and Action-Reaction are the ONLY tools I know of that allow easy and reliable prediction of where prices are going and when. Don’t get me wrong, these are not crystal balls, they allow knowledge of potential targets, one or more of them are very often hit by the market. P. Mikula provides a very good starting knowledge on Andrew’s pitchforks and action-reaction. He doesn’t say how to select the pivots points however but as A. Andrew supposedly said, each 3-points sets have a story to tell. This will lead you to sound meditation for sure!

Even if we are not doing trend trading the way described in this book, M. Covel is a fantastic writer who explains in detail the automated systematic trading, explains what it is to have an edge in the market, how to demonstrate it. Your challenge is to do better than trend followers, which you can possibly do after thinking deeply with book just above.

Fractals are a way to trade with the trend, represented by an Alligator in Bill Williams book. Providing you spent some time understanding his concept, fractal trading is a very good trading system. We are missing statistics but the programming should be easy.

‘Trading with the odds’ was the first book I read that took a deep dive into the probabilistic side of trading. C. Kase explains it all in simple terms, the weekly synthetic bars are a must-have trading tool. Kase Peak Oscillator is also fascinating in its concept.

Last book, bot not least, is not about trading per se but it is a must-read. Nassim Taleb makes the case about luck: even W. Buffet may have been lucky because probability of having someone so successful at trading among billions of humans is not zero! And handling of randomness in trading precisely the core of the system!

That’s all folks. Enjoy these books and trade safely.

The holy grail of indicators

With that title, I am sure you are going to read this post to the end! If you have no time, the summary is: the holy grail just does not exist but it is worth searching for it, for the path is more important than the actual indicator!

If you are looking for the best ever indicator, you can hardly start from a blank page. You need inspiration, you need to get your brain on invention steroids! So, the first thing to do is to consider what is available on trading platforms and make a strength / weakness analysis. you need to get a deep understanding about how the indicator is built, how it behaves under different market conditions, play with parameters and multiple securities like index, stocks, trackers, …

Let’s look at MACD or Moving Averages Convergence /Divergence. I will not explain here the background, but show one graph with 2 different parameter settings.

See? One is smoother than the other. The timing of signals is the same but the eye likes second one much better. Meaning also, if you make it part of trading system fully automated, you will have less noise to manage!

MACD is a trend indicator, therefore it performs quite poorly in horizontal markets. What about a momentum indicator?

This one tells you you are in an uptrend when positive and when it goes green below 0, it may be a counter-trend trade opportunity. It looks good to the eye, not perfect as there are false signal, but looks fairly good. Of course, some back testing would be needed to validate the impressions!

Now if you are looking for the grail of indicators, here is what it should do:

  • It should tell you what is going up in higher time frame (e.g. in weekly for a daily chart). Because it means you have some good move on-going, it is not just noise in current time frame
  • It should be capable to anticipate market turns, through divergences or hidden divergences (in direction of the trend)
  • Signal should be timely, not too early and definitely not too late
  • You should be able to see and professional are doing versus all other market participants
  • It should be smooth and easy to read
  • The number of false signals should be reasonable
  • The signal should be profitable as often as possible. Refer to previous article about winning streaks.
  • Please add here you own requirement: …………………………………………..

A way to look at higher timeframe in a single chart (holy grail surely means only one chart!) is to use synthetic bars. This has been introduced by Cynthia Kase in her book: ‘Trading with the odds’ https://www.amazon.com/Trading-Odds-Probability-Profit-Futures/dp/155738911X

The idea is to to calculate MACD as if on a weekly chart but we do the calculation each day, so we have the information we want before the week is closing. MACD on weekly chart is usually used as one-bar indicator on histogram: when it changes color (goes up after going down in negative territory), it is signal to buy. Just look! Green arrows point to buying opportunities in weekly chart, sometime 3-4 days before end of the week!

There are hundreds of indicators available on the markets. Try to make sense about the most common ones, discards those that are complete nonsense, consider those that work as promised by their author (I have not found any divergence for long time on accumulation/distribution) and continue your trip! It gives food to your mind!

Your next question is of course: did I find the holy grail? I wished! Please remember we are trading prices, not the indicator! My indicator tells me where professionals enter after market bottom (red circles in counter trade), has plenty of divergences, has zero crossings or bumps (smileys) to indicate entries in direction of trend. But it is not perfect, get some false signals from time to time (because the professionals are wrong ;-), jsut kidding of course)

Of course there remains the need for risk analysis (volatility and position sizing, price objectives).

