Keep it sweet and simple. That’s the fit-for-publishing version of KISS. And it is critical for being successful in the long term with quant investing. The biggest mistake I see new quant investors make is over complicating things. This just leads to failure. By focusing on a few high impact factors a quant strategy can still significantly outperform while being relatively easy to maintain. In today’s post I will present a KISS quant strategy has provided plenty of outperformance while being easy to maintain.
Before I dive into the KISS quant strategy let me throw some stats at you. I have a link to Portfolio123 in most of my quant posts and I am also part of their affiliate program. Of all the clicks on my blog over to P123 only 6% of people have chosen to sign up for a free trial. Of those 6% only 9% have actual signed up for a paying P123 membership. That is 0.5% of people who showed more that some passing interest in quant investing have actually committed to use the tool. Needless to sat the affiliate program has been a total bust for me. Now, this is just pure anecdotal data. I have a really small blog and P123 is not the easiest tool to use but it jives with other stats I have seen others quote about quant investing in general. A big part of the problem is people over complicate things. They try to shoot for the stars with super aggressive strategies with high failure rates. It doesn’t have to be like that. By focusing on a few big things you can get ‘enough’ outperformance from a quant strategy to make a huge difference to your wealth.
Here is an example KISS quant strategy. We start with the SP500 index of stocks. 50% of the portfolio is in a pure momentum strategy based on 6 month returns (10 stocks rebalanced every 4 weeks). The other 50% of the strategy is in the Value Composite strategy (VC2, 10 stocks, rebalanced every 4 weeks). A rebalancing between the momentum and value portions of the portfolio is done once a year. We also use a recession crash filter (SPY-UI) based on the unemployment rate. That’s it. What this strategy does is to focus on a few big areas of potential outperformance that have proven themselves over long historical periods.
- Equal weighting
- Concentration (10 stocks for each strategy)
- Momentum factor
- Value factor
- Recession filter
Here are the simulated results for this strategy going back to 1999.
Pretty powerful. Significantly higher returns and lower drawdowns than the SPY. No need to chase complex implementations of factors, or esoteric indicators, or rebalance the portfolio every day/week. Keep it simple. Maybe you’ll actually be able to stick with it.
Whatever your view of stock returns going forward even a few percentage points of outperformance and lower drawdowns consistently applied across a long period of time can be life changing. Most quant investors miss this. Break out Excel and model the impact of even a few percentage points of outperformance and lower drawdowns over a significant period of time. That’s the potential and the ultimate challenge.
In short, try hard not to over complicate quant investing. Focus on getting a few big things right and keep it sweet and simple.
17 Comments
Alex · August 2, 2017 at 6:17 am
Fabulous blog as usual!
So does this mean I can subscribe to your model on P123?
paul.novell@gmail.com · August 3, 2017 at 3:31 am
I have one model on P123. The SPY-UI model. It is free.
Paul
Tony · August 2, 2017 at 7:43 am
Hi Paul,
In regards to P123, I personally did the free trial awhile ago and found it very difficult to get the quant screens I wanted up and running, and even once they were setup I wasn’t sure if they were done correctly. I would happily purchase P123 through your link (as I’m sure others would) if there was some way around this.
Have you thought about offering some of the quant screens you discuss on this blog as a freebie for people who purchase P123 through your link? Or maybe you could do a tutorial series of blog posts on P123. I bet you would get a lot more sales that way.
paul.novell@gmail.com · August 3, 2017 at 3:32 am
Hey Tony, I’m thinking about some way to make quant portfolios easier and more accessible. Plan is to get the newsletter off the ground first then we’ll see.
Paul
Mike · August 2, 2017 at 9:16 am
Paul,
Are you applying the SPY-UI filter on each component (Mom, Val) or are you running it just on SPY where entire portfolio shifts to out of market asset on violation of trend filter? Thanks for all the thought provoking work you do.
paul.novell@gmail.com · August 3, 2017 at 3:33 am
Running it on the entire portfolio. A macro risk-on risk-off overlay.
Paul
Doug Nashif · August 2, 2017 at 11:42 am
Paul,
Just my 2 cents on Portfolio 123…although I’ve followed your articles for years now and would love to be a part of any quant strategy that you have developed I have found P123 confusing and a bit intimidating. I will take a look again and if I find it has improved (or my brain has sharpened) I would be very interested in investing in your strategy.
As a side note…in agreement with the previous poster, if there was someway to be a part of your strategy outside of P123 that would be preferable to me.
