The Portfolio MYTH-----Do you need to change your thinking?

[Wyatt]

Certainly in trend following, buying a reasonable no. of different stocks increases your likelyhood of owning most of the stocks that go on to outperform and may make up for the large no. of losers or non performers.
If you could go straight to these performing stocks , that would be nice but unrealistic, While it has been mentioned by market wizards that winning stocks, tend to go up right from the get go, backtesting suggests tight stops decrease returns.
To be able to make money holding just 1 or 2 stocks, you really must be at the screen full time, hunting the high liquidity action, willing to jump in with size and do a bit of line wiping, hang your nuts out in the breeze and pin your ears back. We've all seen it, but that is not for everyone. Unless your algo is that esquisite

Is the answer momentum, value or high octane day trading?

 

[tech/a]

Post # 1 of this thread. Tech/a is correct. End of thread!!

I think it's done

 

[odds-on]

I personally do not see the issue with holding only a couple of stocks in a portfolio as long as one is aware of the potential distribution and willing to act accordingly.

Most super-investors seemed to be highly concentrated/leveraged in their early years. There are probably numerous others who were similarly concentrated/leveraged over the years but lacked the luck/skill/timing so eventually blew up. This is worth thinking about a lot before going down the concentrated/leveraged path.

 

[tech/a]

Certainly in trend following, buying a reasonable no. of different stocks increases your likelyhood of owning most of the stocks that go on to outperform and may make up for the large no. of losers or non performers.
If you could go straight to these performing stocks , that would be nice but unrealistic, While it has been mentioned by market wizards that winning stocks, tend to go up right from the get go, backtesting suggests tight stops decrease returns.
To be able to make money holding just 1 or 2 stocks, you really must be at the screen full time, hunting the high liquidity action, willing to jump in with size and do a bit of line wiping, hang your nuts out in the breeze and pin your ears back. We've all seen it, but that is not for everyone. Unless your algo is that esquisite

Is the answer momentum, value or high octane day trading?

You know
The long term investment type threads I see here with 20 or so stocks all show a good number with massive drawdowns.
I've often wondered why they just keep holding regardless of how wrong they are!

Your back test statement can you qualify that?

 

[odds-on]

How about what they were doing, say, 40 years ago....(accessable via Chairmam's letters via BRK site if interested). Anyway, it's rather more than 10. And that's when they had a lot more flex than today.

What about 60+ years ago? 75% of net worth in GEICO.

 

[Wyatt]

The long term investment type threads I see here with 20 or so stocks all show a good number with massive drawdowns.

Yes it is the age old problem with trend following, definitely interested in your concept, monitoring just a few stocks gotta be easier than 15-20 and I guess you would have low market exposure.
As a highly respected and active member of ASF, i would love to see some of your backtesting and do please keep the ideas flowing.

Speaking of holding on, my backtesting suggests monthly rebalanced portfolios tend to increase win % and lower D/D. A bit like Radge's US momentum system

http://bettersystemtrader.com/037-cesar-alvarez-studies-stop-losses/

 

[DeepState]

What about 60+ years ago? 75% of net worth in GEICO.

Well, that's true. He was also 21 yrs old and had a total asset base of $20k at the time.

Do you think that situation represents a sensible thing for others to emulate for meaningful investment? It certainly worked for him.

 

[howardbandy]

Deepstate may mis-understand my point.

The distinction between investing and trading is, to my point, that investing is locking up the funds "forever." An example might be buying a property that will be used for a family estate. Or buying a company, such as a winery, planning to manage it as a business "forever." These are very infrequent. I do not have any.

Any transaction where I expect to reverse it during my lifetime -- maybe in years, maybe in days -- is a trade. Even other real estate transactions, such as buying a house to use as a residence or an apartment to manage as rental property, are trades. Stocks of companies I do not manage are clearly trades.

My point is that, for those who follow quantitative principles, trades are entered when the rules indicate that there will be profit, and exited when the rules indicate there are better uses for the funds. The rules create a model. When applied to a tradable issue, they create a system. The system has risk and profit potential. All such systems are alternative uses of funds. By normalizing risk, profit potentials can be compared. And the funds allocated to those with the highest profit potential.

This thread is about portfolios. Continuing on with my point, portfolios consist of several issues held simultaneously. Each of them has its own risk and profit characteristics. My recommendation to identify those separately and use the funds for the best rather than for the average.

Best, Howard

 

[Sir Burr]

So if you select 20 stocks I'd have thought you'd have to be a lot
Better at selecting stock than people who pick one or two.

