As an investor it is quite important to realize how reliable our methods or indicators are. If we don’t know the limitations of our tools, then how are we going to use them well? Some readers may be surprised to know that I have never more than 60% confidence in any of my methods, cycles, indicators or forecasts. I expect them to be wrong at least 40% of the time. And I never have more than 60% confidence in methods, indicators or analysis I find elsewhere on the internet or in books, no matter how compelling the evidence that is being presented. That’s what I call the 60% rule. More on that and on its implications for trading further down in this post.
Let’s start with our customary look at the S&P 500 index (click for larger image):
Last week’s drop saw the market test important support near 1750 before rebounding. We have another week of lunar red period to go, so it’s quite possible that the lows get tested again this week. But several of my indicators now show a nice bottom. The MoM indicator has actually dropped to its lowest levels since May 2012.
So, I think the recent lows will hold and then we will probably see the market climb back until March or April. Sentiment has become especially negative for emerging markets, but my weekly key reversal system is starting to give long term buy signals for markets like Indonesia and Vietnam. Of course, we cannot be more than 60% confident in any of the above.
Why this 60% rule? Well, if we are lucky enough to find some edge in the market it will always be a small edge. There are no big edges to be found in liquid markets. If we find something that works 60% of the time we can do very well already. But our game plan will need to take into account that we will be wrong 40% of the time. Peter Lynch, the famous investor, formulated it like this:
“In this business, if you’re good, you’re right 6 times out of 10. You’re never going to be right 9 times out of 10.”
The advantage of knowing that you will be right only 6 times out of 10 is that you start with realistic expectations and will not suffer from overconfidence. It also comes easier to cut losses short when you know that you will be wrong 4 times out of 10.
So, I always aim for 60% accuracy with my methods and in my forecasting. And that’s hard enough to do. For example, we expect markets to be stronger in lunar green periods than in lunar red periods. How reliable is it? Comparing the green periods to the red periods that come immediately before and after, is a simple way to remove effects from longer term trend and offers a fair comparison. Since 2009, when we started this blog, the green periods have outperformed the red periods that come before and after it 57.5% of the time. So, quite close to 60%.
Good luck,
Danny
Related articles
LunaticTrader 
“system is starting to give long term buy signals for markets like Indonesia and Vietnam”
How much of the rally do you need to see before there is an actual buy signal? Vietnam is up 25% in the last year.
I follow you regularly but to get the 57% performance you need to take every signal and four years of performance is not really enough to have confidence. It’s like asking someone to buy a car from a manufacturer who has only been selling cars for 4 years compared with a company such as BMW. It’s a tall ask. Not knocking it; well done. Just trying to give the perspective of someone who trawls the internet and is constantly bombarded by ‘follow me’ schemes.
Hi Ptolemy,
The long term buy signals do not depend solely on how much a market is up vs a year ago. If you look at a 1 year chart of Vietnam market, you can see it has been in a sideways consolidation pattern and only broke out in January: http://www.bloomberg.com/quote/VNINDEX:IND
The reason why I only gave last 4 years data as an example is because I don’t want to use “in sample” data as evidence. The lunar effect has been in place for more than 60 years, but as a general principle the data that was used to derive a hypothesis is not used to confirm the hypothesis. So, I use the “out of sample” results since 2009, when we started keep track of it here on the blog.
That’s how trading systems should be tested, on the out of sample results. And that’s where it usually becomes difficult to reach or exceed 60% accuracy rates.
Danny
Danny, Thanks for reply. I understand better, now.
Your best post EVER
Because its a fact
When somebody tells me they get 70% winners, its short term
Thank you
bob
Thanks for the thumbs up, Bob.