Market Timing

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
Dillinger
Posts: 43
Joined: Sat Jan 27, 2018 5:45 pm

Re: Market Timing

Post by Dillinger »

Tom_T wrote: Tue Feb 06, 2018 12:20 pm Fall 2008: Dow drops from ~11,000 to ~8,800 in the span of under four months, with many wild swings along the way.
March 2009: Dow bottoms out at 6,594.
July 2009: Dow hits 9,093.
Dec 2009: Dow closes at 10,430.

Jan 26, 2018: Dow hits a high of 26,616.
Feb 5, 2018: Seven trading days later, Dow drops to 24,345.

I'd like to know what methodology would have told me when to buy and when to sell during this time. :)
I just provided you with one.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Market Timing

Post by willthrill81 »

wrongfunds wrote: Tue Feb 06, 2018 4:01 pm assuming market timing actually works, don't you still need to account for the time delta and the price differential between making the decision and executing the trade? once your algorithm decides to take the action on specific price point, the trade itself most likely does not execute at that exact price point.

if you are going to do backtesting, at least account for this and see if the numbers remain similar or not.
Any decent backtest will account for that. And if you're trading mutual funds like most are, there's always at least a one trading day between determining to make a trade and executing that trade. Further, some (including me) only trade once per month, using the closing price on the last day of the prior month. Over the long-term, results show that trading more often doesn't change the situation much either way if fairly long timing periods are used (e.g. 10 month moving average).
The Sensible Steward
Topic Author
Dillinger
Posts: 43
Joined: Sat Jan 27, 2018 5:45 pm

Re: Market Timing

Post by Dillinger »

PVW wrote: Tue Feb 06, 2018 12:20 pm This is the basic idea behind momentum investing. Apparently there is some reward to be had from going along with the crowd. Lots of info in this forum and on the internet. A good article is posted in the link below about the reality of momentum investing.
I'm not very familiar with momentum investing. But to the extent that it relies on following a general trend, it is similar. That being said, I would hasten to add that the system I provided in the OP is a long-term strategy, not a short-term one - although moving averages can be successfully employed in any time frame.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: Market Timing

Post by HomerJ »

Dillinger wrote: Tue Feb 06, 2018 4:10 pmalthough moving averages can be successfully employed in any time frame.
This is false of course.
Topic Author
Dillinger
Posts: 43
Joined: Sat Jan 27, 2018 5:45 pm

Re: Market Timing

Post by Dillinger »

bgf wrote: Tue Feb 06, 2018 12:23 pm back testing is a completely pointless exercise. 100% totally pointless.

or but yea, just go on portfoliovisualizer or whatever and backtest your way to :moneybag :moneybag :moneybag
Past returns certainly do not predict future returns. That being said, what is the rationale or basis for your particular investment strategy?
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Market Timing

Post by willthrill81 »

Dillinger wrote: Tue Feb 06, 2018 4:25 pm
bgf wrote: Tue Feb 06, 2018 12:23 pm back testing is a completely pointless exercise. 100% totally pointless.

or but yea, just go on portfoliovisualizer or whatever and backtest your way to :moneybag :moneybag :moneybag
Past returns certainly do not predict future returns. That being said, what is the rationale or basis for your particular investment strategy?
I dare say that there are close to zero investors with stock exposure that are not being influenced to some degree by historical data. It might be as simple as in determining how much they need to invest over time in order to 'likely' achieve their goals, but the impact is still there. I haven't seen one withdrawal strategy formed apart from backtested data.
The Sensible Steward
Topic Author
Dillinger
Posts: 43
Joined: Sat Jan 27, 2018 5:45 pm

Re: Market Timing

Post by Dillinger »

Fallible wrote: Tue Feb 06, 2018 1:59 pm In his book, The Devil's Financial Dictionary, WSJ columnist Jason Zweig says backtesting can provide important information on how an investment approach worked in the past, but that too often it is "designed to find any technique ... that beat the market over some period, no matter how unusual the circumstances." Here's his overall advice to investors, in particular those presented with backtesting results from their financial advisors: "If you don't ask tough questions about performance that was plucked out of the past, you are likely to end up blindsided by the returns you get in the future."
Anyone arguing for an asset allocation strategy relies on past performance to make his or her argument. (Anyone who is making an argument for any kind of investing strategy does. I would think it would be foolish to select an investment strategy without taking into account how such a strategy would have fared in the past.)
Topic Author
Dillinger
Posts: 43
Joined: Sat Jan 27, 2018 5:45 pm

