ilYzyPKPwhICaq

Discussions specific to trading the stock market.
cryder
Roundtable Knight
Roundtable Knight
Posts: 101
Joined: Tue Jul 22, 2008 5:02 am
Location: Melbourne, Australia

Post by cryder »

Stock filters I use are Market Cap, volatility as a % of price and moving average of volume. Allows the instruments to be grouped in approximately 6 classes.
nodoodahs
Roundtable Knight
Roundtable Knight
Posts: 218
Joined: Wed Aug 09, 2006 4:01 pm

Post by nodoodahs »

Most of CANSLIM's filters are fundamental, athough I think the lookback period used is atrociously long.

The so-called fundamental "anomalies" are legion and all you have to do is choose a small handful to bias the portfolio towards profitability, these could be absolute, or relative to the market, or the index, or the industry.

Liquidity (average volume x average price over some recent period) as a filter has the advantage of directly translating into your slippage.

Index membership as a filter generally guarantees some level of liquidity and provides an easy benchmark for the system's alpha/beta.

@ ecritt, you mentioned the CANSLIM funds falling out of the game, I'm curious, do you think they ignored the market-timing trend component of the system? After all, if they were truly following CANSLIM principles, shouldn't they have been in cash for most of the crash?

[Edit: oh SNAP! Even if the funds HAD been using that component, and HAD been in cash, they'd STILL be out of business! Don't those fund allocators pay those managers to INVEST and not to HOLD CASH?]
ecritt
Roundtable Knight
Roundtable Knight
Posts: 134
Joined: Sat Aug 28, 2004 3:44 am
Location: Phoenix, AZ
Contact:

Post by ecritt »

@ ecritt, you mentioned the CANSLIM funds falling out of the game, I'm curious, do you think they ignored the market-timing trend component of the system? After all, if they were truly following CANSLIM principles, shouldn't they have been in cash for most of the crash?

[Edit: oh SNAP! Even if the funds HAD been using that component, and HAD been in cash, they'd STILL be out of business! Don't those fund allocators pay those managers to INVEST and not to HOLD CASH?][/quote]

All but one of the canslim mutual funds I've observed over the years didn't last long enough to experience the crash. I don't know why they went out of business. As for the canslim "traders" I've known over the years, they talk a good game on the way up but haven't been able to survive on the way down. They seem to be uncomfortable "doing nothing" when "doing nothing" is warranted.
rgd
Roundtable Knight
Roundtable Knight
Posts: 396
Joined: Thu Aug 10, 2006 5:21 pm

Post by rgd »

In all of the research I have done in futures markets, I have found no benefit to constructing a portfolio by anything other than organic methods. By organic, I mean the process of entering and exiting positions based soley on a first come basis. If my risk budget has been consumed, all subsequent signals are rejected. I will have to get stopped out of an existing position before I will evaluate a new one. While my research in this area has not been exhaustive, I just haven't received any indication that the time invested in this area will offer any benefit which may exceed the increase in system complexity.

I have read much about relative strength applied to equities. I have seen the exciting quilt charts which shows the range of performance for various sectors on an annual basis. However, I have yet to find anyone with a published track record that can exploit this opportunity in any meaningful fashion. Especially when compared with a relevant benchmark. In risk adjusted terms, they are not capturing anything as they are merely moving into high-beta sectors.

In my own research, I have found relative strength to offer nothing in risk adjusted terms, prone to extended periods of underperformance, unable to overcome transaction costs when scaled, and very sensitive to the strength measure.

I am interested in hearing if anyone finds an edge in constructing a portfolio by anything other than organic methods, whether in equities or futures.
trender41
Full Member
Full Member
Posts: 22
Joined: Fri Apr 08, 2011 2:19 pm
Location: Washington, DC

Post by trender41 »

when you guys all looked at stocks, did u use only the TB universe? or did you add data?
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

Look at stock indices.
trender41
Full Member
Full Member
Posts: 22
Joined: Fri Apr 08, 2011 2:19 pm
Location: Washington, DC

Post by trender41 »

AFJ,
Thanks! as I mentioned on a different thread, I have worked at mutual fund and a hedge fund, both deep fundamental. However, after months and months of selling the idea, I convinced the team to use DMA's to overlay over the market to better understand risk. That was it, I was sold and now I am here! Thanks so much, your quick and kind responses are GREATLY appreciated.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

