Market Timing Strategy

[eytonxav]

I know Ken Moraif and have a few friends that use his firm and have nothing but positives to report about the performance he has achieved for them over many years. Ken utilizes several model portfolios based on a client's individual circumstances and rebalances periodically to the model. I believe many of his clients are retirees and that most are on a 60/40 model. The market timing only relates to when Money Matters will reduce/increase equity allocation based on the the S&P moving above/below its 200 dma by X%, as well as a few other indicators. He is not out to make anyone rich and doesn't claim that, but he will get you a decent return while protecting your principal.

Some people need help with investing and staying disciplined and Ken is probably as good as it gets if you want an investment manager that will let you sleep at night. He is a very nice individaul, and I feel he genuinely cares about people, and supports many charitable causes. Just to be clear, I have no $s invested with his firm, as I am do it yourself type.

The other firm in the metroplex which I think is a decent one is AssetBuilder Inc. - Registered Investment Advisor which is affiliated with Scott Burns.

 

[GaryInCO]

I don't need luck (it works, and has worked for me/DW, over an extended period of time, over the last 30+ years).

And 20 of those 30 years encompassed the biggest bull market in history. Of COURSE it worked then. (I would genuinely be very interested to hear how you've done since 2000.)

If you believe that history will repeat, then B&H will probably work as well in the future as it did in the past. If you believe (as you might have guessed that I do ) that the next 30 years is unlikely to repeat that performance, then you might want to consider other options.

Anyhow, you come along and state you have a timing method that works. You came on a bit strong.

I guess the "delusional" comment was probably uncalled-for. Other than that I was just trying to say that what has worked in the past is not guaranteed to work in the future, and here's an option that I believe will help people respond to down markets and sideways markets. Then I did my best to document that with hard math.

So, while I also don't believe market timing works (but, for some reason, I still harbor a hope that it does work) what's my next step? What do I look for? But, keep it real simple: no head and shoulders/candlestick stuff. Full moon stuff I'm down with.

Hey, no fair! I'm not supposed to be infecting you guys any more!! Get thee behind me!

But, well, since you asked...

As Dora said, her moon comment was a joke. I actually know a few people who use moon cycles in their trading, and they claim it works. But that's a bit too voodoo for me.

If you believe the simple method I proposed might be worth investigating, then just follow the steps I outlined. Once a month, calculate the average of the last 12 monthly closes. If this month's low is above that average, switch to "bull market" mode. If this month's high is below the average, switch to "bear market" mode. That's it. No candlesticks or patterns. No moon cycles or chicken entrails.

Right now it's in "bull" mode. The average as of the end of July is 126.21 for SPY, and this months' High was 135.70. Until the monthly High crosses below the average, it stays in "bull" mode.

You decide what you do in bull and bear markets. Maybe you switch between SPY and tbills, maybe you adjust your AA, whatever works for you. This is just a tool to help you determine if it's a good time to increase your risk in the market.

Here's a comparison of two simpleminded approaches: "buy SPY and hold" vs. "buy SPY when in bull, hold cash when in bear." I didn't include dividends or taxes. Neither approach is what you'd do in reality but it gives you an idea how this compares.

As you can see, it fell behind a bit in the fire-breathing bull of the 90's. And as I said, if you expect that to happen again, then B&H works great. If not...

But look how it does when the market isn't going into orbit. It ducked nearly all of the big dumps in 2000-2003 and 2008-2009. According to investing gurus like Warren Buffet, rule #1 is "don't lose money." This helps you avoid losing money.

This works because the market is NOT random. It tends to trend -- meaning that it tends to do what it's been doing more often than not. By far the strongest trend has been up, but the down moves also persist pretty well. If it heads down, it tends to keep heading down for a while, and that's a good time to be more conservative. (Much of the underperformance of the late 80's was due to it getting caught by the 1987 crash. It took the crash losses, then said "time to get out" -- and the market instantly took off again. That's a pretty unusual event.)

Now as I said, I have no idea how rebalancing would affect the B&H results. I need to go study Bogleheads. Probably BH&R would greatly outperform the simple-minded "buy SPY and go away" example I show here. BUT doesn't BH&R work better if you have a feeling for the market direction? Might you be able to do better at BH&R if you rebalanced into a more conservative stance when the market is in "bear" mode?