Re: Market Timing

Post by Dillinger »

willthrill81 wrote: Tue Feb 06, 2018 4:29 pm I dare say that there are close to zero investors with stock exposure that are not being influenced to some degree by historical data. It might be as simple as in determining how much they need to invest over time in order to 'likely' achieve their goals, but the impact is still there. I haven't seen one withdrawal strategy formed apart from backtested data.
Exactly! What other kind of basis do we have for choosing one strategy over another?
Da5id
Posts: 5065
Joined: Fri Feb 26, 2016 7:20 am

Re: Market Timing

Post by Da5id »

willthrill81 wrote: Tue Feb 06, 2018 3:48 pm Most Bogleheads would agree that market timing is making buy/sell decisions in direct response to market performance. It need not necessarily be driven by a belief about future returns going one way or the other. For instance, those engaged in rules-based trend following can expect to make many trades, probably more than 50%, that will ultimately generate a lower return than buying-and-holding. But the end goal isn't to make the majority of your trades profitable compared to BAH; rather, it's to minimize downside risk, and the historical data on the topic show that that has indeed been the case.

I actually don't expect to generate higher absolute returns than BAH over the long-term, but I certainly expect to have similar absolute returns and substantially higher risk-adjusted returns to BAH. Even Larry Swedroe has said that he believes that trend following may generate better risk-adjusted returns than BAH.

Here are some of the peer-reviewed papers in the field. There are others.

https://papers.ssrn.com/sol3/papers.cfm ... id=2126478

https://papers.ssrn.com/sol3/papers.cfm ... id=2764933

http://mebfaber.com/wp-content/uploads/ ... 962461.pdf
Got it. Not a fan personally. I think simplicity of buy and hold has a real virtue even if it proves in retrospect not to be the most optimal strategy. Buy and hold requires less leaps of faith than more involved strategies. Less evaluation of competing (and changing) ways to beat the market by whatever criteria (risk/returns). If I ever wanted to go for more complexity, I'd probably just go for a multifactor fund (and Vanguard just trotted out one) and call it a day. I'd certainly not pick one of the plausible factors (time series momentum) to focus on.
User avatar
bottlecap
Posts: 6906
Joined: Tue Mar 06, 2007 10:21 pm
Location: Tennessee

Re: Market Timing

Post by bottlecap »

It's interesting to me that random people show up and espouse market-timing strategies on Bogleheads.org, of all places.

The arguments are all the same. I've done my homework, my testing shows, etc... No theories. No studies. And no proof, of course.

Who cares? Why do they care if people on Bogleheads.org agree with them?

Why not just do your market-timing and if it works out, great for you, then retire happily?

Market timing takes lots of time. Success if far from guaranteed and, in the majority of cases, just the opposite. And what is success? Beating an index strategy by 15 or 30 basis points?

Have fun. I'm going to spend that time on furthering my career or spending time with friends and family. Thanks to Bogleheads.org.

JT
bgf
Posts: 2085
Joined: Fri Nov 10, 2017 8:35 am

Re: Market Timing

Post by bgf »

Dillinger wrote: Tue Feb 06, 2018 4:55 pm
willthrill81 wrote: Tue Feb 06, 2018 4:29 pm I dare say that there are close to zero investors with stock exposure that are not being influenced to some degree by historical data. It might be as simple as in determining how much they need to invest over time in order to 'likely' achieve their goals, but the impact is still there. I haven't seen one withdrawal strategy formed apart from backtested data.
Exactly! What other kind of basis do we have for choosing one strategy over another?
unfortunately, there is some nuance required here.

there is a difference between

1) investing based on taking a specific data set, say SP500 stocks from 1990 to present, and then data mining for patterns that may or may not reproduce themselves, i.e., engaging in some form of statistical analysis, and

