Ray-You've posted about a strategy regarding pulling money out of an index when it drops below the 200 day SMA and buying back in. Perhaps I don't have your strategy correct correct, but I did a historical analysis on a similar strategy some time ago and it showed that the vast majority of the time, the buy in was at a higher price than the sell meaning that you would be better off holding most of the time. Could you elaborate on the strategy for the sake of discussion?-murray
Could you elaborate on the strategy for the sake of discussion?The benefit of a SMA filter is not to improve returns but to improve risk-adjusted returns. That is, you get slightly lower return but greatly reduced volatility and drawdown. Faber discusses this in his paper and on his blog.If you haven't done so already, download the spreadsheet I posted (S&P500 vs. IUL) and look at the SMA section.vast majority of the time, the buy in was at a higher price than the sell Yes. And in the vast minority of the time the buy in was much much lower than the sell. 18% lower for 12/2000 to 4/2002. Then IN for one month (which had a 1% loss) and then another 11% lower from 4/02 to 5/03.It's amazing how much better a portfolio will do if you side-step a couple of 25%-45% losses.It's amazing how much better you sleep when the market is in freefall if your portfolio is sitting in cash.You've committed the classic neophyte mistake of thinking that the # of wins vs. losses is important. It's not. What is important is the SIZE of wins vs. losses.That's why, BTW, most amateur covered call investors end up getting wiped out. They collect dozens and dozens of small wins, winning month after month for as much as a year or two. They think that they've found the route to free money. But then one large loss comes along and they lose as much in one trade as they made in 50 trades.So....95 times of 100 (or whatever the number is) you'd be slightly better off holding. 3 times of 100 you'd be about even. And 2 times out of 100 your face gets ripped off.Don't forget, too, the question of what to do with the money when the signal says to get out. In that 2000-2003 period, 1-year T-bills were yielding 6% to 1% (average 2.6%)
Thanks, Ray, for the deeper explanation. My analysis showed 3 times in the last 50 years that you would have had one of those major events. Granted, if I were on a fixed income, I would have wished I were out of the market. OTOH, looking at the day to day charts, it seems it would be quite easy to miss the precise sell & buy points and significantly sway the results. There's also the matter of taxes with all that buying and selling.Personally, I started moving out of stock funds at a measured pace in 2007. I've evolved my strategy into the following:When the major markets (US, Europe, Asia) are at 52 week highs AND I've not sold in 3 months, I shift 5% towards bonds/cash (from 100% stocks to 95/5%).When the major markets are at 26 week lows AND I've not bought in the last 3 months, I shift 10% towards bonds.I use 0% & 40% bonds/cash as my limits and simply rebalance if the buy/sell triggers hit.The major advantage I see is that I'm buying at lows and selling at highs, significantly more so than using the SMA and I should be reducing volatility. Granted, I did sell during the recent up swing of the market where I would have been better off holding, but SMA trading would have done the same thing.Thoughts?-murray
OTOH, looking at the day to day charts, it seems it would be quite easy to miss the precise sell & buy points and significantly sway the results.Nope. You don't have to be perfect, you only need to be Good Enough. The overall results are similar whether you do daily, weekly, or monthly. Daily is slightly the best, but has 4 times as many trades as weekly. There's also the matter of taxes with all that buying and selling.51 trades (25 round-trips) in 63 years isn't a lot of buying & selling.Personally, I started moving out of stock funds at a measured pace in 2007. I've evolved my strategy into the following: ... How does this backtest?That's the problem, isn't it? You can't backtest this, so all you have is personal anecdotal data over a limited period of time in one phase of the market.Humans are prone to a whole slew of cognitive biases & fallacies, and overweight narratives over actual data. There's been innumerable blog posts, papers, articles, and books written about these.The major advantage I see is that I'm buying at lows and selling at highsThe major problem I see with this meta-statement is encapsulated in the concept of "Would you rather be right or rich?"I wasn't going to toss in any quotes, but what the heck:"By thinking it’s all about strategy, a trader can get way too involved with each individual trades outcome, whereas in reality the success of the strategy relies on the edge provided over a large number of trades."Brian Lund, “The goal in trading is to make money, not to be a perfect trader.”"Looking at portfolios, think deeply about process over outcome. If you do something the right way enough times, you’ll win.—-Dan Loeb"Demosthenes (384-22 BC): “Nothing is easier than self-deceit. For what each man wishes, that he also believes to be true.” In other words, there’s a powerful tendency to believe that which could make one rich if it were true."The first principle is that you must not fool yourself—and you are the easiest person to fool." Richard FeynmanWith respect to your remark: I did sell during the recent up swing of the market where I would have been better off holding."You can’t judge the correctness of a decision from the outcome. This is a concept that many people find nonsensical. But good decisions fail to work all the time – just as bad ones lead to success – simply because it is impossible to predict which alternate history will materialize."
