Stocks A to Z / Stocks B / Berkshire Hathaway (BRK.A)
No. of Recommendations: 15
What with the current mentions of timing with the 325 day SMA it seems worthwhile to examine that using the timing spreadsheet I created a while back.
This is the S&P 500 (including dividends) from 1950 to 2024. This is weekly, not daily or monthly.
To avoid whiplash near the SMA crossover point, the sell signal is when the current price drops 1% below the SMA. The buy signal is at or above the SMA.
When OUT, the return is that of the 1 year T-bill.
Two interesting metrics that are not usually shown are the CAGR return when IN and when OUT. A positive return when OUT is money that is left on the table.
GTT is Growth and Trend Timing (do not sell if two FRED economy indicators are positive).
43 weeks ~= 10 months ~= 215 days
65 weeks ~= 15 months ~= 325 days
Buy & hold (no timing)
When IN When OUT
11.4% N/A 11.4% CAGR
-51% MaxDD
2/5/1950 1.05 Sortino
3/4/2024 15% stdev
===============================================
SMA 43 week (~200 days) -1% sell
When IN When OUT
13.3% 6.4% 9.5% CAGR
-24% MaxDD
2/5/1950 1.28 Sortino
3/4/2024 11% stdev
---------------
SMA 65 week (~325 days) -1% sell
When IN When OUT
13.1% 6.3% 9.7% CAGR
-26% MaxDD
2/5/1950 1.28 Sortino
3/4/2024 11% stdev
===============================================
GTT 43 week (~200 days) SMA -1% sell
When IN When OUT
13.9% -1.2% 11.6% CAGR
-26% MaxDD
2/5/1950 1.90 Sortino
3/4/2024 12% stdev
---------------
GTT 65 week (~325 days) SMA -1% sell
When IN When OUT
13.2% 2.6% 10.9% CAGR
-26% MaxDD
2/5/1950 1.52 Sortino
3/4/2024 12% stdev
No. of Recommendations: 2
Thanks.
I'm still getting the Norgate data but hope to use MI to do an analysis also of which MA is ideal and over what "mound of toast" range. My skills are limited so am using a geeky friend for the extraction. So far has Nasdaq100 to a date in 1993 and is trying to confirm why not all the way back to 1985.
No. of Recommendations: 5
Interesting. Based on your study, the only choice that beats B&H is GTT 43. It improved overall CAGR by 0.2%. Hardly worth the trouble IMO, if you can stand the higher volatility.
No. of Recommendations: 11
Based on your study, the only choice that beats B&H is GTT 43. It improved overall CAGR by 0.2%. Hardly worth the trouble IMO, if you can stand the higher volatility.
Correct.
But...
the purpose of timing is not to increase the CAGR -- although many people think (erroneously) that's the purpose. And then bail when it turns out to not improve returns.
No, the purpose of timing is to trim the drawdowns and reduce the volatility.
That GTT has better return than B&H is just a happy occurance.
Look at all those stardard deviations. B&H 15%, timing 12%.
And those Sortino Ratios. Which is a better metric than the popular Sharpe Ratio.
"The Sortino ratio measures an investment's return relative to its downside risk, offering a more precise risk-adjusted performance metric than the Sharpe ratio by only penalizing harmful volatility.
...
A higher ratio is better, with 1+ considered good and 2+ generally deemed excellent."
Actually, timing is not particularly beneficial in the accumulation years. The period when timing is mostly beneficial is in the retirement/withdrawal years. Because there is not enough time to recover from the loss periods.
No. of Recommendations: 5
This the count of trades and whipsaws of each backtest.
Trades is the number of buys plus sells.
A whipsaw is when the next week after a sell signal is a buy signal.
GTT has the effect of staying IN more of the time and greatly reducing the number of trades. As you would expect with a system which sometimes disallows a sell signal.
B&H
Trades-> 1 0 <-Whipsaws
In % -> 100%
======================================
43 week SMA -1%
Trades-> 167 28 <-Whipsaws
In % -> 72%
65 week SMA -1%
Trades-> 109 20 <-Whipsaws
In % -> 74%
======================================
43 wk GTT -1%
Trades-> 67 9 <-Whipsaws
In % -> 84%
65 week GTT -1%
Trades-> 45 1 -Whipsaws
In % -> 84%
No. of Recommendations: 2
RayVt
For GTT has the buy signal usually been initiated by the index going above the moving average or by the timing signal being turned off by the growth factors going positive?