This is just a tool. I'm not saying you should buy SPY and hold it, or buy SPY in "bull" mode. This is a green-light/red-light indicator to tell you when it's safer to be in the market, and when you might want to think about lightening your risk. Many people try to do that subjectively by watching the market, reading the papers, looking at the economic situation, etc. That's tough to do well. This is an objective method that has a very good track record.

Now if you'll excuse me, I have to go practice my flouncing.

 

[utrecht]

GaryinCo,

You said to calculate the monthly avg of SPY and compare it to the high or low depending on the situation. Are you using the avg of the closing price of SPY each month? If so, how many months back are you going? Or are you using the avg daily closing price of SPY for the present month?

 

[clifp]

If you believe the simple method I proposed might be worth investigating, then just follow the steps I outlined. Once a month, calculate the average of the last 12 monthly closes. If this month's low is above that average, switch to "bull market" mode. If this month's high is below the average, switch to "bear market" mode. That's it. No candlesticks or patterns. No moon cycles or chicken entrails.

Right now it's in "bull" mode. The average as of the end of July is 126.21 for SPY, and this months' High was 135.70. Until the monthly High crosses below the average, it stays in "bull" mode.

You decide what you do in bull and bear markets. Maybe you switch between SPY and tbills, maybe you adjust your AA, whatever works for you. This is just a tool to help you determine if it's a good time to increase your risk in the market.

Here's a comparison of two simpleminded approaches: "buy SPY and hold" vs. "buy SPY when in bull, hold cash when in bear." I didn't include dividends or taxes. Neither approach is what you'd do in reality but it gives you an idea how this compares.

I am a bit confused in the the Green chart, I thought you showed a sell signal around 4/11. Did the market reverse itself and go into bull mode again?

Since I too lazy to actually do the calculation, (and not entirely clear on the process) I would appreciate it if you would start a thread. Then each month or quarter or whatever post the average numbers. Obviously if there is a sell signal then please let us know. While I am not entirely sure how'd I'd use the information I do change AA on periodic basis depending on my propriety valuation calculation (aka SWAG scientific wild ass guess). It certainly would be a factor I'd consider.

Back in the end of 2007 early 2008, Running Man made a number of post about the problems with banks/mortgage etc. Since his posts contained a lot of good analysis (but were against conventional wisdom), when his prediction proved accurate I did not dismiss them as a example of the stop clock being right twice a two. Suffice to say if takes the time to say XYZ is trouble in the future I'll do more than just read his posts.

 

[GaryInCO]

You said to calculate the monthly avg of SPY and compare it to the high or low depending on the situation. Are you using the avg of the closing price of SPY each month? If so, how many months back are you going?

I said "Once a month, calculate the average of the last 12 monthly closes." So at the end of each month, add up the monthly (not daily) closes for the last 12 months and divide by 12.

E.g. next Friday, 7/29, will be the last trading day of the month. That weekend, go to the website below and add up the closing values from 8/31/10, 9/30/10, 10/29/10, ... , 6/30/11, and 7/29/11, and divide by 12. Then compare that result to the just-closed month's High and Low. That's all there is to it.

You can get the last 12 monthly closes, and the current month's H/L, HERE. Use the "Close" value, not the "Adj Close." Yahoo actually displays the first day of the month, but in reality they're showing the closing value of the last day of that month. So the value they call "June 1, 2011" is the closing value for the month ending 6/30/2011.

As Lsbcal correctly observed, you might get very slightly different results if you did it on the 1st, or 7th, or 15th, or whatever. But the overall results should be very similar. You want to use the High and Low of the previous month, ending on the day that you look at the Close, so it's easiest to use the monthly bars that Yahoo creates for you at the end of the month.

I am a bit confused in the the Green chart, I thought you showed a sell signal around 4/11. Did the market reverse itself and go into bull mode again?

Green chart?

It's been in "bull" mode since 10/1/09. See the chart in my first post.

Are you referring to the little dip near the end of the chart? That's when the market declined (and thus the value of the SPY position dropped) from 4/10 to 7/10.

Since I too lazy to actually do the calculation, (and not entirely clear on the process) I would appreciate it if you would start a thread.