Thanks,
Doug
paul.novell@gmail.com · August 3, 2017 at 3:34 am
Thanks for the input Doug. P123 is not any easier…
I am noodling on something wrt to a quant product. Need to get my newsletter off the ground first then we’ll see.
Paul
Ryne · August 4, 2017 at 7:10 am
As with all your articles, it shows just how easy it can be to “beat the market” over time with a few rules in place. Could you clarify the rebalancing of value and momentum every 4 weeks? (e.i. Does this mean every four weeks you are scanning for the best 6 Mo momentum, as in selling stocks that are no longer in the top 10 and buying the new list of top 10?) and then when you say they are balanced annually is that just the total cash balance is rebalanced?
paul.novell@gmail.com · August 20, 2017 at 9:04 am
That hardest thing to do is to ‘beat the market’. Most will fail. Only less than 1% of people I think can do it over time. It’s easy to beat the market in a backtest. It’s entirely a different beast to do it real time over many years.
And yes, a 4 week rebalance means you are replacing stocks which are no longer in the top 10.
Paul
Ken · August 8, 2017 at 4:36 am
Paul,
Like you, I am a retired engineer and have been following this website for a few years. I greatly appreciate the work you put into this website. Like others, I found the learning curve at Portfolio123 too steep for me. However, I did click on your link to AllocateSmartly and am a paid and happy subscriber there.
Count me in for your KISS strategy. I would strongly consider becoming a paid subscriber to your proposed newsletter and/or KISS strategies.
On a side note, has the 1966 Safe Withdrawal Rate been quietly abandoned or replaced by a superior risk indicator?
Many thanks,
Ken
paul.novell@gmail.com · August 20, 2017 at 9:00 am
1966 SWRs are still one of my favorite portfolio choice tools. You can pretty much rank SWRs for different strategies by knowing annual returns and max drawdowns.
Paul
gary · August 8, 2017 at 8:31 am
Hi Paul,
I looked at Porfolio 123, but it seemed more difficult to be sure that I was getting everything that I needed for implementing the VC2 strategy, so opted for StockInvestor Pro instead, which is clunky, but I still found easier to be sure of what I was looking at. I don’t do much in the software itself – just get the data out and use my spreadsheet tools for processing. I’m sure I’ll revisit Pofolio 123 again though.
In terms of KISS and performance – isn’t it hard to judge strategies without testing multiple periods? Without a measure of performance likelihood I wouldn’t be able to tell whether or not extra effort above my super simple strategy (mostly just VC2 re-balanced annually) is worthwhile.
thanks again for a good post,
Gary
paul.novell@gmail.com · August 20, 2017 at 9:01 am
Yes, you need to look at multiple periods. In particular, base rates are very important. Testing back to 1999 is also not nearly enough. That’s why I always use quant strategies based on factors, primarily value and momentum, that have stood the test of time.
Paul
Doug Landry · August 9, 2017 at 4:35 am
As always, thanks for your insight. Like others, I’ve found P123 a little hard to manipulate. At least for me, Value Signals is much more intuitive (albeit not as powerful because it doesn’t allow for back-testing).
How do you think this portfolio would work if you expanded the investing universe from SP500 to something like market cap > 500 mil or something similar? Also, what do you think about adding one more step to your momentum portfolio: take top 5% of firms based on 6 month momentum, then rank by volatility (ie lower volatility as an indicator of a smoother ride per Wes Gray should enhance performance), then buy the top 10? Another perturbation to your value side could be a simple sort of Joel Greenblatt meets Wes Gray: pick your universe, take top5% based on EBIT/TEV, then rank the survivors based on a quality component (could be ROIC if you want to stay true to Greenblatt or Piotrowski F-score if you want to go a little closer to Wes Gray), then buy the top 10. Could also add an O’Shaunessey 6 month momentum to this. All of these would still be pretty simple and might add a performance boost. Wish I could back test to find out.
Also love the downside protection unemployment rate play. Out of curious ity, I forwarded it to Wes Gray. Alpha Architect’s current strategy is to use both an SMA and TMOM as a timing device. Wonder if he will decide to change that strategy, at least for his US funds, based on the research suggesting that the unemployment rate works MUCH better.
Thanks again. Please keep me informed re your newsletter.
Corey Philip · August 12, 2017 at 12:48 pm
With annual returns nearly 20% , this thing just reeks of a flop going forward.
paul.novell@gmail.com · August 20, 2017 at 8:58 am
Sure. Could be. Forward returns should be lower for all equity strategies given current valuations. But for it to underperform the market value and momentum would need to stop working all together. Possible but unlikely.
Paul
Comments are closed.