Aussi
No. of Recommendations: 5
For GTT has the buy signal usually been initiated by the index going above the moving average or by the timing signal being turned off by the growth factors going positive?
The first.
This was one thing I examined in my spreadsheet.
Normal GTT:
When IN When OUT Row #s
13.9% -1.2% 11.6% CAGR
-26% MaxDD
1.90 Sortino
12% stdev
Trades-> 67 9 <-Whipsaws 15 <--Success
In % -> 84% 18 <-- Fail
45% pct
-------------------
Both BUY and SELL signals are gated by the FRED indicators:
12.8% 1.7% 11.2% CAGR
-26% MaxDD
1.50 Sortino
13% stdev
Trades-> 83 9 <-Whipsaws 15 <--Success
In % -> 87% 26 <-- Fail
37% pct
Success is the number of OUT periods where you ducked a loss.
Fail is the number of OUT periods where you missed out on a gain.
The SMA-only non-GTT timing this was 23 success and 60 fail. Only 28% of the out periods was a success.
The way GTT works is that you STAY INVESTED unless the FRED indicators signal a possible recession. I call this -- the two FRED indicators being positive -- "disconfirming a recession". Note that this is not saying anything about a recession, just "recession is unlikely right now."
As you can see from the data, simple SMA timing has a positive return when you are out of the market. The optimal would be the market showing a net loss in the periods where you are out. That way you are ducking a loss rather than ducking a (small) return.
The logic is:
* When the market is going up, you want to jump on. Period.
* If the market is going down, it might be just volatility.
* If it is just volatility, you want to stay in.
* Otherwise, if the market is down *AND* there is a potential recession, that's when you want to be out.
No. of Recommendations: 5
S&P club inclusion rules:
1. US domicile
2. >=10% of shares are publicly floated
3. Market cap >=$22,700,000,000 unadjusted
4. Listed i.e. not IPO for >=12 months
5. Annual dollar volume as % of float-adjust Market cap>=.75
6. >=250,000 shares traded in all months of last 6
7. GAAP net income from continuing operations must be positive for the most recent quarter
8. GAAP net income from continuing operations must be positive for the sum of the most recent four consecutive quarters
No. of Recommendations: 4
Based on your study, the only choice that beats B&H is GTT 43. It improved overall CAGR by 0.2%. Hardly worth the trouble IMO, if you can stand the higher volatility.
Correct.
But...
the purpose of timing is not to increase the CAGR -- although many people think (erroneously) that's the purpose. And then bail when it turns out to not improve returns.
No, the purpose of timing is to trim the drawdowns and reduce the volatility.
As I noted, if your goal is to reduce volatility, then it can make some sense. But I view that as an emotional crutch more than anything else.
Actually, timing is not particularly beneficial in the accumulation years. The period when timing is mostly beneficial is in the retirement/withdrawal years. Because there is not enough time to recover from the loss periods.
On the flip side, as people advance through retirement their desire and ability to constantly monitor the market and trade periodically is diminished due to deteriorating health. And at any age, timing has tax and friction consequences.
No. of Recommendations: 3
The sell signal is clear but for the buy signal it appears that there are 3 possibilities:
- Buy when the index closes above the moving average, regardless of the GTT signal.
- Even if the index closes above the MA, only buy once the GTT signal also turns positive.
- Buy when the GTT signal turns positive, even if the index is still below the MA.
I think you have covered 2 of the 3.
No. of Recommendations: 3
I don't think the second option Is viable.
GTT signal is positive, don't use timing. So when GTT signal is positive, you are long. If GTT signal is negative, you are long if above MA.
Aussi
No. of Recommendations: 8
The sell signal is clear but for the buy signal it appears that there are 3 possibilities:
Nope. I was as clear as I could be.
Buy when the index closes above the moving average. Period.Not much to be said about that. When the market is going up, you want to be in the market. Because, duh, it is going up.
Plenty of people have said that to win at investing just simply avoid losing. The big losses hurt you more than the big wins help you.
"not losing money ends up being far more powerful than making extra money in the long term."
"not losing money in bad times is more important than making it in good times."
