A year ago I wrote an article about why trend following doesn’t work on stocks. That article surprised many people. After all, I was mostly known for writing a book about trend following on futures. Why would I diss trend following? And why does a futures guy talk about stocks?
The fact is that I’ve been working with stocks longer than with futures. I’m a quant. I’m not tied down to any particular asset class or strategy. As most professionals, I’m pragmatic.
The point with that article a year ago was to highlight the differences. There’s trending behavior in both stocks and futures, but it’s a very different game. In drawing a demarcation line between trend following and equity momentum, I wanted to make that clear. My concern was with hobby traders who might make the mistake of applying classic trend following models for futures, on stocks.
What is the Deal with the Word Splitting?
I use different terminology for stocks. I’ve stated before that trend following doesn’t work on stocks, but that momentum does. Why bother with this distinction?
Equity momentum is very different. The classic trend following style models are heavily reliant on diversification. In stocks, you get very little of that. The equity game is largely about beta. The market regime is of great importance. Stock selection increases complexity. Leverage isn’t available as it is with futures. There are suddenly many more variables to deal with, while some key features of futures are suddenly gone.
I use different words to highlight the different characters of these markets. You cannot apply the same trading models. Regardless of what system sellers might tell you.
Stocks on the Move
Stocks on the Move has a very similar concept as Following the Trend. The entire book describes one single trading approach, and it does it in great detail. I walk you through the problem first. I’ll go over mutual funds and ETFs, describing what they’re good for and where they’re a bad idea.
I also explain why equities, contrary to popular beliefs, is the most difficult asset class. It only seems easy, if you don’t look closely enough.
Then I’ll move on to telling you why classic trend following doesn’t work on stocks. Don’t worry, I’ll show you all the details and all the simulations to back that statement up. This of course leads to the point of the book. Momentum investing.
In the next section, I’ll detail the building blocks that we need for a proper momentum strategy. There are some very important components here. How to decide on stock universe, how to measure overall market regime, how to rank stocks, how to set position sizes, when to buy stocks, when to sell them, how to rebalance positions etc. I’ll explain these points and give you a set of rules to start out with.
After I’ve given you all the rules, we move on to the results. Like in Following the Trend, I go year by year in detail. It’s important to understand the dynamics of a strategy, not just the end results of a simulation. I try to give you realistic expectations of deploying such models in reality.
In the last part of the book, I go over some of the strategy components again, to see which ones are really important. Does the trend filter matter? Perhaps the ranking method is the key? Or the time frame?
As with Following the Trend, I show all details and all rules in the book. After all, anyone can make up any performance if they don’t show you all the rules. Feel free to test them, use them and modify them.
Buying the Book
You can buy the book right here on this site. The payment, shipping and handling will be done by Amazon.
Wait, if it’s handled by Amazon, why do I direct you to this link on my own site? Well… Frankly because while you pay the exact same amount as you would on Amazon’s site, for some reason I get to keep a much larger part for myself. No, it might not make all the sense in the world, but Amazon pays a considerably higher royalty per book if they’re ordered through this link that I keep posting on this page. Oh, you can buy it on Amazon too of course, both the paper version and the Kindle.
Help! Review and Share
I have self published this book. Publishers claim that they do a substantial amount of marketing, and that this is the reason why books like Following the Trend sell so well. I’m hoping to prove them wrong. For that, I need your help.
If you like my book, please write a review for it. Amazon reviews is one of the key sales tools for trading books. If you have a blog, please write a review for it there. Trading blogs has become a very important medium for making books visible in the first place. Facebook, Twitter, LinkedIn and such sites is where awareness is so important.
Can you help me out? If my self publishing experiment works, and if this book really does outsell my Wiley published book, then there will be more like this published. Help me make this book visible and I’ll keep producing content!
Seems that if you follow link from LinkedIn using iPad app, CreateSpace order does not succeed (you shopping cart becomes empty after you have entered your address). However, managed to complete order in Createspace by using laptop and Chrome browser. Hope Amazon delivers EU customers orders from Europe, not from US, so I would get book sooner :).
Using CreateSpace makes no sense from customer point of view: Estimated delivery to Finland is 20/7 while I could have had delivery from Amazon UK in few days
Ah, sorry, Arto! I should have made that very clear… Ordering from CreateSpace is the same as ordering from Amazon.com.
