SPY Timing & Scaling System

[bikeeagle1]

Hi folks. I'm just starting this thread for those of us who try to time the overall market using ETFs like SPY. I have developed a system that I use in part of one of my accounts that has been working well, so I thought I would share it with anyone who is interested.

My goals with the system were:

1. To participate in longer-term market uptrends, but to step aside during downtrends (IOW, to engage in overall market timing), while at the same time,
2. To smooth out returns by scaling in and out as the market moves up and down along its uptrend path.

Another way to put that is that I wanted to combine value and momentum, but on an overall market basis rather than on a stock-by-stock basis as it is usually described.

So, while SPY is in an uptrend I want to be onboard, but I want to reduce or scale out some exposure as it runs up toward overbought conditions. Then, when it pulls back, I want to increase or scale in my position to take advantage of the next runup. Finally, when the short-term pullback turns into a longer-term drawdown, I want to get out completely and wait for the next longer-term uptrend.

I also want to do all of this without making any guesses, forecasting, following news, etc. I want it to be a purely mathematical system that could be traded by a computer. I also didn't want to have to monitor the markets every day, so this system ignores all intra-week data and only trades based on weekly bars. I try to follow it at the close on Fridays, but Mondays at the open works well too.

Anyway, without further adieu, here is where the system currently stands (I'll post every weekend with the latest signals, in the form of two questions with answers):

Should we be in or out of SPY? IN

Allocation? 25.8% Cash, 74.2% SPY

As with the usual disclaimers, I am posting this for informational purposes only. I am not a registered investment advisor. I'm just some guy on the internet. Invest at your own risk.

 

[1nvest]

Interesting, i do precisely the same! however not so much week to week, Im (as my name suggests) into much longer so would typically look month to month
Im 40% in SPY, 20% cash however i also suspect i might look at other asset classes perhaps not yourself such as commodities and precious metal
I'd be interested to know more about what method you're using to scale in/out. Momentum i'm a great fan of.

 

[Rufus_Leakey]

Im 40% in SPY, 20% cash

What is the other 40% in?

 

[1nvest]

What is the other 40% in?

commodities and precious metals. i would have hoped to have a bond element here, but that wont be happening anytime soon i suspect with interest rates

 

[bikeeagle1]

So just to clarify, the way I use this system is as a part of a broader blended portfolio. I keep about 33% of my IRA account in this system and the other 67% in the blended portfolio.

In the blend, I am always at a fixed static stock allocation of 40% stocks and a mix of bonds and other assets. Allocating 33% of the account total to this system, which scales from 0 to 100% stocks and the rest in cash, results in my overall stock allocation fluctuating from about 40% to about 60%. In other words, when SPY is in a bear market, I am 40% stocks. When SPY is uptrending but in a pullback, I could be as much as 60% stocks overall.

My goal with this thread is to isolate the SPY system and to forward-test its results by reporting the latest signals here. I am not advocating (nor am I) putting a whole account into this system.

Hope this clears up some questions. Thanks for any input.

 

[bikeeagle1]

At the close today, the system will be 6.7% Cash and 93.3% SPY. Since today is a Friday, I will start a model portfolio for tracking purposes. Since I want to be able to trade in and out of a meaningful number of shares, I will use $100,000 as a starting portfolio value. So, at the close, the order will be:
Buy 212 shares of SPY, Market-On-Close (MOC) order. Keep the remainder in cash. I will update this after I get the closing price.

Edit (after the close):
Fill price 437.98. This leaves the portfolio at $92,851 SPY, $7,149 Cash. As a benchmark, I will compare to a pure SPY portfolio. That would be 228 shares of SPY at 437.98 for a total of $99, 859 SPY and $141 Cash. Here they are in table form:

SPY System:
212 Shares SPY at $92,851
$7,149 Cash
Total $100,000

SPY Benchmark:
228 Shares SPY at $99,859
$141 Cash
Total $100,000

If we get an up week next week, I'll be selling some shares. If we get a down week, I'll be buying. If my downtrend signal triggers, I'll go to all cash and wait for a new uptrend. See you next week!

 

[bikeeagle1]

This morning, I realized that I have a way to compare how well this system has performed compared to SPY from one of my live trading accounts. I have been running this system all by itself (without the influence of the broader blend) in my Roth IRA. That gives me a way to check to see how it has been actually performing in the real world. So, I went back and looked at the first time SPY exceeded its current level, then compared how my Roth account has performed during the same time period. Here's what I found.

SPY's first weekly close above its current level was on 7/23/2021, when it closed at 439.94. With yesterday's close at 437.98, SPY is currently down by -0.45%. During that same time, my Roth account is up +1.34%. That means that the system has beaten SPY over that period by +1.79%.

Also, during this live test phase, I have only been running this system with half the account. The rest has been sitting in cash. So, if I had been running the system using the whole account (which I plan to do after the next exit and then re-entry signal), the account would have beaten SPY by double the +1.79%, or +3.58%. Since the time period involved happens to be exactly 6 months, that's an annualized beat rate of 2 X +3.58% or +7.16% (I know, that's not exactly how to calculate the actual annualized growth rate, but it's close enough for a quick comparison). Again, I'm not saying the account grew by +7.16% annualized, I am saying that it beat SPY by an annualized +7.16%, and with significantly less volatility. I don't know about you, but that's a "win-win" in my book!

