Post New

The no-new-highs-lately bull market signal has been around for a while now. It says you're still apparently in a bull market if there has been a fresh recent market high within the last 99 trading days. Since publication, the times marked "good" have been mostly and accounted for almost all of the market gains, while the rest of the time has gone pretty close to net nowhere. That's pretty much as hoped...it was intended mainly as a bull detector and never had much to say about the rest of the time which is a jumbled mix of good and bad.

However, the average forward return falls surprisingly steadily depending on how long it has been since a recent high. The 99 day cutoff is a little bit arbitrary.

As a little reminder of that, I just wanted to mention that
(a) The signal was long right through the crash of 1987, which took place only 38 trading days after the most recent high, and
(b) The S&P 500 is currently at 41 days and counting : )

Jim

Jim ..... No indicator will ever be perfect - as I wrote before - I really like DBE because - its an uncorrelated time dependent approach - as opposed to price ( Although the new High is price driven). It just sorta provides enough of a long noose for the market to start to hang itself. That being said amongst the other BCs - this one has the highest MDDs - for the reason you mentioned. It also of course was long thru COVID - any sudden event driven risk cant be handled because of the underlying "Bears have a way of working itself to form" assumption.

Here's the current status of the group of DBE indices I track ( This list was once generously shared by Zee)

(1) SP500 : 41 days
(2) NDX100 : 41 days
(3) COMPQX: : 83 days
(4) DJ-30 : 91 days
(5) NYSE : 94 days
(6) IXF-X : 95 days
(7) RUT-X : 97 days
(8) IWV ( Russell 3000) : 41 days
(9) IWB (Russell 1000) : 41 days
(10) SP100 (OEX-X) : 41 days


So the broader market (without the Mag 7 ish ,FANG etc) has been much weaker - with Dow, Small Caps, NYSE all poised to hit in next 10-15 days

Best
AC

That being said amongst the other BCs - this one has the highest MDDs - for the reason you mentioned.

Again, just to beat the dead horse, it's not actually intended as a long/short timing system, and it's not even that great at long/cash, which is how one would find a maximum drawdown. The two states aren't bull and bear, they are best interpreted as "bull market" and "no idea".

The main purpose is to identify those stretches that really smell like an ongoing bull market, as it makes sense to invest in those stretches in a different way. For example, faster timing systems are probably not worth the bother, and fairly aggressive quant screens and growth stocks will probably do well on average. During the other stretches, the recommendation has always been to use a faster and more discerning timing signal, or just be prepared for strong ups or downs.

Jim

And I'll repeat for new readers: the S&P hit a tiny, irrelevant new high in February. Irrelevant, except to reset the clock on this indicator to July.

On the Russell 2K, this signal will go bearish by Wednesday.
On the Nasdaq, this signal will go bearish in about 3 weeks.
On Emerging Markets repped by EMXC, it went bearish in December.

FC

And I'll repeat for new readers: the S&P hit a tiny, irrelevant new high in February.

I'm not following you here. The 99-day channel for the S&P500 had been working its way upward for about nine months. The bump in February was the 23rd in a row. Here is a chart:

https://stockcharts.com/h-sc/ui?s=$SPX&p...

DB2

Jim, IIRC in post #262358 on the old MI Board, you mentioned that perhaps 114 days was a slightly better indicator than 99 days. 99 days had the momentum and we are likely splitting hairs so that's the indicator the Board stuck with.

Given data has continued to fill in since then, do you think 99, 114, or some other number is best? I would guess that any number roughly in this range is within the error bars, but thought I'd ask.

Thank you for any insight.

Chip

Given data has continued to fill in since then, do you think 99, 114, or some other number is best? I would guess that any number roughly in this range is within the error bars, but thought I'd ask.


I am not Jim, but here is how I interpret it. If there has been a recent high lately you can have confidence that the current bull-market is in tact. The longer we go without a new high the more likely we are not in a bull-market. 99 Days was a reasonably long enough period to break the trend, but the difference between 99 and 114 is just adding *slightly* more confidence to the indicator.

As Jim has pointed out many times, its "bull" or "not sure" (not a strong indicator of a bear).

tecmo
...

Please don't get too attached to one specific number on Dying Bullish Euphoria lookback. It can vary by index/market used. The point is, after "about" 4 1/2 trading months if there hasn't been a new high, that is getting statistically close to a bear market indication.

The point was to avoid bailing on a bullish market too early. Zee back in the day determined for him that something like 150 days (another full 2 1/2 months) was the optimal lookback.

If you're looking for more background on the strengths and weaknesses of the 99D, please ask!