Until next time, trade safely and don’t take useless risks with virus!

Are the markets efficient?

It is a theory that has been in the air for quite long time that says that the price at any time reflect all the information available to investors.

Read again what I have just written!

How can this be true? what is the fair price of a security with all information available? Is there any magic formula for this? And why would it be valid for every one and all the time? Why does the price keep changing every day?

True enough, there are organizations (banks, Hedge funds, …) that have implemented in their system a formula based on statistical analysis. The inputs are usually price earnings ratio, price to book value, … and a few dozens others like ‘is surprise earning having any significant impact on price?’. Then they analyze real-time(!) all the information available in media, social networks, … and compute a ‘fair’ price, decide whether the security is under-priced and then they take position. Not on a daily or weekly chart! No on a 3-minutes chart! Because competition is sharp, they come up with same conclusions, and prices usually catch this fair price in a couple of hours. Risk is very limited for them! And so markets are efficient from this perspective.

For the average investor, unfortunately, this does not work unless you can buy yourself one mainframe computer and have programming skills and can analyze big data.

But markets are still moving in other time frames! Because price is made of the opinion of huge number of investors, be they humans or robots or martians, … and nobody thinks exactly like his neighbor. Reasons are linked to the security itself or external, e.g. you need cash to buy yourself a coffee! The result of this processing can only be random and you would be right questioning the market efficiency. I have proven in previous article that trends emerge out this random process and that is one of the best way to make money and beat the market!

So have fun with random processes, dig in whatever good books exist on the subject.

Why you need to master probabilities…

Everybody hated probabilities in high school. The exercises like a bag with 20 red balls and 30 black balls, draw 2, what is the probability of having 2 black balls?

Trading is managing your luck at playing with a system that is random, and any forecast you make has non-zero chance to be wrong! In other words, you are not trading the market, you are trading a belief that the market will go (or continue to go) in one direction for some time to make profits. You must not leave that to anyone having no interest in probabilities.

Take the example of weather. You are not asking climatologist about the weather next week. Climatologists focus on long term climate change and whatever happens in the weeks or even months is insignificant; in other words it is noise! Meteorologists, on the opposite, are concerned by the short term, they have made a model in which they believe, after having tested many weather conditions and they bet on something likely to happen, e.g. sunny weather, and sell this information to their customers. You know weather forecasts are never perfect but they are improving as more and more data and computing power is made available to them!

Same goes with trading. Economists are interested in long term economic cycles (Kondratieff, …) and so their comments on many TV channels why S&P500 is up 0.5% because (un)employment shows an increase of 0.3%. Not only is 0.5% up or down on market index is noise, but there is surely no correlation at all, if you would look at data from the past.

Traders main interest should and must be how to have a system that can say with a level of certainty where the market is going in the short term. The model the trader is using should have a positive expectation, i.e. being profitable over long periods.

Yes, but sometime, the weather is sunny for very long time, and nothing happens? Correct! And for trader, it is the same. When market is too quiet or too much random (like it is now), trader must refrain from trading and risk blowing up his account

Trade safely, use stops, and don’t forget to wash your hands!

Risk / Reward Ratio: my relative take on it!

We have previously discussed the win/loss ratio and the impact on occurrences of winning and losing streaks. You may fix it somehow, especially if you are lucky, but if, like for normal humans, it is not always the case, then it is not enough to be profitable.

Imagine the following:

  • Winning percentage: 40% Loosing percentage: 60%
  • Average performance of winning trades: 20%
  • Average loss of losing trades: 15%

Then your gain expectation over 100 trade is .4*20-.6*15 = -1%

Your expectation is negative, you will not win over the long term! Remember it is an average expectation, the actual can be far lower.

Whatever your win/loss ratio, the risk/reward ratio is here to help counterbalance the effect of luck. Say you bet 1€, you may lose it or gain it back. Is it worth playing it?   NO! You are better off keeping your 1€ in your pocket.

It is the same for trading. If you bet 100€, you wish to get at least 150€ or much more. This is sometimes called asymmetric position. The risk/reward is calculated opposite:

%RR = (potential reward) / (risk taken)

In our case: %RR= 150/100 = 1.5

To be honest, I never take trade with %RR less than 2.