The Traditional Approach


The charts and table below show the results of buying and holding the MSCI AC World Index (Net, Standard, USD) over the past ten years (based on end month data points) and are contrasted to the results obtained by the employment of a systematic, momentum driven strategy based on Equity Index Futures and a Triple Moving Average System (“TMA Systemâ€
Attachments
Table.JPG
Table.JPG (25.08 KiB) Viewed 15454 times
Charts.JPG
Charts.JPG (73.54 KiB) Viewed 15457 times
Last edited by AFJ Garner on Thu Apr 21, 2011 9:01 am, edited 1 time in total.
Chuck B
Roundtable Knight
Roundtable Knight
Posts: 481
Joined: Thu Apr 17, 2003 6:34 am

Post by Chuck B »

AFJ Garner wrote:Buy and Hold based on the MSCI AC WORLD INDEX Net Standard contrasted to a simple triple moving average system and a portfolio of 22 equity index futures. 350, 80 and 25 day MAs. An off the cuff example of what can be done to improve the lot of the hapless equity investor.
...only problem being that said mutual fund manager will immediately boot the system (and the employee recommending it) after the 1st, or assuredly the 2nd, losing trade. :wink: :D
nodoodahs
Roundtable Knight
Roundtable Knight
Posts: 218
Joined: Wed Aug 09, 2006 4:01 pm

Post by nodoodahs »

No, actually the pension fund investing in the manager will fire them because their returns have too much "tracking error."

Or, alternately, the first time the manager winds up holding "too much cash" the pension fund will fire them because "we pay you to INVEST and not to HOLD CASH."

Or perhaps the first time there's a calendar year where the manager under-performs, they'll get canned.
Chuck B wrote:...only problem being that said mutual fund manager will immediately boot the system (and the employee recommending it) after the 1st, or assuredly the 2nd, losing trade. :wink: :D
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

In many respects, a fund passively tracking the MSCI AC World Index constitutes a fine investment and satisfies the vital requirement of wide diversification. The index contains around 2,500 large and mid cap stocks, spread across 45 developed and emerging countries and is calculated to account for the reinvestment of net dividends. It can and should be bettered by sophisticated asset allocation and perhaps a timing mechanism.

How much better off would the man on the Clapham Omnibus have been to have ditched his chinless stockbroker (dimwitted scion of a noble family) and stuck his money in a tracker!

As I said once before, it makes me chuckle when I recall the reaction of my neighbour in Klosters after I unwittingly passed him a copy of an article I had written on the topic. The old boy is far from dimwitted but unfortunately he had spent his entire life working for private client stockbrokers and the like.

He choked; his face went bright red, he coughed, spluttered. He called me a fool and an idiot and he has not really spoken to me (much less forgiven me) to this day. I am indeed an idiot - to have had the stupid tactlessness to have copied him the article.

Sadly I had far from perfected my Buddhist practices and became very upset. I nearly capitulated to the promptings of an evil goblin who suggested I wrap a steaming, fresh turd in newspaper, place it on my neighbour's dormat, set fire to it and watch while he attempted to put it out.

There may be some value in stock picking in emerging or poorly researched markets. A friend reports many years of excellent performance in a Chinese small companies fund. But in general how can the thundering herds expect, as a class, to outperform the market when they ARE the market?

Where, indeed, are the customers' yachts?
doubleR
Full Member
Full Member
Posts: 15
Joined: Thu Jun 12, 2008 3:25 pm
Location: London

Post by doubleR »

I have tested on the 200 european large cap stocks from the DJ 200 LCXP index (on bloomberg) a LTTF (long only) from TB.

The risk per trade is 0,05%

CAGR 13% from 1/1990 until 3/2011 with a 22.3% Max DD

I use zero for interest rates to penalize the system. Divi included in the adjusted price of each stock.

As AFJ mentions this would be a system difficult to "sell" because from mid 1998 until mid 2003 the equity line is basically flat with the risk profile oscillating between 30 and 10%.

The idea to be "invested" or to do "something" would scupper any idea to put research into practice until one has found enlighened investors/firms.
Attachments
LinearEquityDrawdownGraph_P1.png
LinearEquityDrawdownGraph_P1.png (13.23 KiB) Viewed 15174 times
drm7
Roundtable Fellow
Roundtable Fellow
Posts: 96
Joined: Sun Apr 20, 2003 9:02 pm
Location: Richmond, VA

Post by drm7 »

doubleR wrote:I have tested on the 200 european large cap stocks from the DJ 200 LCXP index (on bloomberg) a LTTF (long only) from TB.
Thanks for your contribution! Not to be nit-picky (or maybe I am :P ) I had a few clarifying questions:

1. What is the buy and hold return (with dividends invested) for the index itself for the period?
2. Have you looked at:
a. shorting the index the whole time?
b. shorting all the stocks without long positions?
3. Is there any survivor bias in your stock selection? The components in the index in 1990 aren't the same as the components in the 2010 version of the index. Did you adjust for this? What about companies that were bought/went out of business?
doubleR
Full Member
Full Member
Posts: 15
Joined: Thu Jun 12, 2008 3:25 pm
Location: London

Post by doubleR »

1. What is the buy and hold return (with dividends invested) for the index itself for the period?

SXXR index starts early 1992 at 100 and it is currently at 461 after almost 20 years so an 4.6 x increase. My portfolio starts with 1,000,000 and ends at 13,000,000 (so a 13x increase)

2. Have you looked at:
a. shorting the index the whole time?

YES, it does not pay unless it is 2000-2002 and 2007-2009


b. shorting all the stocks without long positions?