I run a couple of large position number systems and backtesting them with just 2 positions reduces profits substantially and increases drawdown moderately.

In my case I "think" it's simply not having the opportunity of picking up some nice trades while only selecting two. Maybe you can pick those nice trades, who knows

monitoring just a few stocks gotta be easier than 15-20

I'm not sure if this thread is about daytrading, EOD or investing but if EOD pressing the Explore button in Amibroker where it exports to a CSV. Send via the DDE spreadsheet to IB makes no difference in time between 1 or 100 positions.

 

[tech/a]

I run a couple of large position number systems and backtesting them with just 2 positions reduces profits substantially and increases drawdown moderately.

In my case I "think" it's simply not having the opportunity of picking up some nice trades while only selecting two. Maybe you can pick those nice trades, who knows



I'm not sure if this thread is about daytrading, EOD or investing but if EOD pressing the Explore button in Amibroker where it exports to a CSV. Send via the DDE spreadsheet to IB makes no difference in time between 1 or 100 positions.

I have absolutely no doubt that that would be the case.
In a systems test scenario you are waiting for the test parameter cycle to do its thing.
From my own trading in real-time my findings are as follows.

(1) Each selected trade must have a buy trigger.(In futures there are obviously sell triggers). Many never trigger.

(2) I want selected stock to move in my direction immediately.
If it does I move the stop (a fixed technical point generally at the bottom of a pattern)
the distance of the move in that day.(I use fixed $ value for the initial buy)
Ill keep doing that until I reach B/E at that point I only look at expected target or technical signals of
stalling or stopping for evaluation.
This leads to quite a few B/E Trades.
Very rare full loss of initial stop as I raise the stop if the move is against me after a triggered buy.
So I do have turn over pretty quickly before nailing one that goes.

(3) Most stops are between 1.5-.5 % when hit OR at B/E if it takes off and turns.
This I believe is a key. On Average it is quick to kill off accumulated losses with a good win or two.
Similar to Peters methodology in some respect but much quicker on stops and moving of them.

(4) Once in profit and at B/E Ill let it run and add aggressively if it triggers another setup and this can be
from an anticipation off a single bar. Ill only hit it for that day or to a target (Within that day---another key I think). Leaving the
initial trade to run its course--targets or technical pattern for exit, including stagnation.

You do have to be on top of it particularly in the initial phase and if your hitting it. I have real time at my desk and on the phone.
Mind you I miss quite a bit (Timing) as I do my day job---I have an understanding boss.

My findings ----

 

[odds-on]

Well, that's true. He was also 21 yrs old and had a total asset base of $20k at the time.

Do you think that situation represents a sensible thing for others to emulate for meaningful investment? It certainly worked for him.

It depends on your objective. It does provide an example of how to compound capital. Just got to pick the right stock.

 

[DeepState]

Just got to pick the right stock.

Yep. That's what all of this comes down to. If you actually could pick the right stock with high confidence...you don't need much diversification. Actually, scrub the confidence bit and replace that with ability.

Also, you don't need much diversification if buying 'lottery tickets' where no other means exist to achieve your aims or otherwise you really don't care what the outcome is.

 

[tech/a]

Trade enough setups reject enough
And you'll find a lot of right stocks.

If your really interested read the above
If you have no interest.Happy with what
You now do.

Return to village.

 

[DeepState]

Deepstate may mis-understand my point.

Thanks for the clarification.

This thread is about portfolios. Continuing on with my point, portfolios consist of several issues held simultaneously. Each of them has its own risk and profit characteristics. My recommendation to identify those separately and use the funds for the best rather than for the average.

So, for a situation there there is even a low (but strictly positive) degree of aversion to loss and even high predictive ability (say, Info Coefficient of 0.2 at monthly interval), how many stocks might appear when the universe is the ASX 200 or S&P500? And signals are available on the full cross-section.....

Would it be 2? Or closer to 20+

 

[DeepState]

Trade enough setups reject enough
And you'll find a lot of right stocks.

If your really interested read the above
If you have no interest.Happy with what
You now do.

So...I am interested. Which is why I am fielding alternative viewpoints. Just that I also check the maths to make sure 1+1 actually does equal 2, and not 15 as sometimes claimed or imagined.

Trade heaps of setups, reject heaps, there will be heaps left, some of them will be good ideas? And, if by chance, some of them go up and you lock to breakeven, it's amazing how hit rates go up!