You will see in many books or websites that you should never risk more than 1% of your trading capital on a single trade. It tends to be confusing for many because this 1% tends to become bigger as capital grows and since you still want to sleep at night, it is better think in a different manner.

The question to ask yourself is this: if you lose 100€, does it hurt? No? what about 200€? 300€? Oups it hurts too much! So be it, you are just above the pain threshold.

Take note of this value, this is the (absolute) risk, the divisor in the calculation above.

I have discussed previously how-to guesstimate a potential objective (with drunkard’s walk for instance).

So, we are ready to trade!

Let’s take an example. A signal (green arrow) indicates you have trading opportunity. We assume you will buy tomorrow just above today’s high at 3.56€. The stop is at 3.18€ and possible objective is 4.31€

%RR = (4.31 – 3.56) / (3.56 – 3.18) = 1.97

Not too bad, uh?

Now we need to ensure you are not crossing the pain threshold! Let’s assume for a moment that it is 100€. Small but why not.

Now you can size your position accordingly.

The risk for each stock you buy is 3.56 – 3.18 = 0.38€

Therefore, the number of stocks you can buy = 100€ / 0.38€ = 263 (rounded)

So you will invest: 263 * 3.56 = 936€.

Just be sure to sell if the stop is reach so that your pain does not exceed too much 100€.

Should the objective be reached, you will have gained: 263*(4.31 – 3.56) = 197€. In line with our win/loss ratio 🙂

If all goes well, after a few days, the calculated stop will start hiking up, so your risk will be void.  After objective is reached, feel free to take back some profit and let your profits run.

How did it go for this trade? It went far beyond my expectations!

In summary, you may be a stoical kind and only look at the risk in percentage or relative view. For most of us, it makes more sense to look at the risk in absolute value and sleep well at night because we are not risking more than we can afford money-wise and psychology-wise.

The foundation post: fooled by the market … and your brain!

Previous title: Markets are correcting, aren’t they?

Sorry I have not written for a while. Thank you very much for your encouraging comments. Yes it is true, luck is an important factor in trading and it needs to be factored in when working on any trading strategy. The question is always the same: how do you manage the unknown? The fact is that when the brain has seen something, which we may call a pattern, it is looking for evidence that it is there and reveal something. We are not accustomed at at all to challenge the existence of pattern.

Let’s take an example. Here is a list of numbers: 2 , 4, 6. What is coming next?

…

Most likely you are going to say 8, right?

In fact, had you answered 25 it would have been an ok answer because I only about 3 consecutive higher numbers. No other rules.

When you see a pattern next time on a chart, take time to challenge it!

Now let’s look at S&P500 and recent price action.

You do not need 20 indicators. A ROC-L (created by Alexander Elder – it is a ROC 21 days applied to 13 days exponential moving average) is good enough, quite smooth. There was a divergence on the top and signal to sell was there (red color)

That’s it. I am not looking for any other confirmation. Market has been on a long streak of winning weeks. Whatever the reason (Fed printing money by billions, or chinese virus, …), the market was bound for some other direction.

I explained previously (did I?) that the market is statistically NOT a normal distribution so don’t rely on Bollinger bands to anticipate. We are much better off with action-reactions lines (the big idea behind Andrew Pitchfork) because if prices have gone away from center line in one direction, then it is likely they will do so on the other side. Concept is simple, the secret is where to draw the center line. I need to write a book about this, since Andrew never went to such details.

So finance suddenly realizing impact of Covid-19, prices have dropped in 2 days to their second reaction line. The 2 lines below close to each other should prove a better support.

Let’s be clear, we are bearish now, with something that looks like the beginning of a bearish short term trend. I can change my mind above 3300, this number will of course go down in coming days. On a longer term basis, the bullish trend is of course still valid as long as above 3100 on weekly chart, and 2840 on monthly chart.

Due to the weight of index based ETF’s, it does not make too much sense to look for opportunities on major stock. But volume might be small on small caps making trading also risky.

Until next time, trade safely.

How far will the market go?

When you want to know where price is going to go, there are many ways to guess. Among others:

1. Return to average. If price go too far away from average, then they are likely to go back to average

2. Fibonacci retracements and extensions

3. Check analysts price objective

The weakness is the same for all: when? in one swing or after many up and down swings?

With volatility trading, we can answer thoroughly two different ways

1. The drunkard’s walk

The concept  of ‘drunkard’s walk’ was introduced by  William Feller by the following math problem:
If a drunk guy leaves the lamp post he is leaning against, how far would he have gone, on average, after n steps?”