Still, I have a feeling that your book will arrive earlier…
Createspace wouldnt process my credit card so ended up buying on amazon..only downside is that delivery scheduled to my address in the Middle East for july 8th – 28th which is quite long based on my experience with amazon. No option for faster delivery 🙁
Ordered, will receive on June 23. Worked fine on an iPhone with Chrome browser.
On the Amazon site it says only the Kindle version is available and it’s not released until July 1. Createspace was fine though, I prefer paper over Kindle.
Ordered right away! I bet I’ll devour this over the weekend, I’m really interested to see how you construct your ranking models and universes and if they’re segmented/double-sorted by industry/sector, and how your practical trading model squares with academia when it comes to cross-sectional momentum with respect to effective holding periods.
If you want to give the book a bit more of a boost I’d love to hear you be interviewed on Michael Covel or Niel’s podcasts if you have the time.
All the best,
I’m also really looking forward to the Kindle version. As for Michael Covel interviewing Andreas Clenow…The first meeting would go something like this…
Mr Covel: So what do you do?
Mr Clenow: I am a professional trader and write books which teach people how to trade. The books cost $40. So, what do you do?
Mr Covel. I am a professional marketing expert and sell courses which teach people how to trade. The course costs $2000.
Mr Clenow: Uh, ok. Perhaps you could teach me how to market my books better.
Mr Covel: Great. Perhaps you could teach me how to trade.
I love this response! That BS drives me nuts. Covel is a parasite.
He’s a book author, blog and podcast publisher. I can’t see how you could fault him there. His only problem, as I see it, is that he doesn’t produce any track-record for the systems he sells. But no one forces you to spend $3k on one of those.
I ordered a hard copy as soon as this article is out. I cannot wait for the book to arrive. Thanks Andreas, I am confident this will be another invaluable book for your readers.
Paperback on it’s way, arrives Friday night – very nice!
Good for others that Kindle version got accelerated to Monday release.
#1 Best Sellers in Stock Market Investing (kindle)
#1 New Release in Stock Market Investing (paperback)
Sounds like some serious momentum (pun intended)
Why on earth is the kindle version 48 bucks and the paperback 38? Have the amazon guys lost their minds? f…
That’s not right. Is that on Amazon.com or a local Amazon site?
It should be 38 everywhere, for all editions. I set the price and they shouldn’t fiddle with it.
I’ll have a word with them see what’s going on. When I check the Kindle myself on Amazon.com now, I see it listed at 30.47. I don’t know what they’re up to…
It’s very right actually. I checked again, using your link above. Takes me to the amazon.com. The paperback is 38. This really sucks big time from amazon, I dont know what they are doing it. I can send you a screenshot if you send me your email.
Even if I search by the book name in search results for Kindle version they say 48.
I’m discussing with them now. They claim nothing is wrong, I still claim that something is up.
My theory is that they vary the price based on your cookies. Try opening Amazon.com in a different browser, or in incognito mode. That should prevent them from getting your cookies. When I do that, I get 38 again, and if I use my regular browser I get a different price.
I think that the sneaky monkeys are changing price based on your shopping patterns etc. There was an article a few years ago that airlines charged higher prices if you used an Apple computer to book. Presumably because if you already bought an overpriced computer, you’re more likely to buy an overpriced ticket…
edit: just saw your post that you tried what I suggested. I’ll keep chasing Amazon to see what’s going on here.
This is too funny. If I search with a browser where I am NOT logged in @amazon it shows 38… I wonder if it’s because I always buy the kindle editions of books? Or I’m too paranoic about it ? 😉
K talked to support and they said they automatically add me a $9 gift or something so that price is ok in the end. But it’s still strange, they should fix it. Thanks for the prompt reply!
Well, this pricing issue is getting complicated. I’ve been playing email ping pong with Kindle support for a while to try to figure this out.
Based on my parsing and interpretation of their very odd explanations, here’s what the logic seems to be:
* Amazon has what they call dedicated kindle stores for each country. For me, in Switzerland, that would be Amazon.de.
* Even if you go to amazon.com, logic will be applied based on your dedicated kindle store.
