Closure

Loading...

En provenance du cours de Indian School of Business

Advanced Trading Algorithms

166 notes

Indian School of Business

Advanced Trading Algorithms

166 notes

Cours 3 sur 5 dans Specialization Trading Strategies in Emerging Markets

This course will provide back test results for all the strategies in developed and emerging markets. The learner will also be taught scientific ways of back testing without succumbing to either look ahead (or) survival bias. You will learn various methods of building a robust back testing system for the strategies discussed in the previous course. You will be taught how to differentiate between mere data mining and results based on solid empirical or theoretical foundation. Next, you will learn the ways and means of back testing the results and subjecting the back test results to stress tests. After which, you will learn the various ways in which transaction costs and other frictions could be incorporated in the back testing algorithm. Finally, you will learn techniques for measuring a strategies' performance and the concept of risk adjusted return. You will use some of the famous measures for risk adjusted returns such as Sharpe ratio, Treynor's Ratio and Jenson's Alpha. You will see how to pick an appropriate benchmark for a proposed fund.

À partir de la leçon

Strategy - G Score

After completing this module you will be able to understand the meaning of G Score, build a trading strategy based on G Score and test the strategy that you have built.

Rencontrer les enseignants

Prasanna Tantri
Senior Associate Director
Center for Analytical Finance

[MUSIC]

All right, so let's summarize what we have learned in both the modules.

So the first set of trading strategies were sort of building blocks,

basic trading strategies.

And kind of trading strategies we did in this particular module are,

you can call them advanced trading strategies.

The reason being underlying papers are a little bit more complicated and strategies

require you to have some understanding of markets and a little bit of site ethics.

But at the same time none of them require a background per say, you need not

have a degree in finance to execute them I know these are easy to use strategies.

So we started Petrosky, basically all these strategies are two types.

Some of them are momentum types, and some of them are mean reverting type.

So that's the two categories of strategies we have.

For example, a strategy like is a momentum type strategy.

You know what are we doing?

We're buying growth stocks there.

That is a Petrosky is a mean reverting strategy.

By that, what do I mean by mean reverting?

In mean reverting strategies, we expect that the something which has gone up

will come down and something which has fallen will go up.

In fact there is, people do this pairs trading.

We could not cover this yet, maybe in some other module, we'll talk about this.

Pairs trading is all about this, you buy something which is going up, sorry,

going down and sells something which is going up.

And you hope that these points will converge.

That's what you do.

So these are mean rewarding cells.

Whereas in momentum strategies, what you do is

you expect the winner to keep going up and the loser to keep going down.

So these are two broad categories of strategies that we looked at.

So we started with Petrosky, the idea there was there are value stocks.

Within the universe of value stocks, there are some good and

there are some not so good stocks.

And you have this formula of figuring out which is F score,

separating this good from the bad.

And if within these renewals of values talks,

if you can separate these talks based on eccentric characteristics.

Please note all these strategies are based on eccentric characteristics,

it's not exposed.

I already explained the difference between accent and exposed.

So in Petroski,

based on accented characteristics, that means pretrading characteristics.

What you do is you separate stocks into positive gainers and

losers and you create a long short portfolio.

That's value investing.

So from then we move to price on this announcement drift where

companies that announce better than expected results, they keep on performing.

We also discusses across you know where we looked at quality of [INAUDIBLE],

It was slightly different you know in bid price announcements drift we did not

bother about quality of We just looked at earnings share price and those which there

was a positive share price and stocks where there was a negative share price.

We did not say how share price came about.

And we went long on those which share priced in analysts positively and

went short on those which share priced analysts negatively and held this for

a while, whereas in accrual we went into the quality awareness.

So we separated this outperformance in earnings

into those which is caused, that part which is caused by cash earnings and

that part which caused by accrual earnings.

The idea here is simple.

The argument is that that part of earnings,

which is given by cash is likely to be more sustainable.

Earnings are likely to be persistent and

hence the stock reaction is likely to be dominant.

Whereas, if the earning and prices driven by accrual,

then that's not likely to be the case.

That's the whole idea behind this cash earnings and accrual earnings.

And the strategy there also is simple.

We go long on stocks, we tell more cash earnings.

And short on stocks, we'd have less cash earnings and hold them for awhile.

That was the accrual trading strategy.

And we saw that it works well in different parts of the world,

we presented our own numbers to you, it seems pretty nice,

and then we went to another center school called G score.

Now G score is a counterpart of F score school, so if you relate G score and

F score, it's called of pertussis of score.

F score close to value stocks,whereas G score it's

applicable to the other half groups of stocks.

So if you're doing stocks into the market to book which is high value or

I know doing high market value by book value,

those which have high market value to book value qualifies to gross stocks.

So we created this G-score, G score which is based on

growth attributes like advertising, R&D, and all that.

And we saw that the G-score,

you know G-score does well in most parts of the world.

So that's another strategy.

One interesting about G-score and

F-score is that if you try applying G-score to the value then it doesn't work.

Similarly, if you try applying f score to the growth then also it doesn't work.

So f score applies to these value and g score applies to growth inverse.

That's the best part, that's what I keep telling you.

These are not strategies created by data mining Assume that it's created by

data mining, then this kind of a classification would not have helped.

There is reason why this works, you know.

For gross stocks, we look at attributes like R&D,

advertisement and these kind of things which are essential for growth.

Whereas for value stocks, we look at attributes which are important for

value We look at whether there has been profit,

whether there has been some kind of a turn around.

