No. of Recommendations: 20
Good question. The simple answer: It can be.

Over the life of this board an increasing amount of time and effort has gone into discussing how to implement mechanical timing that will improve risk-adjusted returns. The primary focus has been on layering on some sort of timing system on top of standard MI screens.

The focus of my own research has been on identifying cycles, and attempting to pinpoint transitions in the market. It has been a very long pursuit and countless hours have gone into it.

A little over a year ago I decided to take advantage of a third party tracker to validate my own signals. I'm sharing those with you today as an encouragement to those of you who have made a lot of effort to do something similar to what I'm doing. The point is simple -- it most definitely can be a very worthwhile pursuit. Yes, it takes a tremendous amount of time. And yes, it means that you'll run into a whole lot of brick walls as you strive toward the prize. There will be failures. But hopefully, the successes will counteract those failures and keep you pressing on.

The way timertrac works is that I have to enter the actual trades on their site and they execute the trade and store the historical record based on the next close price. If I enter it at night, which I usually do, they enter the trade at the next day's close. Actually, I tend to use a concept I call dream prices which are a limit order x% below the signal day's close. This is not possible to do on timertrac. So their approach does hurt returns somewhat.

The investments are all 3x ETFs, both bullish and inverted.

Anyway, without further ado, here is a link to one of the systems that I use:

http://www.timertrac.com/private/medallion.asp?mlid={DEAC7659-59A8-4AE4-846B-A2B440B26E8E}

It appears that the link will not work correctly so you will have to copy and paste it into your browser window if you are interested in seeing the graph.
Print the post Back To Top
No. of Recommendations: 27
A perhaps more interesting question is found by dividing things up a bit.

Is timing doable for Zeelotes? Sure. I don't imagine many of us doubt that.
Is timing doable for the median fellow at the board here? Less certain.

You make it look easy!
If that's not true, it's a dangerous mistake for us mere mortals to make.
The trades that you make done with the best signals I follow would probably not make money.

Jim
Print the post Back To Top
No. of Recommendations: 2
My own dabbling with timing has produced decidedly less than 120% return.
Print the post Back To Top
No. of Recommendations: 1
Are you curious about how many come to visit the MI board in a 24 hour period? A total of 150 unique IP addresses viewed the graph I shared in this thread since I posted it. It is probably safe to say that close to that number visit the board daily, at least in the middle of the week.

That is markedly less than the activity here just a few years ago. I recall I used to track this, but time constraints these days make that a near impossibility.

I wonder if the graph on:

http://www.datahelper.com/mi/search.phtml

Is anywhere near accurate? It sure doesn't seem to be.
Print the post Back To Top
No. of Recommendations: 41
Jim wrote:
Is timing doable for Zeelotes? Sure. I don't imagine many of us doubt that.
Is timing doable for the median fellow at the board here? Less certain.


Actually, since the middle of 2004 I've put a lot of time and effort into helping the members of this board as it relates to this question of timing. Granted, for the vast majority my success rate probably has not been all that high. But it has not been due to a lack of trying...

Maybe this will help. I've compiled a list of some of my timing-related posts and ranked them in their importance, as far as I see it. Those with a higher ranking almost certainly have some influence on the signals that you see above in the graph.