2) investing based on the thing itself. i invest in global equities because i want to own shares of companies all over the world. that is why i invest in stocks. not because in some prior period they returned 'x' with variance 'y' and sharpe ratio 'z'.
“TE OCCIDERE POSSUNT SED TE EDERE NON POSSUNT NEFAS EST"
CantPassAgain
Posts: 577
Joined: Fri Mar 15, 2013 8:49 pm

Re: Market Timing

Post by CantPassAgain »

bgf wrote: Tue Feb 06, 2018 5:15 pm
Dillinger wrote: Tue Feb 06, 2018 4:55 pm
willthrill81 wrote: Tue Feb 06, 2018 4:29 pm I dare say that there are close to zero investors with stock exposure that are not being influenced to some degree by historical data. It might be as simple as in determining how much they need to invest over time in order to 'likely' achieve their goals, but the impact is still there. I haven't seen one withdrawal strategy formed apart from backtested data.
Exactly! What other kind of basis do we have for choosing one strategy over another?
unfortunately, there is some nuance required here.

there is a difference between

1) investing based on taking a specific data set, say SP500 stocks from 1990 to present, and then data mining for patterns that may or may not reproduce themselves, i.e., engaging in some form of statistical analysis, and

2) investing based on the thing itself. i invest in global equities because i want to own shares of companies all over the world. that is why i invest in stocks. not because in some prior period they returned 'x' with variance 'y' and sharpe ratio 'z'.
Precisely. There is a big difference between understanding what it means to have ownership rights in a means of production or a legal claim on a debt note versus someone who simply looks for patterns in a dataset so they can jump in and out of asset classes like they are trading baseball cards.
LRonHalfelven
Posts: 87
Joined: Tue Sep 29, 2009 7:55 am

Re: Market Timing

Post by LRonHalfelven »

The OP's scheme does have one advantage which many similar ones lack: because it's based on such a limited period, there's a large amount of out-of-sample data (extending all the way back to the inception of the S&P 500 in the early 1950s) against which it can be tested. I haven't done so myself, but anyone thinking of trying this approach would be well-advised to do the test first.
"If you change your strategy frequently you don't really have one." --Garry Kasparov
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Market Timing

Post by willthrill81 »

bottlecap wrote: Tue Feb 06, 2018 5:10 pm It's interesting to me that random people show up and espouse market-timing strategies on Bogleheads.org, of all places.
Why not? With the exception of the market timing issue, I agree strongly with the Boglehead philosophy. And you won't hear me putting down buy-and-hold either. If one is truly capable of sticking with it through thick and thin, then it's a fine strategy. However, the recent movements in stocks and feedback from many (especially on the buy side) are evident that many who frequent this forum aren't completely sold on all aspects of the philosophy.
bottlecap wrote: Tue Feb 06, 2018 5:10 pmThe arguments are all the same. I've done my homework, my testing shows, etc... No theories. No studies. And no proof, of course.
There are actually numerous theories put forward by academics as to why trend following works. There have been many research papers on the topic; many here apparently just choose to ignore them for whatever reason. The following excerpt from a forthcoming paper in the Review of Finance is intriguing.
The evidence of time-series momentum in conventional asset classes directly challenges the random walk hypothesis and renders the theoretical background of market efficiency more puzzling. Besides, its high erturn premium in extreme market movements seems to contradict rational asset pricing explanations. The present findings are therefore more likely to support behavioral explanations, such as theories of sentiment, which further challenge the notion of efficient financial markets. The veracity of rational theories that explain time-series momentum should not be excluded, however, and may be fruitful subjects for future research.
And strictly speaking, it is impossible to prove that any theory is true. Without knowing the future in advance, no one knows what the optimum strategy will be going forward.
bottlecap wrote: Tue Feb 06, 2018 5:10 pmWho cares? Why do they care if people on Bogleheads.org agree with them?

Why not just do your market-timing and if it works out, great for you, then retire happily?
So is it "Bogle's way or the highway?"
bottlecap wrote: Tue Feb 06, 2018 5:10 pmMarket timing takes lots of time. Success if far from guaranteed and, in the majority of cases, just the opposite. And what is success? Beating an index strategy by 15 or 30 basis points?
With free software available today, one can easily employ trend following in under one minute per month.