FC

Fair enough. My point was, consider:
1.) ONLY the S&P 500 out of the major indexes has made a recent new high (in February).
2.) The new high on Feb 19 was a single point that was .7% above the prior high a month before that 3.) the prior high late January was a single point that was .1% above the Dec 6 high after 6 weeks.

The earlier chart shows sequences of new highs in chunks daily, an advancing market.
The later phase of this shows choppiness, toppiness and two contextually meaningless, if numerically correct, highs over a period of 2 1/2 months.

Here is the max pain scenario of 99D/DBE extant, and why it cannot be used reliably by itself to preserve capital.

FC

Here is the max pain scenario of 99D/DBE extant, and why it cannot be used reliably by itself to preserve capital.

Both of the FRED data series are up YOY, thus disconfirming recession.
Latest data published on 4/16/2025
Next Release Date: May 15, 2025

Advance Real Retail and Food Services Sales (RRSFS)
and
Industrial Production: Total Index (INDPRO)

Grit teeth and carry on.

This may be an opportunity to do some tax loss harvesting. Pretty easy to do while essentially still keeping your position. SPY/VOO to VTI, then back in 30+ days. Etc.
Actually, pretty much any market index ETF to any other, since the same dozen stocks are the largest holding in just about every broad index.



Not to get all political, but reading a few of the message boards here, people who are normally calm, cool, and rational are tearing their hair out coming up with reasons that "this time it is different!!!"

"There is now an irrational idiot in the White House!"
Yes, for the 47'th time since the USA was born. Hah!

Fair enough. My point was, consider:
1.) ONLY the S&P 500 out of the major indexes has made a recent new high (in February).
2.) The new high on Feb 19 was a single point that was .7% above the prior high a month before that 3.) the prior high late January was a single point that was .1% above the Dec 6 high after 6 weeks.

The earlier chart shows sequences of new highs in chunks daily, an advancing market.
The later phase of this shows choppiness, toppiness and two contextually meaningless, if numerically correct, highs over a period of 2 1/2 months.

The "Bull Trend" can take some time to recover.

Example:

JUL 16 : 5640 (local top)
AUG 05 : 5170 (local bottom) (-9%)
AUG 30 : 5630 (end of the bull?)
SEP 19 : 5700 (bull trend holds)

Now, you could make an argument that the high of 6120 on Feb 19 is impossible to hit in the "next couple of months" - we are down 16% from those highs and the clock is ticking, which would signal the end of the bull trend. The "buy the dip" bounce from Mar 13 to Mar 25 (5510 to 5750), didn't really recover enough to give a lot of confidence either.

tecmo
...




tecmo
...

Advance Real Retail and Food Services Sales (RRSFS)
and
Industrial Production: Total Index (INDPRO)

These very lagging indicators will tell you there was a recession after we can all see it in the rear view mirror. The market is a leading indicator, and it's screaming Recession!

Elan

These very lagging indicators will tell you there was a recession after we can all see it in the rear view mirror. The market is a leading indicator, and it's screaming Recession!


Well, the FRED isn't yet calling a recession. https://fredhelp.stlouisfed.org/fred/data/understa...

"FRED uses business cycle turning points determined by the National Bureau of Economic Research (NBER) for recession shading on graphs."
"NBER defines recession as a "significant decline in economic activity that is spread across the economy and that lasts more than a few months"


As of the 4/16/2025 report, both the RRSFS and INDPRO indexes are up, so as of yet we are not in a recession.

You must not be hasty: "the stock market has predicted 13 out of the last 7 recessions.”

I truly don't get all these people running around screaming hair on fire and gloating that evil Recession is coming! Recession is coming!
The old TMF board was called UNEMOTIONAL Investing for a reason. Leave your emotions at the door and invest using cold logic without emotion.
Go with what the historical data shows, not with what your emotions are screaming to do. Fear will almost always be detrimental to investment success.



Anyway...
For investing purposes we don't really care about recessions, per se.
For timing purposes, what we care about is avoiding deep losses but not missing large gains.

The statistics show that the standard 200day/43week/10month simple moving average timing reduces the losses but also reduces the gains, while slightly improving the volatility metrics.

In the 2016 GTT paper, he analyzed a bunch of publicly available economy factors and determined that -- for purposes of timing -- the SMA sell signal should be ignored when both these two FRED indexes were positive year-over-year.
Doesn't matter if you call this "recession indicator" or not. Call it "SMA Sell Gate" if that makes you happier.

I call it "Disconfirm Recession", but that's just me. Ever since I went bald I don't have the ability to run around with my hair on fire. ;-)

So here is the statistics I have. I created this spreadsheet myself, using publicly available data.
IN periods include dividends. OUT periods get 1 year T-Bill interest.

