Post New

... for posting on Sep29 your precious major bottom indicator which was so incredibly spot on, calling the very lowest day!

Though on another board someone who's name I keep secret says 100% self-assured "There is no bottom indicator" I just sold Jan'25 $290 calls for $67.8 (+64%) and Jan'25 $300 one's for $61 (+63%). Unfortunately this means I can never visit Monaco again as after DLTR I am now even more indebted to you => 2 bottles of champagne now (minus a few glasses for some insignificant flaws :)

If you get bored managing only your own money and need something to keep you from sleepless roaming Monaco's streets at night ......

... for posting on Sep29 your precious major bottom indicator which was so incredibly spot on, calling the very lowest day!

I second that. I'm surprised how little attention that got given the epistemology:
1. He only made made one call (in the context of prominence / conviction). In other words, we are not data-mining too much over many other random suggestions.
2. The date selected was within one day from the actual very lowest quote.
3. Three months have passed, and the published date remained the market bottom each day of those 3 months.

Jim will be the first to highlight that the lowest quote can be revisited, and far deeper lows might be found (let's say, in the medium term, though unlikely very extreme-short-term nor in the long-term), but the point was after all to find a major bottom, not the lowest value in all future history.

So I congratulation mungofitch furthermore.

- Manlobbi

I prefer to think of it this way:

I didn't call the bottom.
I created a model which tries to do so.
Since then, I merely report what it says, and how it has done.
It's a subtle difference...


FWIW, a long time ago I also published a very simple market timing system that attempts to identify ongoing bull markets.
It tries to identify stretches that are relatively safe to go in the water, leaving other stretches as "not so sure".
Again, a subtle difference between "not so sure" and "the model thinks the market is going to fall" bearishness.
The most useful thing about the signal is that it doesn't change states very often.

For whatever it's worth, these are the results for SPY since the model was posted 14.3 years ago.

The model has been bullish 83.8% of the time.
CAGR for SPY during those days +12.92%/year rate, total return 326.3% not annualized.

"Not so sure", 16.2% of the time.
CAGR for SPY during those days -1.01%/year rate, total return -2.3% not annualized.

It has had slightly less than one signal per year.
(i.e., on average one bullish stretch plus one not-so-sure stretch about each two years)

It hasn't been bullish since 2022-05-26

If you are more of a Nasdaq person, the returns while this model was not bullish were very similar on QQQE (Nasdaq 100 equal weight) at -1%/year.
But the CAGR during the bullish stretches was 3%/year higher at a rate of 15.9%/year.

Bottom line: it's impossible to build a reliable market timing model.
However, it's not impossible to build a model that has statistically significant predictive power.

The trick is to find a way to use that information so that you get an edge from it on average over time,
while not taking on increased risk when it is sometimes inevitably wrong.
You can do very well with even slightly loaded dice, provided you don't bet the rent on every roll.

Jim

I didn't call the bottom.
I created a model which tries to do so.
Since then, I merely report what it says, and how it has done.
It's a subtle difference...

I understand the distinction. That would apply firmly also if you had published every single major bottom, but you happened to publish on a particular date (following a series of major bottoms).

On the other hand, given that you (1) made one prominent observation, then (2) the probability of that one observation being both within (i) 1 day of the major bottom, and (ii) surviving as the bottom for 3 months, I believe to be a fairly low probability. It wouldn't surprise me if it was less than 15% chance, even after being informed by the model and trusting the probabilities given by the model (I'm approximating given the number of outputs by the model).

The major bottom model for the whole series of outputs, I believe is good. But whether in category of luck or skill, the individual call that you made was a lot more than good.

- Manlobbi

<So I congratulation mungofitch furthermore.>

...and for sharing without expectation of gain.

- a rare combination of genius & grace -
:-)Shawn

It hasn't been bullish since 2022-05-26

As correct as that was from then until today let me please invert this (Munger, the Apostel of "Invert, always invert!", just had birthday): So it was bullish before that date.