There is potentially a lot at stake here. For instance, academic research has shown that, using decades of market data, trend following would have significantly increased the withdrawals that could be taken from a portfolio without prematurely exhausting it. I'd call that successful and worth achieving. For buy-and-hold, 2000-2009 was a 'lost decade', but simple trend following strategies achieved good returns during that period.
Last edited by willthrill81 on Tue Feb 06, 2018 6:52 pm, edited 1 time in total.
The Sensible Steward
Topic Author
Dillinger
Posts: 43
Joined: Sat Jan 27, 2018 5:45 pm

Re: Market Timing

Post by Dillinger »

willthrill81 wrote: Tue Feb 06, 2018 4:02 pm Both Paul Merriman and Meb Faber do this. Half of their portfolios are buy-and-hold and half are trend following.
I believe I read Merriman's "Market Timing with No-load Mutual Funds." (It was a long time ago.)
Fallible
Posts: 8798
Joined: Fri Nov 27, 2009 3:44 pm

Re: Market Timing

Post by Fallible »

Dillinger wrote: Tue Feb 06, 2018 4:34 pm
Fallible wrote: Tue Feb 06, 2018 1:59 pm In his book, The Devil's Financial Dictionary, WSJ columnist Jason Zweig says backtesting can provide important information on how an investment approach worked in the past, but that too often it is "designed to find any technique ... that beat the market over some period, no matter how unusual the circumstances." Here's his overall advice to investors, in particular those presented with backtesting results from their financial advisors: "If you don't ask tough questions about performance that was plucked out of the past, you are likely to end up blindsided by the returns you get in the future."
Anyone arguing for an asset allocation strategy relies on past performance to make his or her argument. (Anyone who is making an argument for any kind of investing strategy does. I would think it would be foolish to select an investment strategy without taking into account how such a strategy would have fared in the past.)
From my post, you can see that Zweig points out the upside ("important information") to backtesting - but also the downside depending on how it is done and how well it is done. The entire definition, including examples, needs to be read to fully understand his points about backtesting gone wrong.
"Yes, investing is simple. But it is not easy, for it requires discipline, patience, steadfastness, and that most uncommon of all gifts, common sense." ~Jack Bogle
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 2:17 pm
Location: USA
Contact:

Re: Market Timing

Post by willthrill81 »

Fallible wrote: Tue Feb 06, 2018 7:05 pm
Dillinger wrote: Tue Feb 06, 2018 4:34 pm
Fallible wrote: Tue Feb 06, 2018 1:59 pm In his book, The Devil's Financial Dictionary, WSJ columnist Jason Zweig says backtesting can provide important information on how an investment approach worked in the past, but that too often it is "designed to find any technique ... that beat the market over some period, no matter how unusual the circumstances." Here's his overall advice to investors, in particular those presented with backtesting results from their financial advisors: "If you don't ask tough questions about performance that was plucked out of the past, you are likely to end up blindsided by the returns you get in the future."
Anyone arguing for an asset allocation strategy relies on past performance to make his or her argument. (Anyone who is making an argument for any kind of investing strategy does. I would think it would be foolish to select an investment strategy without taking into account how such a strategy would have fared in the past.)
From my post, you can see that Zweig points out the upside ("important information") to backtesting - but also the downside depending on how it is done and how well it is done. The entire definition, including examples, needs to be read to fully understand his points about backtesting gone wrong.
The point remains, however, that there is no clear line in the sand as to 'how much' backtesting is needed and appropriate and when it is counterproductive.