So, these are the kind of attributes we look for value stocks.

So, that's Mohanram's F, Mohanram's G-score I'm sorry.

So, we looked at that fact and then we entered the world of momentum.

Starting from the paper, the idea here being winners will outperform losers.

And we saw that such a simple strategy, how it works.

And we saw the kind of returns that this strategy makes.

So after having done that, then we went into this momentum crashes.

Now, within all the earlier strategies also, we discussed when these strategies

are likely to work and when these strategies are unlikely to work.

That is what we said.

So momentum crashing is an example of failure of momentum.

Now we saw that during times, if you remember those four conditions,

during panic times When the market has crashed, when volatility is high,

and the market is in the rebound phase, a momentum strategy is unlikely to work.

And we're actually sure that momentum strategies can really,

really go wrong in these times.

To take an extreme example, in the 1932, recall it,

we saw that a momentum strategy would have wielded, August 1932,

if I remember correctly, Momentum strategy would have entered a negative 72%.

That's what we saw.

So once we complete it,

then we devise a strategy to take advantage of momentum crash.

So we learned a little bit about options.

And we said, that if you have a strategy, when you go long gone stalls.

And go short, instead of going short, by put options, you will be insulated for

momentum crash.

But we said it, it has a cost, and the cost being, put option premier.

And finally, we also spoke about this betting against bidder.

The idea here being, hybrid stocks tend to be built up by constraining investors.

So they tend to be more valued.

And a strategy of going short on high bidder stocks

going long on lower is likely to make money.

So these are the strategies that we have discussed in detail.

We have seen the papers, we have read the abstract.

And generally you know, before that I should have told you this,

before that we also discuss how to go through a paper and what part of

an academic paper one has to look at in order to understand any studies.

More importantly, we also said what kind of a paper one can safely ignore

without harming one's cause much, of trading.

So that's all we did here.

Now in the process, we also learned a lot of other things.

We learned, as I've told you before, we learned about options.

We learned about futures, basically an introduction of options and futures.

We learned concepts about alpha and

beta, measuring performance, and so on and so forth.

Now the question is, what next?

How do you go from here?

We have another module were we'll talk about performance measurement portfolio

and so on and so forth but regarding this studies, what do we do with this studies?

An important thing to be done I think we'll put it up as part of the project

is to see how this strategies work together, that's very important,

I have discussed this before, so if say for example More.

If two strategies are negatively correlated,

that's the best thing that you can hope for.

And what happens is when one strategy does badly, the other will compensate for it.

Of course, you will not have this best of the best situation and

both strategies will do it on an average.

But at the same time this negative co relation will reduce the risk of a loss.

So it's very important that how these strategies,

it's important to understand how these strategies move together.

So we will talk about it in the next module but

I think when you back calculate, when you do this back testing of this strategies,

you should also see how this strategies move together.

So what I would recommend is go through this papers all this are main

papers that we have covered, got through at least the trading studies section,

figure out the market when you're reading and try to get pass data and

the first thing to do is to test how this strategies have what.

Second, please understand in your sort of testing, in a back testing.

There is no transaction cost you don't have to pay transaction cost

When you test.

However, in reality there are transaction costs.

You will have your brokerage, you have your various forms of taxes, and most

importantly, there is something called impact, I have to discuss about this.

If you place a large order,

the stock price is likely to go up If you, you know, by order.

If you placed a large sale order, the stock price is likely to go down.

This is more important in liquid stocks.

Liquid stock,

a large stock where there are thousands of traders at any point of time.

An order here and there will not make much of a difference Think of a small stock.

There are only five buyers.

There is demand only for five stocks at any point of time.

And you suddenly go and buy 500 stocks.

So the sellers will ask a very high price.

So this impact is very important.

So you have to consider all this.

We will talk about it in the next module, but

at least to start with, consider number, brokerage, costs and

tax which is applicable to market in which you want to trade.

Consider this and

deduct that from your national profit that you get from your back testing.

So when you do back testing it's important to, what should be your mindset.

The mindset is not to find the result.

You should always be looking for rejecting a hypothesis.

You should try to proof that this trading strategy does not work.

try in a different way You don't try to prove that it works.

So what you do is you break your sample period into months, quarters,

years and see in how many sample periods this works so have that mentality.

So when you fail to reject this hypothesis that the trading strategy does not work,

then you'll have some confidence that there must be something here.

And the next step is to do some kind of a mock trading.

Once you do mock trading, assume that you're buying a particular stock on

a particular day, using life prices, but don't invest Wait for

a year or so and see how the results are.

And then only think of starting with real money.

Remember, there is a crucial difference between mock trading and using real money.

When you're doing mock trading, it's very easy to be rational And

do not get carried away by sentiments.

But when you do real reading, it's really, really hard.

Your mind will start playing games with you.

So you should be careful about this.

So it is not necessary that you will be able to translate

your mock reading result into actual reading.

The experience may be different.

So, I wish you all the best in this.

So you can always contact us, any question on these strategies.

It leaves you in

the next module.

[MUSIC]

Explorer notre catalogue

Rejoignez-nous gratuitement et obtenez des recommendations, des mises à jour et des offres personnalisées.

Coursera propose un accès universel à la meilleure formation au monde, en partenariat avec des universités et des organisations du plus haut niveau, pour proposer des cours en ligne.

© 2019 Coursera Inc. Tous droits réservés.

Télécharger dans l'App Store

Disponible sur Google Play