Why was it able to beat SPY so well? Because by scaling out on the way up, I have reduced the impact of pullbacks. By scaling in on the way down, I have increased my participation in rallies. Less money in during pullbacks and more money in during rallies equals better returns and a smoother account curve.

I have to admit, I am encouraged by these results and I am pretty excited to see how it does over the longer term! Of course, during times when SPY runs up week after week, the system will lag behind because it will be scaling out. But, those are exactly the times when SPY is overdue for a short term pullback, which is what this system is designed to take advantage of. Where the system really shines is in reducing the volatility of the portfolio, and in making money during choppy conditions, as I have shown in this post.

 

[1nvest]

If we get an up week next week, I'll be selling some shares. If we get a down week, I'll be buying. If my downtrend signal triggers, I'll go to all cash and wait for a new uptrend. See you next week!

Mind if i ask your definition of an up/downtrend?

 

[bikeeagle1]

Mind if i ask your definition of an up/downtrend?

I'm sorry but I'm not sure I'm willing to share that yet, because of all the work and time I put into developing it. Also, I am still evaluating it and may make some changes as time goes on.

As a simple publicly available example though, one could use some form of closes above or below a major moving average, like the 10-month, for example. Here's a test of that system from portfoliovisualizer (this is based on trading monthly, and either all in or all out, depending on the signal with no scaling):

Test Tactical Allocation Models

Test tactical asset allocation models based on moving averages, relative strength, dual momentum, the Shiller PE ratio (PE10), and target volatility and compare results against different buy-and-hold portfolios

www.portfoliovisualizer.com

Honestly, I may evolve toward something that simple, and one of the things I am watching is to see how well my signal compares to that one. I don't know about you, but I would be pretty happy with my account tracing the blue line on that chart, but a little smoother thanks to the scaling.

 

[1nvest]

I'm sorry but I'm not sure I'm willing to share that yet, because of all the work and time I put into developing it. Also, I am still evaluating it and may make some changes as time goes on.

As a simple publicly available example though, one could use some form of closes above or below a major moving average, like the 10-month, for example. Here's a test of that system from portfoliovisualizer (this is based on trading monthly, and either all in or all out, depending on the signal with no scaling):

Test Tactical Allocation Models

Test tactical asset allocation models based on moving averages, relative strength, dual momentum, the Shiller PE ratio (PE10), and target volatility and compare results against different buy-and-hold portfolios

www.portfoliovisualizer.com

Honestly, I may evolve toward something that simple, and one of the things I am watching is to see how well my signal compares to that one. I don't know about you, but I would be pretty happy with my account tracing the blue line on that chart, but a little smoother thanks to the scaling.

yep, im aware of the 10mth/200 day used pretty much extensively by the likes of Meb Faber, Paul Novell and alot of the tactical allocation strategies within Allocate Smartly
If you'd rather not say, no problem

 

[bikeeagle1]

yep, im aware of the 10mth/200 day used pretty much extensively by the likes of Meb Faber, Paul Novell and alot of the tactical allocation strategies within Allocate Smartly
If you'd rather not say, no problem

It sounds like you and I are on the same page as far as who we are fans of. In fact, I am a subscriber to Allocate Smartly, and Paul Novell's SPY-COMP model is one of the indicators I use to determine whether SPY is in an up or down trend. I basically combine that with some other indicators to arrive at my overall signal.

 

[bikeeagle1]

Hi folks. Just checking back in to update. My SPY system has been 100% cash since the close on February 25th, 2022. I'm just wondering how many others have been out of the market for awhile and when did your "out" signal trigger?

 

[1nvest]

glad you're back...
as soon as it closed below the 10month, i was vastly out by then. Interesting Allocate Smartly's aggregate asset allocation (which i find a great barometer for my own allocation) still 15% in SPY. However i put that down to its mix of buy and hold strategies
as soon as the 12month ROC was negative, that was it for me, which is my last indicator
all weekly indicators by then had all given sell signals by week 7th march
unlike you, i scale in as we go higher, and scale out as we get lower

 

[bikeeagle1]

Thanks, 1nvest. I've had the same thoughts about the aggregate from AS. I like to watch it out of curiosity, along with their Meta strategy and also the SPY-COMP indicator. I've gotten comfortable enough with my own strategy though that I no longer trade based on any of the AS systems. I probably should just cancel my subscription to declutter my process, but it's too much fun reading about the new ideas they uncover.

 

[1nvest]

Thanks, 1nvest. I've had the same thoughts about the aggregate from AS. I like to watch it out of curiosity, along with their Meta strategy and also the SPY-COMP indicator. I've gotten comfortable enough with my own strategy though that I no longer trade based on any of the AS systems. I probably should just cancel my subscription to declutter my process, but it's too much fun reading about the new ideas they uncover.

same as you..i've already cancelled the subscription. Its just the aggregate which i find interesting. Indeed my own strategy has evolved over the years, but in no small way from the information from the site