YES, it does not pay AT ALL, you just lose constantly money and you recover something in the above mentioned periods but not enough to cure the grief...

3. Is there any survivor bias in your stock selection? The components in the index in 1990 aren't the same as the components in the 2010 version of the index. Did you adjust for this? What about companies that were bought/went out of business?

totally - I use the current composition of the index. Guilty as charged.
However consider that 3/4 of stocks move together and those dead should been hot stocks first and then bust later. A TF approach would have allowed you to run the upside and stay out of the declines.

however this test more or less compare favourably with the test on the SECTOR indices from 1988 where the information re: dead stocks is there included.

not perfect... but good enough for me
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

doubleR wrote: However consider that 3/4 of stocks move together and those dead should been hot stocks first and then bust later.
Which is why an index tracker is in itself a type of mechanical system. You are only buying and holding those stocks which were in the index both on the date you bought and at the end of your holding period. Dying stocks fall out of the index (the ETF provider sells them) and stocks on the increase come into the system ( your provider buys them).

It seems to me to be valid therefore to devise and back test a higher level system using index futures or cash indices. It is something you could actually have traded. Can actually trade. You can argue the same for sector or sub indices.

If you want an accurate story, I believe you have to look at the full picture. What you have done is certainly an interesting exercise but it lacks a measure of reality in that it is extremely unlikely without the benefit of hindsight that you would have conducted TF on the current constituents from day one to date.

I'm not at all sure how useful it is to take an index's current constituents, back test those as individual stocks and come to meaningful conclusions as to what performance you would have achieved. If you want to back test a system on an index, it seems to me you have two choices: back test the index itself (or an index future) or back test all historic constituents of that index as they changed over time.

I believe a back test should, to the extent possible, mimic reality. No criticism at all. But when looking for some indication that past results may bear some predictive value I would rather "trade" in my back test as I would have had to have traded in reality through the period of the test.
Moto moto
Roundtable Knight
Roundtable Knight
Posts: 427
Joined: Mon Jun 01, 2009 4:12 am
Location: once again in the UK

Post by Moto moto »

Mr Garner makes a great point and the perfect example to be able to understand it is always the NASDAQ bubble.....some of the big winners then are gone...but no one knew which ones would survive
(and his point is also particularly relevant when it comes to trying to backtest equity options trading - very hard to replicate reality)

Trading stocks now there are certainly more opportunities to hedge than there were - eg; there are now small cap indexes, small mining indexes available..... particularly relevant here in Australia as the index is made up of 200 stocks but 80% of the index is in 20 stocks....hard to use as a hedge. While this maybe an extreme example, its something else to take into consideration. Often the hedge is not worth much.

then you have the issues of shorting......many stock shorts were banned during the GFC, often you cannot borrow the stocks, or even worse you are short and then the stock is recalled, forcing you and others into a short squeeze....too many issues, to historically backtest maybe

On saying that one other idea to test as food for thought is a filter of the index as a whole before taking longs the index needs to break, or you only hedge and short the index or take shorts when the index breaks some levels. (at a guess this might smooth some whips, but from experience too often the stocks that have been flying the most and the ones you are long, have drawdowns before the market triggers any shorting opportunity)

Another thing, that stems from many years trading stocks, is that often the first cut is the cheapest. if the stocks do not run on a break get out, which always implies there might be value in position sizing variations based around pyramiding and cutting quicker right at the start.....which then raises issues of liquidity as you get larger....but not such an issue at present.
There is an on going discussion in our office at present as there are lots of placements (capital raisings) in equities at the moment, and this is another source of value that is often not shown in historical data. The point we are discussing is about when participating in placements the value of keeping them and running them or cutting them quickly.....ultimately you run out of money if you run too many of them, as many have to be funded 100% - not as good as futures - which may be a good thing :)
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

Another point I would make is that using a fixed fractional position sizing strategy on cash instruments can lead to vast distortions if a narrow stop is set. Using a bet size of 0.5% on a basket of 200 stocks can be realistic but only if a stop is used of such a width that you are remaining within realistic margin constraints.