* User Gicu in the comments above, lives in Eastern Europe and has Amazon.com as ‘dedicated kindle store’.
* His country has a VAT of 24%.
* His price is then 47.12 USD.
* So far, reasonably simple, huh?
* I see a price of 30.47 USD when I look at Amazon.com.
* This is because my ‘dedicated kindle store’ is amazon.de.
* At Amazon.de the book is priced at 26.98 EUR, because with their 19% VAT that approximately matches the 38 USD on Amazon.com.
* Here’s the kicker: Amazon has an auto price matching system, which really makes for weird effects here.
* When I check Amazon.com, since I’m in Switzerland, the system checks my price in Germany, which is then 26.98 + VAT. Our VAT in Switzerland is much lower, resulting in a price matching output of 30 USD, which is what I see on my own screen.
Makes perfect sense, Amazon. Thanks for keeping it simple.
andreas, i ordered the book and it arrive sat. read it in one sitting. excellent work, very easy to read.
i had one question though about the risk parity sp500 strategy on p. 250 of the paperback with the annualized return of 13.1%. did you run this with an exit for all names when the sp500 dips below its 200ma to see how that would improve results? or did you run it with an exit for each individual name if its dropped below its 100ma? this would seem to complete the cycle and one could compare the risk parity strategy to the momentum strategy on an apples to apples basis. thanks!
Thanks, Justin! Glad you liked it!
I’d very much appreciate if you want to write a review on Amazon. These days, that’s the main marketing space for books.
Vola parity sim: That sim holds all stocks, all the time. No stops, ever. It’s a highly unrealistic strategy, unless of course you’re running it as ETF, but it makes for a good research demo strategy.
I didn’t run with the index exit, or the individual stock SMA exit. Mostly because it’s already an unrealistic strategy. Also because it opens a new can of worms: What triggers reentry? Do you keep cash or rescale remaining positions? Maintaining vola parity positions would then, probably, require calculating the theoretical sizes of non-qualifying positions and reducing asset base with that amount before sizing the rest. Hm… can be solved, but adds complexity to an already not-so-realistic strategy.
i’ll be sure to write a review.
re: the strategy, i was thinking more along the lines of shorting SPY in an amount equal to the notional amount of the open positions whenever spx dipped below 200ma and taking off the hedge when spx returned above it. i know it’s not realistic, but i was just curious to see how that type of strategy would do compared to the momentum strategy outlined in the book. thanks!
Shorting the SPY creates some problems too, as it’s a cash instrument with limited leverage possibilities. Using the futures would be easier from a practical point of view, but you’ve still got beta to deal with. It would be worth trying, but I’d use a beta adjusted ratio for the hedge.
Of course, the first thing to try is that short strategy in a vacuum.
Still waiting on Amazon delivery but bought a digital copy on smashword.com too..read it page to page in one go. The cost of 2 books was lower than what I would potentially lose on a single trade so I figured its worth it!
Solid effort, logical explanations, and very entertaining to read through! Good job! I have a feeling you will be moving up from top 5% to top 1% fairly soon!
What I really appreciate is that you share ideas which widen my frame of thought and views towards creating ideas and trading. This book should be a “must have” on every equities traders shelf (or PC/phone).
Hope the new website stocksonthemove.net picks up and looking forward to the continuous education you kindly bestow on us!
Ah! Now I know the source of 50% of my Smashword sales! 🙂 So far, I’ve sold exactly two copies on that site.
Soon I have to write a post about my somewhat absurd findings in the world of self publishing… The reason I put it on Smashwords was just that they help me put the book on the Apple store. Apple iBook store being the second largest in the world, after Kindle. But… Apple is being Apple about it. While you can upload a pdf via a website to Amazon, Apple require not only that you use iTunes to upload, but also that you use an Apple computer to do it.
Yes… It’s not that they require a specific format, which one could understand. They require that authors buy an Apple computer to be allowed to upload and sell on their site. Why do they do that? Because fuck you, that’s why. Anyhow, that’s why the book is on Smashwords. And I’m starting to like Smashwords… They actually pay authors almost the whole amount, while Amazon only pays 35% of the Kindle price.