That's all true. Of course. It actually is. Now...where is the flaw in that argument that leads to a belief that this amazing investment ability is real? Howard would know. And, if you've read all his books, you'd also know.

Return to village.

Don't like your ideas actually tested by someone with evidence and a knowledge of how this stuff works for real? Thought you might be more awesome than that.

 

[Quant]

The people with the answers to this will never give the keys to the palace away . An edge is only an edge until the masses see and use it . The Howard way is the key , statistical patterns in the market exist and the inefficiencies allow these patterns to exist and be exploited . Now i can claim this is factual and all the usuals will demand evidence . The thing is a smart operator keeps the keys to the palace in his pocket so really this whole thread will continue to go in circles , Its possible to beat the index by 100% year in year out with some very basic strategies and not go outside the top 50 stocks with only approx 50% time in market . Mean reversion allows this exploitation , you just need to find the patterns . The more stable companies are actually the easiest ones to exploit . No-one with the answer is going to share it is basically the point i'm making . Circles is the best you can expect .. good luck to all

 

[craft]

Yep. That's what all of this comes down to. If you actually could pick the right stock with high confidence...you don't need much diversification. Actually, scrub the confidence bit and replace that with ability.

Also, you don't need much diversification if buying 'lottery tickets' where no other means exist to achieve your aims or otherwise you really don't care what the outcome is.

Considering everything we know for certain relates to the past but every decision we need to make about risk relates to the future. Unless your second paragraph applies scratch ability and go back to a "special" kind of confidence.

Its amazing how many traders on the internet think /act as thought they are the trading equivalent of Alex Honnold - hopefully they don't have too much on the line when they discover they are not.

 

[DeepState]

statistical patterns in the market exist and the inefficiencies allow these patterns to exist and be exploited . Now i can claim this is factual and all the usuals will demand evidence .

I accept this as a fact without further evidence required from you. The question proceeds as to whether, even in this instance, we need to reconsider the portfolio myth. That is the question the Duck posed in the header. I have posed the question to Howard with a scenario that most strongly favours the response of a concentrated portfolio. If the coefficient were much higher than 0.2 (and I have not even asked about frictions to trade - so this is net of frictions), you'd be trading so well that ASIC would be watching.

Circles is the best you can expect .. good luck to all

Let's shrink the circle.

Duck says two stock portfolios. I agree in certain situations. These include complete indifference to the outcome or risk seeking behaviour (likes thrill). An alternative includes having insufficient skill to meet a target saving level and thus resorting to highly speculative behaviour. This is a lottery ticket, and it is rational in this situation. So it may be fine for duck and many others.

So let's talk about something less trivial, where this stuff actually matters. Someone who lives off this stuff or where the assets and how they are invested has a material impact on them. What then? Two stocks? Even if you can generate 100% per annum via statistical reversion (which does exist and is the realm of HFT and leverage) these portfolios are very diversified. The individual edges are tiny. Your get these results by diversification across stocks and trading very frequently...and applying leverage.

In this much more narrowed situation, where the evidence load has been dropped to ankle height, what is your view about the portfolio myth? I want to know because this is my situation. I care and I am never happy about what I do and very much like to find where my assumptions are wrong...but arguments to such ends need to actually make sense. It took me five years to unlearn what University taught me in many areas (some time ago now...I'm not a freshie). Maybe the importance of diversification is another one of these.

Or, have we now devolved in to "Just 'cos" in which case...yes, good luck to all...this has been fun and useful in that context alone.

 

[Quant]

Duck says two stock portfolios. I agree in certain situations. These include complete indifference to the outcome or risk seeking behaviour (likes thrill). An alternative includes having insufficient skill to meet a target saving level and thus resorting to highly speculative behaviour. This is a lottery ticket, and it is rational in this situation. So it may be fine for duck and many others.

Im all for small portfolios but in the case of say 2 stocks id much favour indice trading . My idea is a small universe to follow , know them well and deal with them as dynamics change , i primarily follow top 50 as they are stable , less prone to unexpected movement and they have very distinct cycles/patterns/price imbalances that have high probability . I look for what WILL happen , im not looking for lotto wins just a heap of 10-20% moves inside a basically 4-8 week time frame , i wont put all my money in a one position but i may put up to 30% depending whats available . I look for predictability , top 50s over shoot down and up , identifying when the downs and ups happen is the mean reversion bit , just look at say a 120 day rolling range in top 50 stocks , thats the available range to exploit

