Market Timing Strategy

[Lsbcal]

Quote:

Originally Posted by ERD50 True, they are different. But my point was that in any data set, one will find patterns if they look hard enough. Do those patterns have any predictive value, or is it just co-incidence? So does the approach mentioned in this thread really have predictive value, or is it just a back test which found patterns in the past data set? I don't know, but I've seen so much of the latter that I remain skeptical.

I agree that filters are not predictive. My feeling, which I cannot mathematically prove, is that they move you in the direction of the current momentum vector. As I understand it there has been some research on momentum by academia and it apparently exists. But is it exploitable? That momentum vector can change from month to month (I look at end of month data).

Again, one should design the filter with some reasonably sane criteria. Shoot for low frequency of trades and accept the Buy-hold default on very sharp market declines. It is nice to beat the market, but these techniques should be viewed primarily as risk reduction ones.

Quote:

Painful for B&H? The market was up for 1987 (6%) and 1998 (29%). If I could get the average of those going forward, then call me a masochist! Bring on THAT kind of pain! What would have happened if one tried to side-step those? They'd probably miss a run up. -ERD50

I'm assuming you went through those markets. I did and it was painful and scary too. Those did turn out to be rewarding for those years. What I was saying is that the methodology I would design would have you hold throughout those types of declines. It would not try to side step those extremely sharp drops.

The most well publicized methods like the 200 day moving average might have got you out in the 1987 crash (barely), but that method did not do well in the 1990's. See Siegel's analysis in Stocks for the Long Run.

I've done one study to look at the 1929, 1987 and current market (including inflation and dividends). The chart shows that the market movements were indistinguishable for many months. The 1929 initial drop was very sharp (they allowed extreme leveraging back then). The 1987 drop was also just about as extreme.

BTW, the last 10 trading days (2 weeks) was the 5th worst isolated drop since 1950.

I'll post that chart in a followup.

[Lsbcal]

Comparing some market declines. The peaks are lined up so the 1929, 1987 and 2007 peaks are coincident. Each dot on a colored line represents 1 month. The X-axis is for the 1929 decline line.

Notice how 1987 decline was pretty indistinguishable from 1929 decline until many months into the decline. Yet the outcomes were quite different. There was no major business decline in 1987.

This chart is not intended to predict anything moving forward in time. Just trying to show how difficult it is to separate the wheat from the chaff.

[BigNick]

Quote:

Originally Posted by utrecht OK, market timers. Here you go: News Headlines Time to buy.

Wow. Can there be a bigger example of BS than a system in which the decision to enter a market is dependent only on the internal history of that market ?

Specifically, if the whole market is overvalued by 300%, following that system will have you piling in when, after 200 days of steady dropping, it's only overvalued by 250%.

Of course, if enough people believe in it, then it may serve to create a bubble to allow some people to recover some of their losses, if they cashed in. But it doesn't tell us anything about the long term. (Frankly, I don't think that any market timing tealeaf analysis tells us anything, and certainly not beyond the end of the current week.)

[utrecht]

The guy didnt claim that the signal meant anything long term. He said the market would be higher both 2 weeks later and 3 months later. The last time this signal was hit was March 2, 2009 and the market bottomed 4 days later. If you happen to be someone who looks at market signals to do any short term trading, this might be of interest.

[GaryInCO]

Quote:

Originally Posted by Bestwifeever So where are you putting your investments?

Sorry, haven't been in for a while. I still have most of my money in stocks, though there's a lot of mining stocks, GLD, etc in there. I bought some SDS (-2x S&P500) about 10 days ago just as a hedge, and that's done a fabulous job of driving the market up.

For anyone interested: as I said before, the end-of-August signal couldn't switch to short because August started above the 12-month average. However so far the September bar is WELL below the average, which sits at about 127.50. SPY is currently at 117.85. So unless SPY increases at least 127.50/117.85 = 8%+ from here during September, the September bar will be entirely below the average -- and that would flip it into "bear" mode.

[GaryInCO]

If anyone is still watching: September has closed out, and the entire bar was below the 12-month average. This model has now officially switched to a "bear" mode. In the past 50 years a "short trade" in bear mode would have won only 1/3 of the time, but that's because many of those 50 years were during a hyper-bull market. In a bull market the "bear" signal might miss a little bit of the resumption of the bull move, but it tends to reduce the volatility of the portfolio. In a bear or sideways market, the "bear mode" can avoid significant drawdowns.

For anyone wondering, I've lightened up my stock positions considerably, dropping some stocks that had already lost quite a bit and I should have exited weeks ago. Many of my remaining stocks are metals and mining stocks, which are getting hit but should do OK if gold/silver wake up again, energy stocks, and high-dividend REIT stocks like NLY. I also have some SDS (-2x SPY) and FAZ (-3x financials) which are doing OK as the market tanks. Those are basically hedges on the rest of the portfolio.

I haven't been spending much time here lately -- too many other demands on my time. Now that the model has switched modes, I think I'll bow out. Cheers all.

[samclem]

I was finally woo'ed over to technical analysis and now my guru has abandoned me. I guess it's back to plain old rebalancing.

Gary, thanks for taking the time to post. Good luck with the system(s).

[REWahoo]

I'm confused - did the system work?

[ER Man]

For retirement accounts and very long term investing, I do NOT time the market. I stick with a plan.

For short-term investing, timing is important. You can use Options Trading also for that.

[clifp]

Quote:

Originally Posted by REWahoo I'm confused - did the system work?

I think we will know another 6 months or so. If the market continues down from here then the sell signal would have been a timely one. Obviously it would have been better if the signal had come at the end of Aug. but...