Forgive me if I am teaching a grandmother to suck eggs but if you are relatively new to back testing and have simply used the built in money management strategy which comes with TB (and which was designed for use with the highly leverage futures market) then you had better take a look trade by trade through the position sizes taken in your back test.

If you use a normal stop of somewhere between 1 and 10 ATR then you may find that unless you have deliberately set the leverage parameters to extraordinary (and unobtainable levels) (1) you are still seeing high leverage because of the way new positions are processed by TB and (2) you are achieving great (and probably unintended) concentration on a small number of positions.

I have explained the point elsewhere on the forum in great depth. In the coding used for the systems I wrote about in my book I went to great lengths to design a different position sizing routine for cash instruments. You may well have taken a look at it. It aims to avoid unintended use of margin.
Moto moto
Roundtable Knight
Roundtable Knight
Posts: 427
Joined: Mon Jun 01, 2009 4:12 am
Location: once again in the UK

Post by Moto moto »

yes....While I have not tested stocks using TB - I dont have the access to accurate data for the stocks that are all back adjusted, so I end of fudging things to take a best guess - I figure I dont have the time to spend to do an inaccurate test - What I then do is to run things in a discretionary, but systemised manner....however, one of the fudges I take is to try and through in a volume filter, and just avoid many illiquid stocks, or take very small amounts, but the main filter is with having to set a wider stop in order to get some sort of quantity to trade based on volatility, and then set my actual "I am wrong" stop much closer in the market.
It does mean smaller positions....eg; a $25,000 stock position will often return or loose as much as 1 equity market contract, yet far quicker, and more turnover, but the old adage of once in profit, let it ride still works.

the other trap as you mention when using 100% cash v margining that gets interesting is when the margin provider/broker/clearer provides varying levels of margin for various stocks. Meaning something like a BHP, RIO, requires only 15% margin (ie; $15 required for every $100 of underlying), while a smaller cap or less liquid - but still a $2bill, more volume turnover in stock per day, is 80% margin. This can cause problems if you get stuck with the low margin stock causing outsized positions compared to your cash balances.....Especially if the margin goes from 15% to 80%, (and people complained about doubling the futures margins :))
.I assume everything is 100% cash, and work on the basis, anything not receives some extra cash in the bank. After margining costs and long/short cash carry costs, this still ends up getting you - so its easiest to assume 100% cash.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

I had originally drafted the above post for an article in a UK based wealth manager magazine which the editor had enthusiastically requested. Since he has not had the courtesy to come back to me on the article despite several follow up e-mails and a telephone message, it seems a pity to waste it so I post it here by adding the text to the charts already posted above.

Note that different periods will show very different results and that a system, whether based on futures or cash equities, will not always necessarily beat buy and hold in either relative or absolute terms.

It is a surprise, after so many years just doing my own thing on a proprietary basis, to be faced with the offhand impoliteness and couldn't give a toss attitude of the much of the outside world.

The moral (if any) is that if such behavior offends you then don't look to raise outside money. Let it come to you or not at all. Of course, only the relatively well heeled can afford to take this attitude but my sympathy goes to those who must walk the streets and beat the path of "smile and dial" to keep the home fires burning.
Chuck B
Roundtable Knight
Roundtable Knight
Posts: 481
Joined: Thu Apr 17, 2003 6:34 am

Post by Chuck B »

AFJ Garner wrote:I had originally drafted the above post for an article in a UK based wealth manager magazine which the editor had enthusiastically requested. Since he has not had the courtesy to come back to me on the article despite several follow up e-mails and a telephone message, it seems a pity to waste it so I post it here by adding the text to the charts already posted above.

Note that different periods will show very different results and that a system, whether based on futures or cash equities, will not always necessarily beat buy and hold in either relative or absolute terms.

It is a surprise, after so many years just doing my own thing on a proprietary basis, to be faced with the offhand impoliteness and couldn't give a toss attitude of the much of the outside world.

The moral (if any) is that if such behavior offends you then don't look to raise outside money. Let it come to you or not at all. Of course, only the relatively well heeled can afford to take this attitude but my sympathy goes to those who must walk the streets and beat the path of "smile and dial" to keep the home fires burning.
Thanks for the updated post. I'll have to check it out in detail in a bit.

That's not surprising about the dude failing to even bother with replying back to you. He wanted to use you for some need he thought he had. He now realizes he doesn't want to use you, so he just drops you like you never existed. It's a sad fact of "today's world" that people are so focused on the short term and what they can get immediately rewarded for that they neglect common courtesy, common manners -- in the long run, those like that will feel the effects of their decisions imo.

In the corporate environment, I've long held the opinion of "what gets rewarded is what gets done". Perhaps add to that "what doesn't get punished also gets done". Sad, but true.
Post Reply