Stoken Survival of The Fittest book Sims

Research
1

I was playing around with some simulations based a book I read a couple of years back. Dick Stoken’s 2011 book Survival of the Fittest for Investors ( http://www.amazon.com/Survival-Fittest-Investors-Darwin-2019s-Evolution/dp/0071782281 ). In the book, Stoken discusses the merits of combining negatively correlated assets, gets into the use of some tactical switching methods between these assets, and models these using data back to 1972. I am using these to try and better learn P123, put some of these out in the public sims in case anyone wanted to look at them.

One portfolio Stoken calls PCA - a Passive Combined Asset portfolio of equally weighted S&P, Real Estate, Gold, and Long Term Treasuries, 25% each, balanced yearly. (It seemed to me that it was really just a close cousin of Harry Browne’s Permanent Portfolio approach).

http://www.portfolio123.com/port_summary.jsp?portid=1205504

Another portfolio is ACA – Active Combined Asset portfolio – that switches on 1 year and 6 month breakouts between 3 sets of pairs – 1) S&P / Short-intermediate Treasuries, 2) Real Estate / Short-intermediate Treasuries, 3) Gold and Long Term Treasuries. 1/3 each pair.

http://www.portfolio123.com/port_summary.jsp?portid=1205426

I also took the PCA portfolio and added a simple Trend Filter to each of the assets, and used a 4 week rebalance. My attempt at trying to make the approach more of a Faber IVY type.

http://www.portfolio123.com/port_summary.jsp?portid=1205949

And then combined them in a book along with the Risk On Sim I had created the other day.

https://www.portfolio123.com/port_summary.jsp?portid=1205956

Nothing eye-popping, but steady return with modest drawdown with very liquid asset classes.

P123 is very good for rapidly creating models and seeing if the concepts hold up (during the new century)

–Tom C

2

Tom,

Your Sim, Stoken - PCA – Public is not selling any stocks. It is just rebalancing every year to equal weight. That is because you have Position Size set to 4 stocks and have 4 stocks in the ticker list. So it buys all 4 stocks every rebalance. I changed the Position size to buy 2 stocks so it would buy the 2 highest ranked each rebalance. The annual return was 12.56% & 16.64% DD & 100% winners.

Denny :sunglasses:

3

Denny, thanks, if I understand what you’re saying, according to the book, that’s all PCA is supposed to do. It’s just strategic allocation rebalanced between 4 asset classes. ACA is supposed to switch – aka simple tactical allocation. The PCA with trend I introduced was supposed to hold all 4 slots, but go to cash if the asset doesn’t meet the trend filter.

But what you did as a variation improves the return. That is interesting – I wouldn’t have known what to expect there, I chose the ranking system randomly, since the way I was running it rank didn’t matter because it wasn’t being used, just the Buy/Sell criteria. Another way I have tried back when I modelled some of this kind of stuff in ETF replay was to choose between the 4 asset classes and rank based on momentum and volatility. That always worked well.

Combining some of these methods seems to smooth the equity curve.

–Tom C

4

Denny, is your revised sim out there as public? Thanks.

–Tom C

5

Tom,

it is now: https://www.portfolio123.com/port_summary.jsp?portid=1205982

Denny :sunglasses:

6

The performance graph looks horrible for Stoken - PCA - Public (Denny’s Mod).
Here is the MAC-system performance for same period. CAGR and draw down are same as for Stoken.
What is more, MAC-system has been backtested for 65 years. So I would rather rely on the MAC.
http://www.advisorperspectives.com/dshort/guest/Georg-Vrba-140605-MAC-System-Backtest.php

MAC SPY-IEF.png

7

Georg,

I don’t understand your point. Tom proposed a few ideas for investigation, and all I did was make 1 change to add to the discussion. There was no time spent trying to make a great Sim.

Besides, the MAC systems has a lower annual gain, higher DD, and 50% higher turnover. What make my Sim look horrible relative to the MAC system? Oh, I know, you didn’t compare them on log scales. That would have shown my Sim has a smother curve. I know the MAC system was tested over many decades, but the purpose of this thread was to address some new ideas. Do you have anything positive to say about them?

Denny :sunglasses:

8

Georg,

Wouldn’t say horrible – it’s just another variation, based on some ideas in the popular literature – among non-professionals, like me… I like looking at all of the ideas out there, and seeing what is possible. And seeing if all the new subscribers to P123 can learn and whether they have something to add. And this is coming from one of your subscribers!

–Tom C

9

Excellent. Thanks Denny.

–Tom C

10

Guys, I did not want to upset anybody. But when you have only 17 trades from 2011 to 2014, you better have a backtest period longer than 14 years to support a model. Performance of the Stoken model looks “horrible” from 2011 to 2014. And look at the risk measurements of the Stoken model for the trailing 3 years - would anybody want to put money into such a system?

Risk Measurements - Trailing 3 Year
MAC SPY-IEF
Total Return (%) 60.86
Annualized Return (%) 17.17
Max Drawdown (%) -16.45

11

No, certainly not, I don’t trade it myself. But I use some of the concepts (adding non-correlated asset classes and tactical allocation). Stoken tested ACA from 1972-2010. It reduced drawdowns significantly while adding return to S&P buy and hold and 60/40 stock/bond. My goal was to try and model it in P123,and shared what came out of it.

–Tom C

12

Thank you for sharing, Tom.

13

THANK YOU .VERY REFRESHING TO SEE SOMETHING ELSE THAN R2G

14

When you look at something like the Permanent Portfolio fund (PRPFX), where they use their own similar strategic allocation, it certainly has underperformed for periods, just like the last three years, where buy and hold (and any strategy that is long equities) has ruled. But since its inception in 1984, has been a fairly safe place to park long term funds in a hands-off approach for the last 30 years. It’s added about 4% in alpha, while keeping drawdowns very reasonable. Whether or not you invest money in it, it’s a good launching point for investigating one’s own strategies.

http://www.permanentportfoliofunds.com/pdfs/perm/prpfx.pdf

–Tom C