To clarify: When did it went bullish? Was it bullish the whole time while the S&P from it's Dec'21 high of 4818 went down to around 4150 on that date (which would mean bullish during 5 months of the index going down --- apart from some weeks in March --- and losing -15%)?

As correct as that was from then until today let me please invert this (Munger, the Apostel of "Invert, always invert!", just had birthday): So it was bullish before that date.

Indeed.
If you're interested in looking it up, these are the signals. (at market close on the day shown)

2008 [wasn't bullish for quite a while, followed by...]
2009-06-01 Bullish
2010-09-15 End of bullish
2010-10-05 Bullish
2011-09-21 End of bullish
2012-01-10 Bullish
2015-10-09 End of bullish
2016-04-20 Bullish
2018-06-19 End of bullish
2018-07-13 Bullish
2019-02-14 End of bullish
2019-03-13 Bullish
2020-07-10 End of bullish
2020-07-20 Bullish
2022-05-26 End of bullish

It's not a great market timing signal.
But, it seems to achieve its goal: identifying the stretches which are relatively safe: it seems like a still-ongoing bull market.
It sorts out good times from relatively uncertain times.
I tend to invest in slightly different ways depending on its state.

This is using the signal exactly as described in this old post
http://www.datahelper.com/mi/search.phtml?nofool=y...

By construction, it is always bullish for around 5 months after any market top, so all the "end of bull" signals are late.
But rather surprisingly, the stock market is usually down only around 8% from its peak week by the time the "bullish" tagged stretch ends.

Jim

Great! Thank you. This table is more than I did hope for :-)

it seems to achieve its goal: identifying the stretches which are relatively safe
Looking at that table and an S&P chart I can't see that. Such a signal is useful only if it helps to avoid the times with the big drawdowns, right? Which requires that it does not wrongly identifies those as "relatively safe/bullish".

I tried to compare your table with an S&P chart. Apart from the current bear in that chart since the start of your table, 2009, I see 3 "biggies" where the S&P lost 15% or more:

- The middle of 2011 (+-3 months from the middle of the year)
- The second half of 2018
- And as the real "biggie" of course the start of 2020

But I can't see from the table that it would have avoided these stretches. As you say it's "bullish for around 5 months after any market top, so all the 'end of bull' signals are late" (Thanks for the link. So it's your famous 99 day rule :)

But as all bears that happened during the interval of your table (apart from the current one) were not longer than those 5 months I can't see what use it would have had since the start of your table in 2009 as it would have stayed "bullish" during the full time each of those 3 bears lasted.

Where am I wrong (bad flu since Jan 1; brain not working properly)?

But as all bears that happened during the interval of your table (apart from the current one) were not longer than those 5 months I can't see what use it would have had since the start of your table in 2009 as it would have stayed "bullish" during the full time each of those 3 bears lasted.


While it would have been nice to successfully dodge out and back in to the market just before the drop and in at the bottom, it turns out each of those drops were short-lived and if you just stayed the course with your investment plan you were back up to prior highs in about a year or less.

My biggest mistake in the market has been correctly reading the tea leaves in Feb 2020 that the market was in for a serious drop and selling out ... only to sit on the sidelines as the market came roaring back and raced higher and higher highs. I'd have been far better off sticking with my prior allocation.

Jim's indicator may be useful to sidestep some part of the longer grinding bear markets not the quick dips.

Jim's indicator may be useful to sidestep some part of the longer grinding bear markets not the quick dips.

That's actually my point, that I can't see how to assess that from his table starting 2009 as there were not a single one longer than 5 months (apart from the current one).

To assess that one would need a table with the signals going back much further --- but then the question would be whether what was valid 20, 30 years ago still is valid today.

That's actually my point, that I can't see how to assess that from his table starting 2009 as there were not a single one longer than 5 months (apart from the current one).
To assess that one would need a table with the signals going back much further --- but then the question would be whether what was valid 20, 30 years ago still is valid today.

Again, it's important to define "success" in an agreed way.
All of the broad US market returns, and more, occurred in the subset of the time tagged as bullish.
The *average* return during those times was good.
It's certainly true that individual bullish stretches had some unpleasant times, but overall you got all the returns in the intervals expected.
Being in the market the rest of the time was, on average, a low-return proposition. Negative, in this stretch.