On one hand, we can safely say that forming a strategy completely devoid of how that strategy fared in the past is probably unwise. Also unwise would be a strategy developed using 1 month of market data to optimize a portfolio around some goal. We probably all recognize that we need to be somewhere between those extremes, but precisely where we should be is debatable and probably largely dependent on the investor. For that reason, I do not deride those who adhere to buy-and-hold, but I also believe that those who do should not deride those who conscientiously choose another road to Dublin.
The Sensible Steward
Topic Author
Dillinger
Posts: 43
Joined: Sat Jan 27, 2018 5:45 pm

Re: Market Timing

Post by Dillinger »

wrongfunds wrote: Tue Feb 06, 2018 4:01 pm assuming market timing actually works, don't you still need to account for the time delta and the price differential between making the decision and executing the trade? once your algorithm decides to take the action on specific price point, the trade itself most likely does not execute at that exact price point.

if you are going to do backtesting, at least account for this and see if the numbers remain similar or not.
Buy or sell orders are placed after the market closes and before it opens again. So, there is plenty of time to place an order. And whether you buy or sell, your order will be executed at market price whenever the market opens again. (My input data includes both the opening price and the closing price for every trading day during the time period tested. So, whenever you get a buy signal (or a sell signal), you will buy (or sell) at the opening price of the next trading session. And, yes, my program most definitely takes that into account.)
Fallible
Posts: 8798
Joined: Fri Nov 27, 2009 3:44 pm

Re: Market Timing

Post by Fallible »

willthrill81 wrote: Tue Feb 06, 2018 7:11 pm
Fallible wrote: Tue Feb 06, 2018 7:05 pm
Dillinger wrote: Tue Feb 06, 2018 4:34 pm
Fallible wrote: Tue Feb 06, 2018 1:59 pm In his book, The Devil's Financial Dictionary, WSJ columnist Jason Zweig says backtesting can provide important information on how an investment approach worked in the past, but that too often it is "designed to find any technique ... that beat the market over some period, no matter how unusual the circumstances." Here's his overall advice to investors, in particular those presented with backtesting results from their financial advisors: "If you don't ask tough questions about performance that was plucked out of the past, you are likely to end up blindsided by the returns you get in the future."
Anyone arguing for an asset allocation strategy relies on past performance to make his or her argument. (Anyone who is making an argument for any kind of investing strategy does. I would think it would be foolish to select an investment strategy without taking into account how such a strategy would have fared in the past.)
From my post, you can see that Zweig points out the upside ("important information") to backtesting - but also the downside depending on how it is done and how well it is done. The entire definition, including examples, needs to be read to fully understand his points about backtesting gone wrong.
The point remains, however, that there is no clear line in the sand as to 'how much' backtesting is needed and appropriate and when it is counterproductive.

On one hand, we can safely say that forming a strategy completely devoid of how that strategy fared in the past is probably unwise. Also unwise would be a strategy developed using 1 month of market data to optimize a portfolio around some goal. We probably all recognize that we need to be somewhere between those extremes, but precisely where we should be is debatable and probably largely dependent on the investor. For that reason, I do not deride those who adhere to buy-and-hold, but I also believe that those who do should not deride those who conscientiously choose another road to Dublin.
My quoting from Zweig is not derisive but informative. Again, there are upsides and downsides to backtesting; Zweig has pointed to both and I've included both in my post and also advised reading from the book his entire definition, with examples of the downsides and advice for investors receiving backtest results from their advisors.
"Yes, investing is simple. But it is not easy, for it requires discipline, patience, steadfastness, and that most uncommon of all gifts, common sense." ~Jack Bogle
michaeljc70
Posts: 10843
Joined: Thu Oct 15, 2015 3:53 pm

Re: Market Timing

Post by michaeljc70 »

Dillinger wrote: Tue Feb 06, 2018 10:46 am
chevca wrote: Tue Feb 06, 2018 10:25 am You will? So, do you employ this strategy?
I will in one of my nontaxable accounts.
...
I conducted this study five years ago.
You figured out this system 5 years ago and still haven't employed it yet? What are you waiting for?
User avatar
LadyGeek
Site Admin
Posts: 95686
Joined: Sat Dec 20, 2008 4:34 pm
Location: Philadelphia
Contact:

Re: Market Timing

Post by LadyGeek »

Thread locked for moderator review.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.
User avatar
prudent
Moderator
Posts: 9085
Joined: Fri May 20, 2011 2:50 pm

Re: Market Timing

Post by prudent »

After review, the thread will remain locked.
Locked