Using SPX (S&P500 index) weekly, 2/5/1950 to 3/4/2024

Buy & Hold:
11.4%	CAGR
-51%	MaxDD
1.05	Sortino
15%	stdev

SPX with 43 week SMA:
9.2%	CAGR
-28%	MaxDD
1.18	Sortino
11%	stdev

Same, but with FRED indicators gating the sell signal:
11.5%	CAGR
-26%	MaxDD
1.87	Sortino
12%	stdev

So I grit my teeth and ignore my emotions and won't sell unless & until one of the FRED indexes turns south.

I haven't run it using 99-day-high timing instead of SMA, I don't have that data. From what I recall, all of these simple timing schemes give about the same overall result, so it probably doesn't matter much.

"There is now an irrational idiot in the White House!"
Yes, for the 47'th time since the USA was born. Hah!

Nope. This POTUS is most definitely different.

There is a reason historians ranked him near the bottom after his first term. Given his incompetence, it can only go down from here.

It's possible he will listen to reason from those smarter than he is, but that is not a trait that narcissists normally possess.

I am not Jim, but here is how I interpret it.

And me as well, techmo. I agree with everything you said.

With no datahelper, I could not go back and read the old post. At that time, the Board was still figuring out what period to use. It decided on 99 days. Jim expressed a preference for 114. That may have been because it gave fewer signals for statistically the same return.

In any case, I was just curious if he, or any one else, still tracked 114 days or any period other than 99.

Chip

I don't track it per se but I calculate the SP500 114–day date as 7/28/25.

:-)Shawn

I call it "Disconfirm Recession", but that's just me. Ever since I went bald I don't have the ability to run around with my hair on fire. ;-)

I still have hair, and it's not on fire. I've been selling very gradually and I've gone from 100% stocks to about 75% in my main portfolio in the last few months. I've sold most of my call options, and I'm not buying any new ones for now. So no, no hair on fire. Maybe, though, when you have no hair it's easier to bury your head in the sand. ;-)

Elan

"There is now an irrational idiot in the White House!"
Yes, for the 47'th time since the USA was born. Hah!

To keep it a tiny less political, let’s just call Trump a transformative president. But Ray, given the list of things I mentioned, wouldn’t you say he is more transformative than most of our recent presidents.
1. Shutting down USAID
2. Trying to close the Dept. of education
3. Cutting of funding to major universities.
4. Firing masses of Federal workers
5. Threatening to invade Greenland and Panama
6. Talking about making Canada the 51st state
7. Saying Ukraine invaded Russia when obviously that is not true
8. Imposing larger tariffs than the Smoot Hawley tariffs, but exempting Russia
9. Arresting international students
10. Talking about firing the Fed Chairman in the middle of his term

Please stop this thread right there. Move any policy "debate" to US Policy and get it gone from MI.

Please stop this thread right there. Move any policy "debate" to US Policy and get it gone from MI.

Probably a good idea and I appreciate the sentiment.

But what this administration is doing has a significant effect on world markets and so does relate to a person's approach to mechanical investing.

It's been interesting to me to see how people adjusting their investing decisions based on current politics.

Please stop this thread right there. Move any policy "debate" to US Policy and get it gone from MI.
But what this administration is doing....

The problem with drawing the line on politics at any level above zero on this board is that it will become polluted with politics--or snark that's passed off as politics--in short order and lose any connection with investing, mechanical or otherwise. It's only necessary to see the METAR board on the site from which we all migrated and the Macroeconomic Trends and Risks board on this site to see how bad it'll become.

Eric Hines

The problem with drawing the line on politics at any level above zero on this board is that it will become polluted with politics--or snark that's passed off as politics--in short order and lose any connection with investing, mechanical or otherwise.

Agreed. My beloved BRK.a board has quickly gotten infested with it too.

Yes politics has an effect on the markets, but a lot of people can't discuss it without getting emotional. That's why we have a board for politics. Please take it there.

...so does relate to a person's approach to mechanical investing

How so? I've seen mostly emotional responses self-justifying individuals' decisions to ignore backtests.

Which some often do. I did some selling at the bottom in 2020. When Jim called the bottom in 2009, I felt the market had further to drop. And there are other times I regret that I did not have the nerve to follow the data and thought it was different that time.

Chip

will become polluted with politics
.......
Agreed. My beloved BRK.a board has quickly gotten infested with it too.

EXACTLY! I just posted there and labeled my post "OT" as it is about Berkshire --- sarcasm (unfortunately I before repeatedly found many people don't notice/understand sarcasm), with the goal to maybe wake some up to what's going on there, which for me is horrible.

This board always was even more "pure". Please let it be that way.