The answer is he would have gone the square root of n multiplied by the average length of his stride:

√n * Average length of stride

Here is how to translate for price objective setting: start from a major price bottom, the first objective that should be reached within n days is the average price change (or Average True Range)

Example: if we look at an ATR over 25 periods (days, weeks), the first objective that can be reached in 25 days is lowest low (the lamp position) plus 5 x ATR25.

Easy, right? If objectif is not reached, then just give up.

When we say first objective, it is because we are considering a random walk, but if a trend is given birth (or the street has a slope for our drunkard), then prices may go much further away and you need to re-estimate the new objectives with each swing low.

2. Action-reaction

The usual way to use action-reaction is enclose price action into channels, so as to get high probability entry points. Also the distance between lines gives the potential objectives in straight lines.

I am removing previous lines for clarity but I am now drawing an other set of lines. Distance between the lines also indicates potential

By having two sets of AR lines on the chart, crossing of lines indicate precisely where prices are going to go. There may be several possible objectives, but many can be discarded (too sloppy change required, or objective in the past)

See here yellow highlighted target reached almost on time (depending on accuracy of lines drawing)

That’s it! Because we are analyzing price and time together (aka volatility), we are reaching high quality trading

Winning streak? It is all in the figures!

Here we go for a technical discussion today. We have all heard that markets are more or less random, as it represents the opinions of millions of investors or traders, and of course, they don’t usually agree with each other, the proof is that there seems to be always some who wants to sell you some stuff when you want to buy it.

So prices may look random but there is a positive bias because prices can not be negative. So normal curve distribution is not applicable. See when Bollinger bands go into negative territory, you know the objective is invalid!

So let’s assume for the exercise that we have a 55% chance that tomorrow price will be higher than today, so 45% risk they are lower. What is longest series of positive consecutive days?

Any idea?

Hint:  we need to use the law of large numbers.

The formula is:

-ln(number of observations)/ln(probability)

So the longest streak of consecutive winning days for 100000 days observed is -ln(10000)/ln(.55) = 19 days and same calculation yields a 14 consecutive loosing days.

Yes, that is what swing trading is about. Anticipate these streaks of consecutive winning or loosing days and play them accordingly.

The same goes with goes with your trading strategy. The probability lays in your own skills to define a winning system. You need to backtest with as much data as possible. A system that is winning 85% of the time can still yield 6 consecutive loosing trade (but also 70 consecutive winning trades ;-))

What happens with robots trading? There are 2 points to consider:

1. Robots have been programmed by human. So they just replicate our trading strategy. So same as above applies

2. Robots can be taught to learn (machine learning, neural networks). After some time they will all more or less reach the same strategy and so they will ‘see’ same thing, so market will become flat and cancel out robot trading efficiency if any.  Some news (real or fake) will be needed to move the market. Robots will learn to cheat and will need support from humans!

So as conclusion, you need a strategy with an edge to catch swings or trends, they are predicted by the random nature of the market. You can protect your strategy by:

1. Not sharing it. Keep it secret.

2. incorporate elements that can’t be automated (e.g. action-reaction lines)

3. Incorporate strong psychological elements. Trend traders for instance do not have more than 30% success rate for their trades. Meaning they may have to endure 32 loosing trades in a row! Without excellent money management and psychological force to endure this, you blow out your portfolio and yourself in no time! But you can also catch the huge moves!

With volatility trading, I can reach 60% success rate on trades, so easier to stand psychologically, up to 85% in very volatile markets.

Until next time, trade safely

The hidden beauty of volatility bands?

You may have wondered why in previous post I talked about action-reaction lines for trading. This is very technical and manual but the principle will be re-used.

Volatilty bands do already exist. There are Bollinger bands (based on standard deviations) or Keltner bands (based on ATR). Is that it?

No someone created also the ‘better Bollinger bands’ (sic). See some code here: https://www.prorealcode.com/prorealtime-indicators/better-bollinger-bands/

Volatility still is little explored. You know when you are on the right path when it becomes simple.

Look at those bands. If green, go long! (Consider long term trend of course)

This one is even more fun. Formula is deceptively simple. And there seems to be linearity in the market!!! Of course, humans think in straight lines! Just follow them and trading becomes profitable

Apple Weekly

Until next time, trade safely

Previous Next