51 trades (25 round-trips) in 63 years isn't a lot of buying & selling.Perhaps I don't understand the strategy, I count 3 dips of the S&P below the 200 day SMA in the last 2 years alone.How does this backtest?That's the problem, isn't it? You can't backtest this, so all you have is personal anecdotal data over a limited period of time in one phase of the market.Why can't I back test it? The logic is simple to implement using historical data. In fact I did back test it and it did improve returns over a straight 60/40 balanced portfolio.I merely brought it up for discussion and possible analysis as well as trying to understand the SMA strategy you've mentioned.The major problem I see with this meta-statement is encapsulated in the concept of "Would you rather be right or rich?"By almost any measure, I'd consider myself "rich" already. I'm not overly concerned with being "right", I don't even think there is a single "right", but I do want to continue to learn.-murray
Perhaps I don't understand the strategy, I count 3 dips of the S&P below the 200 day SMA in the last 2 years alone.First off, I don't use 200-day SMA. I use 43 weeks or 10 months. I evaluate only weekly, not daily. For monthly, I do it on the last Friday of the month, not the last calendar day.And I use hystersis. Normally, my sell signal is a close 3% below the 43-week SMA. If you use 0%, then you get a flurry of signals right around the crossover point.Here's the counts I get, 1/1/50 to 3/15/13 (all round-trip trades):daily, +0% & -0% (buy & sell): 177weekly, +0 & -0: 95weekly, +0% & -3%: 39monthly, 0/0 and 0/-3: 47 & 45Why can't I back test it? The logic is simple to implement using historical data. In fact I did back test it and it did improve returns over a straight 60/40 balanced portfolio.Maybe you can backtest it. I looked at your criteria and thought it had too many subjective pieces. The market does what the market does, and whether or not you've bought or sold anything in the last 3 months doesn't enter into the picture.I just has the feel of what I was attempting to do many years ago. I eventually discarded it because it was just too subjective. I generally try to write down the concrete steps of any method I'm considering, as if I was going to write a program to implement it. That's a great way to find out if there are any holes or handwaving that you didn't spot at first. Often a narrative description seems clear, but then you write it out and discover that there are either hidden assumptions or something missing, or something that can't actually be done.Some questions that came to mind: What if the US is at a 52-week high but Asia isn't? What if Asia then hits a high but US has dropped below the high? If you insist that all 3 must hit a high (or low) simultaneously, you might *never* get a signal. So you have to shade it, and say to yourself, "Well, US is at a high and Asia was at a high just a few weeks ago and it looks like Europe is going to hit a high any day now, so I'll call a signal." So instead of a bright-line rule, you are back to just deciding by your gut feel.I know that many people like to shade their allocation as you mentioned, by slowly shifting to higher (or lower) bond allocations. That's always struck me as being too hesitant, and lacking conviction. Works for some people, but I'm not of that mind.I've been playing around with investigating various timing signals & methodologies for several years now. After awhile it came clear that it was an "indifference of the indicators" type of thing. What you are trying to determine is if the market is going up or going down. Or going sideways -- but if it's going sideways (that is, nowhere) then it doesn't matter if you are in or out. All those fancy methods were just different ways to measure if the market was going up or down. So it didn't matter which method you used, because they all said essentially the same thing. What I usually see when looking at different signals is that the timing differs by a week or two between one signal and another. The difference in overall performance is just the luck of the draw of