If other hand the market is up before we get the a buy signal than it would have cost you money.

[ER Man]

Anyone do Options Trading?

[Nords]

Quote:

Originally Posted by GaryInCO I haven't been spending much time here lately -- too many other demands on my time. Now that the model has switched modes, I think I'll bow out. Cheers all.

And... there goes another diehard market timer, in search of greener suckers pastures.

Somehow they never stick around to explore all the parameters of their models, let alone let Mark Hulbert review their tax returns. At least Dixonge had the guts to stick around and preside at the post-mortem.

Quote:

Originally Posted by REWahoo I'm confused - did the system work?

I think if it had worked then you'd have heard all about it.

When the market opens on Monday I need to take a good hard look at the prices on selling Berkshire puts. I've been "too busy" to get around to it, but maybe this could help pay for some of the home improvement project that we've been so busy with.

[target2019]

I am a participant and observer for about 5 years now. Before that time, I let it all ride, and cannot say for sure if that was better strategy for me than market timing, rebalancing, or whatever others may follow.

I read quite a bit, and also follow several discussions long-term on Morningstar.

My personal experience is that you can mitigate the downside by using timing signals. However, how do you get back in? That is vastly difficult if you have job(s), family, etc. It is not possible for many of us to successfully follow through with attention and execution when funds pierce a timing signal. If you execute a sell, you can miss a serious run-up.

Treating all investments as a single portfolio also presents challenges. In this month alone, there are two major changes to my 401(k) and wife's 403(b). These are structural, and for a period of time I am actually not permitted to make changes to a large portion of 401(k).

For now, I am quite content to buy low in my 401(k), keep reading, and perhaps use some technical indicators if I ever get to retire.

[Snidely Whiplash]

Quote:

Originally Posted by GaryInCO September has closed out, and the entire bar was below the 12-month average. This model has now officially switched to a "bear" mode.

I was intrigued following this thread and the strategy presented. Not to bust on GaryInCO but anyone who was tempted to give this system a whirl at the end of September really shot themselves in the foot by going into "bear" mode and missing this spectacular October rally.

[REWahoo]

Quote:

Originally Posted by Snidely Whiplash I was intrigued following this thread and the strategy presented. Not to bust on GaryInCO but anyone who was tempted to give this system a whirl at the end of September really shot themselves in the foot by going into "bear" mode and missing this spectacular October rally.

It has been almost a month since GaryInCO last posted and a couple of weeks since he last visited the board. If he follows the typical path of the folks who post here about their market timing system and run into problems, we will not hear from him again.

Should he prove to be one of the very few exceptions and post again it will be interesting to see what he has to say.

EDIT: There was a one month interval between his last posts, so maybe he'll be along soon for an update.

[urn2bfree]

I will just point out that my adjusted DCA strategy of buying a small amount every month but increasing it every time a month is down worked beautifully again. This ratchets up each and every down month. The longest down streak in history is 9 months. By that time the monthly investment gets pretty big compared to the base 9 months earlier, but then pays off bigger when the market turns positive as it always eventually does...and the you go back to the low base amount.

[haha]

Quote:

Originally Posted by Snidely Whiplash I was intrigued following this thread and the strategy presented. Not to bust on GaryInCO but anyone who was tempted to give this system a whirl at the end of September really shot themselves in the foot by going into "bear" mode and missing this spectacular October rally.

You are certainly corrct about what happened in this instance. But it really means nothing. Even quite valid systems will often give misleading signals. That is why they may stiill work. They do not work all the time, just well enough to beat buy and hold over time.

Ha

[arebelspy]

Quote:

Originally Posted by Snidely Whiplash I was intrigued following this thread and the strategy presented. Not to bust on GaryInCO but anyone who was tempted to give this system a whirl at the end of September really shot themselves in the foot by going into "bear" mode and missing this spectacular October rally.

Yeah, it seemed to switch really late (to bear when in bull) then only switch right before the rally. Ha makes a good point though, long term is the important results.

Quote:

Originally Posted by urn2bfree I will just point out that my adjusted DCA strategy of buying a small amount every month but increasing it every time a month is down worked beautifully again. This ratchets up each and every down month. The longest down streak in history is 9 months. By that time the monthly investment gets pretty big compared to the base 9 months earlier, but then pays off bigger when the market turns positive as it always eventually does...and the you go back to the low base amount.

Sounds like the martingale gambling "system."

[haha]

Quote:

Originally Posted by arebelspy Yeah, it seemed to switch really late (to bear when in bull) then only switch right before the rally. Ha makes a good point though, long term is the important results. Sounds like the martingale gambling "system."

You are correct. However one difference may be that stock market returns over a reasonable time period seem to be mean reverting, rather than random.

It is still a risky strategy unless the investor has an unlimited bankroll and unlimited nerve. Return to the mean strategies can get you killed. For example, one strategy that made good sense and held up for many years has been killed over the past few years. Anyone who doggedly persisted would be broke. That is natural gas futures vs WTI. Shale gas has turned that one into a big time loser.

Another example though less dramatic is WTI vs Brent.

Ha

[urn2bfree]

Not quite the same as martingdale which doubles every turn and is always random in a system with no memory. You dont have to double the bet, because you dont usually lose all your monthly input each month. Also the market does have memory- which is why you rarely have long streaks one way or another because of corrections, etc. Not the same as a random dice toss or roulette wheel which can go against you everytime forever. You do have to accept lower returns in up markets. You do have to have a steady money supply. So far never have their been mote than 9 down months in a row and those are rare and followed by big gains.