The process is really trivial, as explained above. I'll be watching this and I'll try to remember to post a notice when it switches to "bear" mode.

 

[ERD50]

Here's a comparison of two simpleminded approaches: "buy SPY and hold" vs. "buy SPY when in bull, hold cash when in bear." I didn't include dividends or taxes. Neither approach is what you'd do in reality but it gives you an idea how this compares.

As you can see, it fell behind a bit in the fire-breathing bull of the 90's. And as I said, if you expect that to happen again, then B&H works great. If not...

But look how it does when the market isn't going into orbit. It ducked nearly all of the big dumps in 2000-2003 and 2008-2009. According to investing gurus like Warren Buffet, rule #1 is "don't lose money." This helps you avoid losing money.

Like clifp, I'm confused - help me to understand this. Seems odd to dismiss B&H because a ten year bull market had much to do with the results, when you are using a specific ten year period to measure your own proposal.

You say it 'ducked nearly all of the big dumps in 2000-2003' - OK, but it also started about 40% behind. So it seems that all that did was get you even again, and only for a very short time. Looks like the strategy then lagged behind for another 6 years. So roughly 3 years of over-performance since 1988 (hard to tell the lines prior, but I can see that the blue lagged for some if it) isn't exactly a big confidence builder for me.

The problem I have with such simple methods like this - what if it misses just one big move? You could be sitting on the sidelines for the next bull. It could very well just be 'data-mining' that such activity matched the cycles in the past. What makes one think the pattern will fit future scenarios enough to do better?

-ERD50

 

[clifp]

In the graph posted above there is a spot from a few months ago marked Sell #2, what is the significance of that point?

I thought that you needed to add up the daily averages rather than the monthly averages. You are correct adding up monthly averages is much simpler.

 

[GaryInCO]

Ah, I see. clifp, I just told it to close out its open trade at the end of the test so I had a record of the current open profit. Sorry for the confusion.

In reality the system is still in "bull" mode and will continue to be until the market turns down sharply, so the monthly High is lower than the yellow line (which is the 12-month average):

Does that help?

ERD50, you're absolutely right that it did fall behind a bit during the sharp bull moves. When the market turns down, it takes a little while to realize that, and suffers a bit of a drawdown as a result. More importantly, when the market turns up, it takes a little while to realize that, and it misses some of the recovery.

But look at the shape of the equity curve compared to B&H. It's steady growth with no significant drawdowns, which is a dramatic difference from B&H. I think most people would find that to be much less ulcer-inducing. Unless you get really lucky and choose to switch your retirement funds out of the market and into bonds just at the end of one of those gigantic bull moves (ooh noo! market timing!! ), the slightly-slower-but-steady growth is likely to work out better for you. Rule #1 is "don't lose money."

So yes, it can fall behind a bit in a huge roaring bull move. It more than makes it up if the market has any bumpy ups and downs. If you're certain the market is going straight up, then of course buy and hold forever is the smartest move. But the market *doesn't* go straight up forever, even in the best of times. It's my opinion that with the economic disaster we're staring at now, we're not looking at "best of times." I don't expect the market to go straight up. I think we'll see a lot of sideways like the last 10 years, and probably some significant down moves.

The problem I have with such simple methods like this - what if it misses just one big move?

It *can't* miss a big move. The math won't LET it. That's like asking "What if 2 + 2 is 5?? Then what??"

If the market turns up significantly, which I'd say is the definition of a "big move," then it passes the average and goes into "bull" mode. That's how the math works.

If the market goes into "bear" mode, then turns up and grinds up slowly at a percent or two a year, it might never move fast enough to switch this method into "bull" mode. (I don't see any cases of that in the last 90 years, but it's theoretically possible.) But "grinds up slowly" is not a "big move." If "grinds up slowly" is the best it can do, you're probably better off in a more conservative stance anyway. If all you do in "bear" mode is to rebalance your AA into a more conservative position, then you're not going to miss much if anything.

A bigger and more realistic concern is the "whipsaw" problem. What if the market goes up a little, switches into "bull" mode, then immediately (while you're still in a bullish position) turns down a little, causing losses before it switches into "bear" mode, then immediately (while you're still in a bearish position) turns up a little, so you miss some profits before switching into "bull" mode, etc? That can certainly happen, and the method can take some losses. From the mid-60's to the late 70's the market seesawed up, down, up, down, and this system made a few losing calls. Still, from 1964 to 1978 it made about 30% profit, while the S&P was up about 13%.