happening to catch (or miss) a particularly good/bad week right at the transition point.Roger Nusbaum:"I don't really think it matters which trigger is used as no single trigger can be the best for all times but they can be effective which is the priority as I see it. Here effective is simply defined as avoiding the full brunt of a large decline. Aside from my belief in its effectiveness, the 200 DMA is simple to explain and understand.No one rule is always correct. they all give false signals."Simple is better, because simple is more robust.FWIW, Faber claimed that the market was more volatile when it was under the 200dSMA, and that most of the best & worst days happened then as well. Early on I backtested that, and confirmed that it was indeed the case -- so I don't bother to look at it anymore.And since the most important thing is to avoid large drawdowns, I don't particularly care to try to fine-tune the signals. My stats show for SPX that with the 43-week SMA +0%/-3% trigger:Avg buy was 7% above the bottom.Avg sell was 7% below the top.And that's good enough for me.'course, I don't use any of this to invest in the S&P500. I use other screens & strategies, and just use the SMA signals to decide when to move to cash.
Maybe you can backtest it. I looked at your criteria and thought it had too many subjective pieces. The market does what the market does, and whether or not you've bought or sold anything in the last 3 months doesn't enter into the picture.There's nothing subjective about it, sell when the market is high, buy when the market is low. The 3 month waiting period is to keep from selling again next week when I just sold on Friday.To be honest, I just backtested the S&P, not all the indices, but that's just because I'm lazy, not because it can't be done.I generally try to write down the concrete steps of any method I'm considering, as if I was going to write a program to implement it.That's what I did. I have it programed into a spreadsheet that tells me to buy or sell.What if the US is at a 52-week high but Asia isn't? What if Asia then hits a high but US has dropped below the high? If you insist that all 3 must hit a high (or low) simultaneously, you might *never* get a signal.I should have been more clear, if the average of the indies are at 95% or 5% of their 52 week/26 week trading range. (average of their ranges, not prices) The S&P had hit it's high last week (I believe), but it was getting dragged down by Europe which still isn't at a high (about 80%), but my average hit 98% on Friday so sell, or rather in this case rebalance since it's at my target bond to stock balance. Simple is better, because simple is more robust.I tend to think it's quite simple, more simple than how I interpret the SMA strategy.-murray
I know that many people like to shade their allocation as you mentioned, by slowly shifting to higher (or lower) bond allocations. That's always struck me as being too hesitant, and lacking conviction. Works for some people, but I'm not of that mind.Just a comment on this, I look at in a completely opposite light. I'm "managing" 20+ years of maximum 401k contibutions from two incomes plus monthly after tax contributions. I'm not prone to change course due to the latest blog entry or market report because it "lacks conviction". I didn't get to the position we're in by being spontaneous. I view it as an ocean liner pushing through the rough waters vs a speed boat trying to navigate every wave. OTOH, I could be headed for an ice berg :)-murray
I didn't get to the position we're in by being spontaneous. I view it as an ocean liner pushing through the rough waters There's lots of ways to do good in the stock market. I think the important things is to have a solid, robust, preferably well-tested process that you can stick to. And Don't Panic. ;-)The single biggest thing is to avoid making stupid mistakes, and to invest with your head not your heart.And, of course, to not give your money to somebody who is so afraid of taking any loss that they'll give up the long-term profits. ;-)
Best Of |
Favorites & Replies |
Start a New Board |
My Fool |