"The moving average timing strategy makes the majority of its money by avoiding large, sustained market downturns. To be able to avoid those downturns, it has to accept a large number of small losses associated with switches that prove to be unnecessary. Numerically, more than 75% of all of MMA’s trades turn out to be losing trades." -
http://www.philosophicaleconomics.com/2016/01/movi...The logic behind the "G" of GTT is perhaps a bit hazy. It took me quite a while before I got a real understanding of it.
The times when it makes sense to be out of the market when it is NOT going up, is when the economy is not doing well. There are plenty of possible indicators to look at to decide if the economy is not doing well.
GTT as published uses two. Industrial Production and Retail Sales. If both are up year-over-year then the economy is doing ok. If either is down, then MAYBE the economy isn't doing so well. Importantly, these are not subject to possible political pressure, unlike employment/unemployment statistics.
What happens is that GTT keeps you in the market much more than simple SMA. as I posted, in 84% of the time vs. 72% for SMA. What GTT is going here is staying in the market and not "accept[ing] a large number of small losses associated with switches that prove to be unnecessary".
Using the same SMA parameters, since 1950 for the S&P500, for the periods that it was out of the market:
SMA had 23 successes and 60 fails
GTT had 15 successes and 18 fails
"Fail" is a switch that proved to be unnecessary.
No. of Recommendations: 6
- Even if the index closes above the MA, only buy once the GTT signal also turns positive.
I recall reading that the market generally turns up BEFORE the economic indicators do.
Also that the powerful returns after a bear market occur in the early stages of the subsequent recovery. If you miss that you leave a lot of money on the table.
"Prices don’t change when fundamentals change. Prices change when expectations and perceptions change. From a trend follower’s perspective, the main indicator that signals changes in expectations is price. "
"Prices change when expectations for future profits change. Expectations often change before fundamentals change." -- ivanhoff
Dang, this stuff is so complicated! If it didn't pay so well nobody would do it.
No. of Recommendations: 4
Buy when the index closes above the moving average. Period.
Yes, I understand your rule set, and it's probably the one that makes the most sense.
What I was wondering is if you tested the theoretically possible case that after a sell signal (when you are out of the market) you already buy if GTT turns positive while the market is still skeptical, i.e. stays below the moving average. Probably a rare event, and risky if there is a double dip recession for example.
No. of Recommendations: 6
What I was wondering is if you tested the theoretically possible case that after a sell signal (when you are out of the market) you already buy if GTT turns positive while the market is still skeptical, i.e. stays below the moving average. Probably a rare event...
Interesting supposition.
Seems unlikely if "Prices don’t change when fundamentals change. Prices change when expectations and perceptions change" is true.
Goes against the rationale of GTT also.
Which is "normal volatility sometimes tells you to exit when you should actually stay in".
It inverts some of that to be "get back IN early, even though the market is still in the crapper."
After thinking about it overnight I was able to tweak my backtest spreadsheet to examine that. After a couple of false starts. ;-(
From 1/1/1950 to 3/14/2024, weekly, 3873 weeks, there were 14 occurrances; IN for 153 more weeks.
Many more trades, 85 vs. 67
Lower CAGR 11.3% vs 11.6%
From 1/2/2000, 7.8% vs. 8.5%
No. of Recommendations: 3
From 1/1/1950 to 3/14/2024, weekly, 3873 weeks, there were 14 occurrances; IN for 153 more weeks.
Many more trades, 85 vs. 67
Lower CAGR 11.3% vs 11.6%
From 1/2/2000, 7.8% vs. 8.5%
Thanks for testing it.
No. of Recommendations: 2
Update 4/2/2026
What I was wondering is if you tested the theoretically possible case that after a sell signal (when you are out of the market) you already buy if GTT turns positive while the market is still skeptical, i.e. stays below the moving average. Probably a rare event...
Yeah, that didn't work out. It had worse results.
On the other hand....
I did another tweak to test the case where you were OUT (due to the normal GTT sell rule) AND both the FRED indicators have turned positive YOY,
AND (here's the tweak) the 10 week SMA was positive.
The idea here is that even though the 10 month SMA is still bad, maybe the 2 month SMA is showing that we are coming off a bottom.
There were a lot more trades, 83 vs. 51, also more whiplashes.
And about 1% higher CAGR. And higher Sortino Ratio.