So let's talk about something less trivial, where this stuff actually matters. Someone who lives off this stuff or where the assets and how they are invested has a material impact on them. What then? Two stocks? Even if you can generate 100% per annum via statistical reversion (which does exist and is the realm of HFT and leverage) these portfolios are very diversified. The individual edges are tiny. Your get these results by diversification across stocks and trading very frequently...and applying leverage.

lets be clear im not saying 100% returns just > 100% indice outperformance with NO negative years and no leverage , with leverage that 100% return is doable but as we know that double edged sword is what explodes accounts . Can only speak for myself but the edges arent tiny

In this much more narrowed situation, where the evidence load has been dropped to ankle height, what is your view about the portfolio myth? I want to know because this is my situation. I care and I am never happy about what I do and very much like to find where my assumptions are wrong...but arguments to such ends need to actually make sense. It took me five years to unlearn what University taught me in many areas (some time ago now...I'm not a freshie). Maybe the importance of diversification is another one of these.

The portfolio myth , its not a simple answer , markets change and are dynamic , in recent time very much so . Investors refer to Buffet a lot and looking at BRK in the greatest bull run from mid 70's to 2000 the results are spectacular and not so much since 2000 . Now sure Buffet beats the index return by 100% but he does so with 50% drawdowns , not accepatable in my world and obv easier to deal with not owning 10% of the market . Buy and hold is a flawed approach in todays world and having 30 stocks isnt ideal , people bury the risk and dont deal with it and that is crazy . Its hard work getting the right strategies and i would never say its easy but it is doable . Dealing with risk is mentally easier with a smaller portfolio as you are basically forced to do it . If you arent accurately proactive buy an index and beat the majority of funds , you cant rely on luck . Mediocre methods get mediocre returns .. old cliche but it really applies ... dont confuse a bull market with genius , its the **** years that seperate the good from the bad .

We are very lucky here in OZ with the high div paying stocks and tax laws that create " events " ... slight hint fwiw ... good luck to all

Quick addition with the right strategy avoiding the 50% dd with even just a breakeven (although i still think positive results are doable under these circumstances ) you are 100% ahead of market right there with the cash/ability to exploit to explosive runups that occur after such events , the 10-20% 6 week returns turn into 20-40% briefly ( year or 2 ) FWIW i blew up my account spectacularly during GFC , best thing that ever happened to me, lead me to where i am today , making sure that never happens again . Time in market is risk and while the market/stock is retracing/flat its time to do something else with that coin

You only want to be in the market when something happens , you learn that pretty quickly as a daytrader and it applies to ANY timeframe

I now have strategies that even in GFC returned very solid returns . these strats test positive back to 2000 with incredible expectancy and accuracy

 

[DeepState]

Cool.

Im all for small portfolios but in the case of say 2 stocks id much favour indice trading . My idea is a small universe to follow , know them well and deal with them as dynamics change , i primarily follow top 50 as they are stable , less prone to unexpected movement and they have very distinct cycles/patterns/price imbalances that have high probability . I look for what WILL happen , im not looking for lotto wins just a heap of 10-20% moves inside a basically 4-8 week time frame , i wont put all my money in a one position but i may put up to 30% depending whats available . I look for predictability , top 50s over shoot down and up , identifying when the downs and ups happen is the mean reversion bit , just look at say a 120 day rolling range in top 50 stocks , thats the available range to exploit

So, in significant part, does portfolio concentration arise from signals actually firing? For example, you might have 2-10 signals firing at any given time and hence the portfolio concentration varies. That's totally fine. However, does that mean that the portfolio invested amount varies with the number of signals as well? eg. you put less in the market when only two stocks are active than when 10 stocks are active? That would represent a very reasonable response to longitudinal signal construction of the type you have described (at least I think).

lets be clear im not saying 100% returns just > 100% indice outperformance with NO negative years and no leverage , with leverage that 100% return is doable but as we know that double edged sword is what explodes accounts . Can only speak for myself but the edges arent tiny

I presume you are long-only on stocks. Your time frame for reversion is much longer than I had imagined. It is normally extracted for periods of less than a month and largely much shorter than that. Either that, or it is long term reversion of the 3 year variety. Nonetheless, I am not questioning your alpha.

Dealing with risk is mentally easier with a smaller portfolio as you are basically forced to do it .

Can you please expand on this? This looks important to your perspective.

We are very lucky here in OZ with the high div paying stocks and tax laws that create " events " ... slight hint fwiw ... good luck to all

You need to watch for the offshore market activity now who don't get credits. They are increasingly crushing the trade before ex. That is their way of playing this arb.