Or look at the Dow. This approach doesn't work quite as well on the DJIA as on the S&P, but I have more history on the Dow. From 1930 to 1949 the Dow went up and down, and at the end of 1949 it was down about 25%. This system made a few losing calls, getting fooled by the up/down motion, yet at the end of 1949 it had about a 44% gain.

It's not trying to predict the future -- except to say "what the market has been doing, it will probably keep doing for a while." It's not guessing when a big move will happen, so it can't guess wrong. WHEN the big move starts, it WILL get on board, pretty soon after the start. That's what the math does. 2 + 2 = 4, every time.

 

[samclem]

This is not a criticism, it is an observation.
When I see systems that depend on a moving average, I generally wonder why that specific moving average period was chosen. 200 days is a typical moving average used by "technical analysts." Why 200 days? Why not 180 days, or one year? Did 200 days produce the best results in the past? So, the present system we're discussing uses the previous 12 months--why was that period chosen? Because there's a clearly identifiable factor in stock prices based on 12 months (annual reports?), or simply because that is what backtesting showed to work. In the past. And that's the same past that we are now saying is very different from the future.

So, if future returns and the volatility of US equities are going to be so different in the future, on what basis should we hang our hat on 12 months being the right moving average period? That's not going to change, too? Why not a week, a month, or 3 months?

 

[Nords]

So, if future returns and the volatility of US equities are going to be so different in the future, on what basis should we hang our hat on 12 months being the right moving average period? That's not going to change, too? Why not a week, a month, or 3 months?

Gary Smith, one of the best market timers I've ever read, used to say "An indicator will work-- until it stops working."

He didn't pretend to know (or even care) why-- he'd just move on to another indicator.

 

[GaryInCO]

Very good question, samclem. As it happens I looked at backtest results to see how it behaved. Everything from 6 months to 40 months or more "worked." Everything from 9 to about 18 months worked about equally well. 12 months was near the middle of that range. I like that it lines up well with a fundamental thing like the length of a year. That makes intuitive sense.

Will a 12-month average continue to be the best? I can't guarantee that. Of course, I can't guarantee the stock market will open on Monday, either. I can just say that from 12 years of experience writing and trading systems like this, this looks pretty stable to me.

12 months actually wasn't the optimal value for the Dow. 5 makes a lot more money, and 15-16 had much less drawdown than 12. But 12 months does work fine. The method doesn't work as well on the Dow, and it doesn't beat B&H in 1960-2011 like S&P does, even on its optimal settings.

Nords, yup, nothing works forever. Even B&H.

 

[Coolius]

I have been following Gary's posts with great interest. As he correctly says, this is a tool to provide an indicator of general expected directions, not a hard signal to buy or sell.

Following his description of the calculation, I have developed an Excel spreadsheet to indicate whether this month's reading is in "BULL" or "BEAR" mode.

This is a very simple and basic spreadsheet, it's NOT a work of art ( so please excuse the lack of refinements ).

By plugging in the appropriate values , a simple "BULL MARKET" or "BEAR MARKET" indicator will appear on the Results line.

Enjoy!!


Update: Since this is my first attempt to attach a file, I don't know if I did it correctly. If not, can someone tell me how to do it? I would also be happy to send it as an attachment to a conventional email upon request ( just PM me ).

 

[Lsbcal]

... From the mid-60's to the late 70's the market seesawed up, down, up, down, and this system made a few losing calls. Still, from 1964 to 1978 it made about 30% profit, while the S&P was up about 13%.
...

For the period Jan 1964 to Jan 1978, the SP500 index was up around 23%. But with dividends the data I have shows it was up 92%. You can correct for the SP500 index dividends using Schiller's data.

I think a good way to present your data would be to show CAGR's for some non-overlapping periods (like 10 year increments) plus number of trades/year compared to buy-hold.

Regarding sharp declines: What about May 1940 invasion of France? What happened in October 1987 panic? How about the LTCM crisis in Aug, Sept 1998? Did the system revert to buy-hold when the declines were very sharp?