                    Title                         Date                 Data             Zee Rank                                                  Link
Trend Change Indicator #1: Market Breadth 7/30/2004 52Wk Highs-Lows 6 http://boards.fool.com/trend-change-indicator-1-market-bread...
Trend Change Indicator #2: Volatility - VIX 7/30/2004 VIX / VXO 4 http://boards.fool.com/trend-change-indicator-2-volatility-v...
Macroeconomic Indicator: Yield Curve 8/3/2004 Treasuries 10 http://boards.fool.com/macroeconomic-indicator-yield-curve-2...
Great Bull Markets & the P/E Ratio 8/14/2004 P/E Ratio 4 http://boards.fool.com/great-bull-markets-the-pe-ratio-21161...
Macroeconomic Indicator: Inflation Rate 8/18/2004 Inflation Rate 6 http://boards.fool.com/macroeconomic-indicator-inflation-rat...
Trend Change Indicator #3: Volatility -- VXN 8/19/2004 VXN 3 http://boards.fool.com/trend-change-indicator-3-volatility-v...
Executive Summary: Conservative Investments 2/6/2005 PE Ratio, Yields & DR 4 http://boards.fool.com/executive-summary-conservative-invest...
Elan's Asset Allocation Method Backtested 2/9/2005 PE Ratio & Yields 3 http://boards.fool.com/elans-asset-allocation-method-backtes...
Market Cycles & Asset Allocation Systems 2/13/2005 PE Ratio, Yields & DR 4 http://boards.fool.com/market-cycles-asset-allocation-system...
Backtest: MI Screens Using Technical Analysis 3/12/2005 Indexes 2 http://boards.fool.com/backtest-mi-screens-using-technical-a...
A Day of Pain or Opportunity? 10/6/2005 NYSE & Nasdaq Issues 8 http://boards.fool.com/a-day-of-pain-or-opportunity-23127768...
Identifying Market Tops & Bottoms 4/13/2006 52Wk Highs-Lows 10 http://boards.fool.com/identifying-market-tops-bottoms-23967...
A Flight to Junk? 4/27/2006 VL Fundamental Data 10 http://boards.fool.com/a-flight-to-junk-24025741.aspx?sort=w...
A Flight to Junk or a Flight to Safety? 10/6/2006 VL Fundamental Data 10 http://boards.fool.com/a-flight-to-junk-or-a-flight-to-safet...
Speculative - Defensive Indicator 10/11/2006 VL Fundamental Data 10 http://boards.fool.com/speculative-defensive-indicator-24685...
Nasdaq / NYSE Volume Indicator 1/30/2007 Indexes' Volume 2 http://boards.fool.com/nasdaq-nyse-volume-indicator-25107977...
A 90% Up Day Extreme Event 9/4/2007 Adv - Dec Data on Exchanges 9 http://boards.fool.com/a-90-up-day-extreme-event-25868925.as...
The Hindenburg Omen 10/19/2007 52Wk Highs-Lows 10 http://boards.fool.com/the-hindenburg-omen-26013312.aspx?sor...
Breadth Extremes -- Again! 12/18/2007 New Highs - New Lows 6 http://boards.fool.com/breadth-extremes-again-26202621.aspx?...
NYSE Bullish Percent Index 1/13/2008 $BPNYA from Stockcharts 5 http://boards.fool.com/nyse-bullish-percent-index-26277326.a...
Zee's Moose 1/29/2008 Price Data 8 http://boards.fool.com/zees-moose-26328915.aspx?sort=whole#2...
Gold and Riding the Wave 3/24/2008 Price Data 7 http://boards.fool.com/gold-and-riding-the-wave-26500070.asp...
Wave and DM Compared 3/30/2008 Price Data 5 http://boards.fool.com/wave-and-dm-compared-26518670.aspx?so...
Market musings on breadth and things... 4/2/2008 Index Issues 6 http://boards.fool.com/market-musings-on-breadth-and-things-...
The Best & Simplest Bear Catcher 10/24/2008 New Highs - New Lows 10 http://boards.fool.com/the-best-simplest-bear-catcher-271256...
NH-NL on Various Lists 5/10/2009 New Highs - New Lows 9 http://boards.fool.com/nh-nl-on-various-lists-27670733.aspx
NYSE NH-NL Great Depression 5/21/2009 New Highs - New Lows 9 http://boards.fool.com/nyse-nh-nl-great-depression-27694142....
SMA Slope -- Great Depression 5/21/2009 Index Price 5 http://boards.fool.com/sma-slope-great-depression-27696436.a...
Dying Bullish Euphoria -- Great Depression 5/22/2009 Index Price 5 http://boards.fool.com/dying-bullish-euphoria-great-depressi...

It is certainly true that to some degree many of the methods I've developed are a bit too complicated for the "median fellow" on the board here. But there are many indicators laid out in the posts above that are not all that difficult to follow. Probably the larger question is that most simply do not have the time or energy to pursue things that consume the amount of study and effort that some of my research entails. I can fully appreciate that. In fact, this is part of the reason why I've posted so little of my research on this forum since 2009.
Print the post Back To Top
No. of Recommendations: 0
Zee

Where you provide your ranking list, is it correct to assume that a rank of 1 is better than a rank of 12. (In your message you stated a higher rank was important, and I am guessing that means a lower number).

Also, is ther a particular reason there is no rank = 1?

Thanks

Craig
Print the post Back To Top
No. of Recommendations: 2
Is timing doable for Zeelotes? Sure. I don't imagine many of us doubt that.
Is timing doable for the median fellow at the board here? Less certain.


Timing is doable for everyone. Just like the old mantra of "follow the screen", it just now becomes "follow the timing model". Now how complicated the model (or screen) probably leads to the success in following it.

I "time" by using a 10 month moving average. A simple, blunt tool, that does the job I need it to do.

JLC
Print the post Back To Top
No. of Recommendations: 0
Zee

Where you provide your ranking list, is it correct to assume that a rank of 1 is better than a rank of 12.


I believe he meant the opposite.

StevnFool
Print the post Back To Top
No. of Recommendations: 3
Craig asked:
Where you provide your ranking list, is it correct to assume that a rank of 1 is better than a rank of 12.

10 = Best
1 = Worst

Anything ranked seven or above probably plays a crucial role in my evaluation of the market.
Print the post Back To Top
No. of Recommendations: 7
I "time" by using a 10 month moving average. A simple, blunt tool, that does the job I need it to do.