I will respect the boards wishes and not discuss politics. I guess because it was allowed on the BRK board, I thought it would be ok here. Also, Ray and Ges were a little political.

Given data has continued to fill in since then, do you think 99, 114, or some other number is best? I would guess that any number roughly in this range is within the error bars, but thought I'd ask.

I wrote some python code about 6 years back to explore questions like this. You can find it here:
https://github.com/jcmiii/Dying-Bullish-Euphoria

I've also got some spreadsheets you can download there.

There are actually two parameters: M (looking for a new M-day high)and N (in the last N days). This program found that SPY returns are maximized for M in the range M=138-142, and N=214. If the algorithm is restricted to M=N my recollection is that 114 was best, but just by a hair. But in 6 years the optimal parameter values could have changed. And of course they change significantly when considering a different index (e.g. IWM, NAS).

John

But in 6 years the optimal parameter values could have changed.

If the optimal parameter values have changed, they were just tuned for the data. Which means they are useless going forward.




And of course they change significantly when considering a different index

Ditto.

If the optimal parameter values have changed, they were just tuned for the data. Which means they are useless going forward.

Your conclusion does not follow from your antecedent. It's certainly possible that the "99-day rule" becomes the "100-day rule" some years later, when optimized on additional data, and that both rules are useful.

The "mean January temperature" changes each year, based on additional data. But knowing the mean January temperature to date is not useless - it can guide me when getting dressed.


John

But in 6 years the optimal parameter values could have changed.
...
If the optimal parameter values have changed, they were just tuned for the data. Which means they are useless going forward.

It's not useful to look at short periods for a signal that triggers only once per year, but you can look at very long periods.

For relatively recent history, since the 1970s say, then a somewhat longer "timeout" on the signal like 115 or 120 days would have worked a bit better, as the key major tops have been pretty rounded except for the crash of '87. It would have avoided a couple of whipsaws. This was analyzed in some detail about 15 years ago.

But the original signal was built using data back to 1916 and 1930. The goal was extreme robustness for all kinds of market conditions. In older decades, markets were not just more volatile, but also more jagged and trendless. Being a bit more nimble worked a lot better. So the 99 day lookback was chosen as what would have worked best on the WHOLE period, sort of a compromise between the older jagged times and the recent trending times--after all, who knows what the market will look like in the next 20 years? Again, the idea being that for a signal this slow, there just aren't enough signals since the 1970s to believe that tuning to the smoothly trending markets since then would be a choice well supported by the data.

In fact, lots of different lookbacks still seem to add a lot of value. For example, "no new trailing-six-month high in the last 3 months" works pretty well for identifying stretches that are particularly clearly ongoing bulls, which is a "timeout" parameter of only 63 trading days.

Jim

So the broader market (without the Mag 7 ish ,FANG etc) has been much weaker - with Dow, Small Caps, NYSE all poised to hit in next 10-15 days

Are we there yet?

https://schrts.co/svSqNcSm

In my mind COMPQX is the swing factor/casting vote - if you believe that the ensemble of indices marginally betters the outcomes ( it does) ....

Because nowadays I only count the pool of 9 indices ( sans SP&100 which alongside the SP500 and the NDX to a large extent is all dominated by the same cap wtd stocks)

So COMPQ will put the voter to majority 5/9 - and its currently at 88.

Best

Are we there yet?
Yup. Bearish on the NYA.

Advance Real Retail and Food Services Sales (RRSFS)
and
Industrial Production: Total Index (INDPRO)

These very lagging indicators will tell you there was a recession after we can all see it in the rear view mirror. The market is a leading indicator, and it's screaming Recession!

Elan

Above was Elan's comment to Rayvt regarding turning off timing if no recession is indicated. I thought I would look at the Conference Board's Leading Economic Indicator to see if it was any help in turning off timing. I wasn't able to find anything, but perhaps others may have better luck.

The LEI is issued in about the 3rd week of the month and it constantly revises previous months. Some sites list the value, but they list the revised values so you have no way of knowing what was actually known on the day of the release. I went back and pulled all the press release back to 2001. If anyone is interested, I can send a file. It has either the current month and 3 previous months, or current month plus previous 2 months data plus the difference between current month and 6 months ago. The conference Board says that a recession is likely if the LEI declines by 2.1 over the previous 6 months and the diffusion index is 50 or less. I looked at 6 month changes and 1 and 2 month changes using QQQ or SPY with a 200d moving average check.

Send me an email if you would like a copy of the data that was obtained from these two sources.

https://www.prnewswire.com/news/the-conference-boa...

https://www.conference-board.org/pdf_free/economic...

Aussi

Post New

Mechanical Investing FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.