I don't mean to be overly critical, but these are questions I ponder about my own data analysis. Details, details .

 

[haha]

This is an example of the generic system called a filter. In the past they have tended not to work, because they need trends. Choppy markets whipsaw them.

However, in today's world where things go up largely because governments are throwing money at them, and markets go down because something causes this process to slow or cease temporarily, we may get more trends and filter methods may work better.

It does not attract me, because I think it is likely to be a distraction, but it could work.

Ha

 

[Aeowyn]

I have successfully timed the market this year by watching the threads when I see a " Whee " thread then I sell and wait till I hear doom & gloom and buy back . It's been 100% successful so far . Thanks forum members !

+1

I haven't used this forum in particular - but will certainly factor it in in the future.

I basically have a buy and hold mentality. But when I'm buying stock funds, I like to look for times to buy when people are freaked out... This time is different .... The sky really is falling... Worst recession since the great depression... It's not going to be a "V" shaped recovery, it's going to be a "W"....Here comes the second dip... It's not going to be a second dip, it's going to be a whole new recession... If bank XYZ goes down it will be catastrophic... If the automakers default, it will be catastrophic... If Greece defaults it will be catastrophic... If the US doesn't raise the debt ceiling it will be catastrophic.... (In the last few months, the markets seem to have cycled on and off worrying about several of these).

And when I'm selling stock funds (either because I want the money for a purchase or because a fund exceeds my target allocation), I look for times that market seems to be "happy".

This year there has been plenty of opportunity for buying and selling.

 

[rescueme]

And 20 of those 30 years encompassed the biggest bull market in history. Of COURSE it worked then. (I would genuinely be very interested to hear how you've done since 2000.)

I know that the argument is always that an “old-timer” (such as me) benefited from a “good run” in the market.

That’s true.

But looking at the long term (which also includes the perceived downturn of post 2K, as you alluded to), my personal returns have not been too shabby.

Looking strictly at my IRA account (rather than my 401(k)/rollover account, which has variable contributions over the years, and the account of which I use for retirement expenses over the last 4+ years), I show an XIRR return of +9.08% for the period of 1/1/1982 (or more specifically, the first business day of the year, when I always made my annual contribution) to 7/22/2011 – roughly 29.5+ years.

If I look at the returns of 1/1/2000 through 7/22/2011 (the period of which you speak), my XIRR return is +5.67%, just under the 6% long term target that I have set for myself.

The point is that while I may have benefited from a good market, my total returns also include the period of which you speak. While you may have only been in the market since 2000, your “total history” is only a speck of time within the total possible (universe) of your investing period of time.

While I have “history” to show where I’ve been, nobody knows what the next 50+ years will bring. Assuming you are much younger than me (I’m 63), the possibilities of the unknown future will weigh greatly on both our investment possibilities.

I started investing in 1982, at the age of 34 (when my defined benefit – e.g. pension, was eliminated). My retirement income plan goes to age 100 (regardless if I make it, or not). That’s a period of 66 years. Will I be invested “in the market” till I pass? Of course I will. Maybe not with an AA that was part of my earlier years, but still counting on the equity side to provide me with inflation protection over the long run.

The age old argument (as you have stated), is only made on a very short period of time. Nobody knows what the future will bring, and the possibility exists that you may have returns that may greatly exceed my “history”, in the future – long after I’ve been turned to ash. Nobody knows what the future will bring (except death, and unknown taxes, of course).

I stand by my previous statements on the subject. Short term investments (along with short term market changes), do not suit the long term investor. If you disagree, than so be it. Just don’t suggest that I am taking the “incorrect path” by my view, or try to suggest my “success” is just because of my investment scheme, over my lifetime, based upon my date of birth. The future has yet to be written.

Can you possibly gain in short term moves - within a short term market? Maybe. Can you possibly gain with short term moves over a long term market? I cannot be convinced of that "fact". Over a 66 year period, I rather like my "method" (if you can call it that)...

 

[GaryInCO]

For the period Jan 1964 to Jan 1978, the SP500 index was up around 23%. But with dividends the data I have shows it was up 92%. You can correct for the SP500 index dividends using Schiller's data.