On average through the years this signal is bullish 73.2% of the time.
In this post-publication era it has been bullish 83.8% of the time.
For a simple reason: we have seen mostly bull markets since 2008, which it seems to have noticed correctly.

It was originally tuned on data 1950-2008, then validated out of sample 1930-1949, then validated post discovery 2008-2022.

Can it be used as a trading system? Maybe. Long if the market appears bullish, cash otherwise.
If one were in the market only during the bullish times, the data suggest you'd have a portfolio with under 65% of the risk of buy-and-hold and returns about the same or a bit higher.
The risk metric I use ignores volatility. It considers only the chances of having a negative rolling 3 month stretch,
and the chances of having a rolling years stretch with a nominal return under 10%.
There's a squared penalty on bigger shortfalls below those thresholds, and more importance given to rolling years than to rolling quarters.
I don't use the signal this way, and don't advocate it, but it seems to work.
For example, in the data from 1930, the worst rolling two-year period buy-and-hold was a real total return of -35.9%/year compounded.
For the hypothetical long-or-cash person, the worst two years would have been -18.9%/year. Certainly bad, but better.

Jim

Bullish as of today?

Jeff

Bullish as of today?

Yup, my slow "99 day" bull market detector now says "bullish".

That doesn't mean it's right, of course.
But it's saying that, statistically speaking, we seem to be in a bull market for now and likely the next few months.
Big lasting drops shortly after fresh recent highs are quite rare. It takes a while for depression to set in.
Usually : )

This signal does not turn on a dime, it's more of a big-picture investing regime kind of thing.

I haven't really tested, and don't use, the same signal calculated from the S&P Equal Weight index.
But FWIW, that has been bullish since Jan 17 and is up 3.1% since then.
This tells us that the cap-weight index switching to bullish now isn't due to simply a few giant firms having a good week.
The reverse, in fact...the strength is somewhat broader than the cap-weight index would lead you to believe.


One last note: the (overcomplicated) version of the "99 day" signal that I watch personally is very slightly different from the standard one.
It hasn't gone bullish yet, but would with maybe another .5% rise in the S&P.

Jim

the "99 day" signal that I watch personally is very slightly different from the standard one.
It hasn't gone bullish yet, but would with maybe another .5% rise in the S&P.

Huh. S&P500 is +1.56% at 12:30 CST today.

Yahoo!! (??)

Jim: the "99 day" signal that I watch personally is very slightly different from the standard one.
It hasn't gone bullish yet, but would with maybe another .5% rise in the S&P.

Me: Huh. S&P500 is +1.56% at 12:30 CST today.


I was just listening to a clip of Fox Business which had Ken Fisher on.
2/2/2023 (https://video.foxbusiness.com/v/6319645714112#sp=s...)

FWIW, he said that we've been in a bull market since October.
Yahoo says the S&P500 is up 12.5% since Oct 3.
BRK-B up 13.1%.

FWIW, he said that we've been in a bull market since October.

Sure.
That's pretty clear, if you take into account either
(a) hindsight, or
(b) my major market bottom detector buy signal Sep 30 : )

So, if he said we're in a bull market since October, that's more impressive if he said it in October than if he said it in February.

The "99 day" bull market signal is very conservative, and signals pretty rarely.
All of its bullish signals are late...but they're conservative.
By the time this thing signals, it's pretty certain that the market is behaving in a bullish way, sufficiently so that (statistically) it's likely to go on for a while.

A bull market is more or less axiomatically a string of fresh [recent] market highs not too far apart.
If you've been seeing that effect lately then it's by definition a bull market, at least for the moment.
If you aren't, it might not be.
The less obvious extension of that:
The longer it has been since a fresh market high, the dodgier the average forward market returns get.
Check out the table in this old post
http://www.datahelper.com/mi/search.phtml?nofool=y...

Jim

Post New

Berkshire Hathaway FAQ

---

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