It has done the job since the Asian financial crisis, during the golden age of simple trend-following timing.
Any system looks like a genius if it said buy-sell-buy-sell-buy since the mid 1990s.
Are you happy with how it does in more normal more jagged/unpredictable times?
e.g., 1930-1997?
If you're happy with the typical rather than recent results, then that's great.

The long bull market ending in 2000 created millions of believers in buy and hold, buy on all dips.
The eerily smooth cycles since then have similarly created a horde of trend following market timers.
I fear that many of these people may be disappointed.

Jim
Print the post Back To Top
No. of Recommendations: 5
Zee,

I for one thank you for sharing all of this work over the years and
pushing the board into the examination and, for many, use of timing
methods.

Personally, I have replicated much of the research you have shared,
added my own tweaks on many approaches, and have developed a few
other unique timing methods. I use timing for a number of my investing
strategies and they have made and saved me a lot of money and let me
sleep better at night :)

Sincerely,

Wes
Print the post Back To Top
No. of Recommendations: 6
Any system looks like a genius if it said buy-sell-buy-sell-buy since the mid 1990s.
Are you happy with how it does in more normal more jagged/unpredictable times?
e.g., 1930-1997?


As it happens, I have daily SPX data from 1950. Not dividends, but VFINX includes dividends and goes back to 1987.

SPX from 01/03/50 thru 12/20/12 with simple 43-week timing:
B&H: CAGR: 7.31% stdev: 15.5%
SMA: CAGR: 7.39% stdev: 10.9%
Stdev when in: 12.6% when out: 21.9%

# trades: 40 # for gain: 24 # for loss: 16
Tot trades: 40 Avg gain: 14.18%
Win trades: 24 Avg gain: 26.75%
Loss trades: 16 Avg loss: -4.69%
CAGR: 7.39%
MaxDD: -23.6%
Avg # trades per year: 0.6
Avg # hold days: 429.
Hold 17163 of 22997 calendar days (75%)


Other time frames:

SPX from 01/03/50 thru 12/31/96
B&H: rate: 8.36% stdev: 13.1%
SMA: rate: 7.84% stdev: 10.8%
Stdev when in: 12.3% when out: 15.4%


and

SPX from 01/02/97 thru 12/20/12
B&H: rate: 4.20% stdev: 21.0%
SMA: rate: 5.50% stdev: 14.1%
Stdev when in: 16.3% when out: 30.6%


It's not about looking like a genius. It's about having roughly the same return but with substantially lower volatility, resulting in higher risk-adjusted return and smaller Max DrawDown.

Notice that the stdev when OUT is twice as large as the stdev when IN. When the market is gyrating like mad, you are sitting on the sideline.

VFINX (SPY) is similar:

VFINX from 03/27/87 thru 10/07/
B&H: rate: 7.92% stdev: 19.4%
SMA: rate: 8.52% stdev: 14.1%


People who (wrongly) think that timing is going to give them double return are going to be disappointed. And like always, that flock will then run around and jump on the next Hot Thing Of Today.
Those of us who understand the above statistics will be lying on the beach, sipping rum smoothies and watching others chasing after will-o'-the-wisps.
Print the post Back To Top
No. of Recommendations: 0
Zee

Thanks for clarification on the ranking system, and thanks for sharing your research.

Craig
Print the post Back To Top
No. of Recommendations: 6
It's not about looking like a genius. It's about having roughly the same
return but with substantially lower volatility, resulting in higher
risk-adjusted return and smaller Max DrawDown.


Good on ya!
Nothing about my comments should be construed to mean that I thought what
you were doing was inappropriate in any way. As mentioned,
If you're happy with the typical rather than recent results, then that's great.
Your testing is good and much more importantly your expectations are sane.

I worry more about the people who do 10 or 15 year backtests and think
it's going to give them the same lift in future that it did in this golden age.

Jim
Print the post Back To Top
No. of Recommendations: 1
Zee, it appears that the link you shared only works for those people who are paid subscribers to timertrac.

Would you perhaps have a chart or table available that you could post here or on your blog which shows the entry/exit dates for those of us without a subscription?
Print the post Back To Top
No. of Recommendations: 3
Knighted asked:
Zee, it appears that the link you shared only works for those people who are paid subscribers to timertrac.

Would you perhaps have a chart or table available that you could post here or on your blog which shows the entry/exit dates for those of us without a subscription?


I have Timertrac set to show trades with a one day delay, which means that if I allowed that link to continue to work someone would be able to pretty much follow my system using the link. For obvious reasons, I had to remove it.

Also, I cannot continue to share my trades with Timertrac for much more than 12 months. I typically stop after the 12 months have passed for the same reason as above. Since anyone on their site with a subscription could discover my system and begin to follow it.

I'll try to put a chart on my blog for those interested, but it will only show about 12 months as well.
Print the post Back To Top
Advertisement