I think a good way to present your data would be to show CAGR's for some non-overlapping periods (like 10 year increments) plus number of trades/year compared to buy-hold.

Yes, the B&H model really should include dividends for an accurate picture. Of course, the timing model also benefits from dividends during its "bull" period, and from tbill or other income during its "bear" periods. I don't have the time or data to do that comparison. I've already spent way too many hours on this. SOME of us still have to work for a living.

Regarding sharp declines: What about May 1940 invasion of France? What happened in October 1987 panic? How about the LTCM crisis in Aug, Sept 1998? Did the system revert to buy-hold when the declines were very sharp?

Between 8/1937 and 11/1942 the market went sideways to down. The system tried to go long 3 times, and got burned 3 times. In those 3 losing trades it lost 55.21 Dow points. During that same period the Dow lost 55.25 points -- almost exactly the same, by coincidence.

In 11/1942 the system went long again ("reverting to buy-hold") and held that for 4 years. After a couple of false starts in the sideways market of 1947-1949, it jumped on board again in 10/1949, stepped aside for a few months in late 1953, got on again in 1/1954, etc etc.

This is an example of the generic system called a filter. In the past they have tended not to work, because they need trends. Choppy markets whipsaw them.

Yes, that's exactly what happened e.g. in the 1937-1942 period. Note, though, that even with the whipsaws, in that case the system performed the same (in terms of Dow points gained or lost) as if you'd held on the whole time. So I guess you could say it "didn't work" during that choppy period -- but neither did B&H.

I mentioned that this system doesn't seem to work as well on the Dow as on the S&P. But from 1930 to today, the Dow has gained 12165.52 points -- and this system made 108717.47. Choppy markets and all, it did almost as well as hanging on the whole time.

And that's in a strong and pretty consistent up market, where B&H excels. In a sideways or down market, a green-light/red-light approach like this can save your bacon.

So, as I said: if you're certain the market will go straight up, then the smart thing to do is buy and hang on for the ride. If you think the market is likely to have significant sideways or down periods, IMHO the smart thing to do is step onto the sidewalk when the bus comes barreling down the street at you.

 

[GaryInCO]

If I look at the returns of 1/1/2000 through 7/22/2011 (the period of which you speak), my XIRR return is +5.67%, just under the 6% long term target that I have set for myself.

Nicely done. That's a very good return for a period when the market (as measured by the S&P500) has lost about 8.3%, or roughly 0.76% per year. You've beaten the market by over 6% per year for over 11 years. Most mutual fund managers would kill for a record like that.

Just don’t suggest that I am taking the “incorrect path” by my view, or try to suggest my “success” is just because of my investment scheme, over my lifetime, based upon my date of birth. The future has yet to be written.

The path you've taken has worked well for you. That was partly due to skill and good choices, as your returns since 2000 demonstrate. But you also admit you had the benefit of a pretty stiff tailwind for much of that time. That is certainly partly due to the time and place of your birth. You would likely be in a different situation if you had been a Japanese investor -- or even if you had built your retirement account in a market like we've seen for the last 10 years (when you made almost 6%/yr) instead of the market we saw in the 20 years before that (when you made something like 12-15%/yr).

You're absolutely right that the future is not yet written, and no one can know for sure what it holds. If the market continues to act like it did in the 20th century, the same techniques that worked well in the 20th century will continue to work well. If the future market looks more like the last 20 years in Japan, the techniques that worked well in the 20th century will not work as well in that more challenging market. I don't think anybody would dispute that.

I think the only question is whether a more nimble approach is a better strategy for an uncertain market environment. That's a decision each person has to make for themselves.

 

[rescueme]

That was partly due to skill...

I doubt it - I'm not that smart (but I did stay in a Holiday Inn Express last night) ...

Seriously, I did not look at the S&P500 as a measurement. I use the W5K as my "measure", along with the MCSI-EAFE benchmark.

Maybe that's why I did a bit better, as a "global investor" (before global was a bad name)...

 

[Lsbcal]

Yes, the B&H model really should include dividends for an accurate picture. Of course, the timing model also benefits from dividends during its "bull" period, and from tbill or other income during its "bear" periods. I don't have the time or data to do that comparison. I've already spent way too many hours on this. SOME of us still have to work for a living.

Being retired does give me some time to do this stuff. I appreciated what you've said recently as you've taken a lot of time with your analysis. However, if backtesting shows inaccuracies and long time periods of failure I'd take that as a warning that you need to revisit the subject and improve your methods over time. We have to be rigorous with ourselves about this stuff ... lest we loose our hard earned money. My take on this anyway, don't mean to be preachy.

Between 8/1937 and 11/1942 the market went sideways to down. The system tried to go long 3 times, and got burned 3 times. In those 3 losing trades it lost 55.21 Dow points. During that same period the Dow lost 55.25 points -- almost exactly the same, by coincidence.

In 11/1942 the system went long again ("reverting to buy-hold") and held that for 4 years. After a couple of false starts in the sideways market of 1947-1949, it jumped on board again in 10/1949, stepped aside for a few months in late 1953, got on again in 1/1954, etc etc.

Here is a Dow chart covering the lead up to war and the early WW2 period. I had wanted to study the performance of the popular 200 day moving average (magenta) and another moving average technique (50/200 day exponential moving average in yellow). The swings up and down in these colors represent buy/sells (when magenta is high we are in market, when low out of market). The 50/200 EMA had more manageable trades but missed the Sept 1939 rise (US market thought war would be good for business) and the May 1940 fall of France. Nasty whipsaws missing both a buy and a sell.

So far every trend following system I've analyzed, including some of my own that do not use straight moving averages, will miss these sharp moves. It's because as filters they will miss some of the high frequency stuff. The last decade has been an ideal one for slow trend timing. Who knows what the next will bring.

BTW, I do have a timing methodology so am not completely buy-hold or anti-timing oriented.

 

[easysurfer]

About buy and hold vs market timing:

Buy-And-Hold Investing Vs. Market Timing

Interesting, the part about, yes, there are proven successful market timers, but they aren't around for a long time. In otherwords, the difficulty for continued success.

Interesting stuff.

 

[Lsbcal]

...Interesting, the part about, yes, there are proven successful market timers, but they aren't around for a long time. In otherwords, the difficulty for continued success.

When I think of using timing it has as a goal to be almost buy-hold. So no more then one trade (buy or sell, not both) every four years or so. The more common approach is much more frequent trading.

 

[samclem]

However, if backtesting shows inaccuracies and long time periods of failure I'd take that as a warning that you need to revisit the subject and improve your methods over time. We have to be rigorous with ourselves about this stuff ... lest we loose our hard earned money.

Understood. But all this is logically inconsistent with the idea that "B&H (with rebalancing) won't work going forward, because the future will be different than what we've experienced." We can't build a system based on detailed study of the past and simultaneously argue that the past is not useful as a guide to the future.

 

[temsike]

Timing the stock market is like being a player in a casino. First, one must guess right twice. That's a 0.5*0.5=0.25 or a 1 in 4 chance of being correct. Then, one must pay commissions, spreads, and short-term capital gains taxes on every transaction.

I prefer to invest in the stock market as a casino owner which means buying and holding index ETFs or funds for the lowest cost and widest diversification possible.

The main consideration is asset allocation. It should be such that there are enough stocks so one eats OK and enough fixed income so one sleeps OK.

 

[Lsbcal]

Understood. But all this is logically inconsistent with the idea that "B&H (with rebalancing) won't work going forward, because the future will be different than what we've experienced." We can't build a system based on detailed study of the past and simultaneously argue that the past is not useful as a guide to the future.

I don't think there is an inconsistency. We have to recognize which market movements we can avoid and which ones we cannot. Ha mentioned filter theory. You can build a filter to pass some frequencies and not others. That is what market timers try to do. Unfortunately people get carried away with this and try to avoid sharp declines which cannot be easily done without missing sharp up moves.

I'm suggesting that to avoid a lot of whipsaws the only really practical timing approach tries to remove slow trending downward movements as we saw in markets like 1929-1932, 1973-74, 1981-82, 2001-2002, 2008-2009. Most of the time you should be in the market and only trying to avoid these slow downward trend markets.

The fall back is buy-hold which is a darn good strategy assuming we have historical returns like past decades and not a repeat of the 1930's or the 2000's. Probably a decent bet IMO.