Rants aside, I’m glad you like the book, Kamran! Would be great if you want to write a review for it.
The new, and so far utterly useless site, StocksOnTheMove.com will improve soon. My plan is to automate stock analytics and charts there, including uploading daily reports on the ranking tables, current portfolio holdings, position sizes etc.
I have written reviews for the book on both Amazon & smashwords.
Looking forward to to how the website develops!
Hi Andreas, great book.
I finally got it yesterday via Kindle (didn’t know about the smashwords version, or I would have got it in epub, I much prefer that to being stuck to Amazon’s app), and after a single read-through I’m very happy.
The graphs are a little small in the Kindle version, and while if I squint really hard I can almost make out what I think the labels say, it would have been nice to have them a bit larger. Also, I found an error for you to fix for the next version, the paragraph following Figure 2-1, the text says:
“Here’s the simple explanation. Start with the index at 100. The first day, the index drops by 10%. Now the index is at 100, the short ETF is at 110 and the double short at 120.”
it should be (underlined to highlight)
“Here’s the simple explanation. Start with the index at 100. The first day, the index drops by 10%. Now the index is at _90_, the short ETF is at 110 and the double short at 120.”
Now to just go and write the code to test the system.
Also notice that the flow chart in Figure 10-1 doesn’t strictly follow the flow the rules given three paragraphs above – the flow chart, if it isn’t the second Wednesday, finishes after checking sell conditions without buying more positions with available cash. The text version always buys on Wednesdays if there is available cash and if the index is above the moving average.
Thank you, Jani! Updating the files now.
Just finished reading the book on Kindle. Great book, just like the first one. Thanks for all the work.
Had a question for you – the strategy looks like it has a high turnover and hence might incur significant trading costs. It would be really great if you could you shed some light on this from any simulations that you might have done or from your own experience? How worse does it get from the back-tested results presented in the book?
I am based out of India and planning to do a local version of it. Here the transaction costs are pretty high for retail traders as compared to the developed world so I need to pay attention to the turnover while assessing back-tests of a frequently trading stock strategy.
Thanks a ton. Looking forward to the website.
Trading cost isn’t that bad, at least if you trade US or EU stocks. I don’t know anything about India though.
I used realistic costs, based on actual recorded costs that we had at our shop. I’ve been trading this and similar models for quite some years now, and my experience is in-line with what the book presents.
Thanks for the comments on the errors. I’ll take a look today and make the updates.
On Kindle, you should be able to zoom in by double clicking the images, filling your screen.
Thanks Andreas for the quick reply. Appreciate it.
So if I understand correctly, the simulation statistics for the strategy shown in the book already incorporates a reasonable estimate of trading costs?
Also another question – was going through your public research library and found this paper from UK titled ‘Can small investors exploit the momentum effect’ (spoiler: yes) in which they form Winner – Loser portfolios of extreme winners and losers (from as small as 1 stock on each side all the way up to 50) and in ALL the cases the returns are driven by the Loser part of the portfolio.
I am finding this difficult to square-off with what I just read in your book. What gives? I know they are somewhat different momentum strategies but this seems like a conceptually different outcome altogether. Can you please take a look at that paper and resolve this apparent conundrum for me?
Thanks a lot. Really appreciate your taking time out to answer all this.
Don’t take academic papers too seriously, Max. They can be interesting to read, but the assumptions and conclusions are almost always flawed. I wouldn’t try to implement these kinds of portfolios…
They’re going long strong stocks and shorting losing stocks. It’s very different from what’s in my book. They’re attempting to avoid beta, while my strategy is beta heavy.
I just looked through it now to see how they allocate capital, but didn’t find it. We probably have a large difference there too. Perhaps it’s in there, but their way of writing is making me lose concentration… If I read anything more about endogeneity issues contaminating characteristics of companies, I’ll fall asleep.
All in all, their strategy is dramatically different.
I had this concern as well. More notably, even if the results would have been negative, I would have appreciated data on the inverse of the actual model.
In other words, we gave long stocks the benefit of a regime filter, rankings and avoiding high-volatility situations when these procedures would arguable benefit short positions even more. Thus we only saw that a raw version of shorting doesn’t work.
For the purposes of comparison, it might have been good to see how the shorts would have done with these same modifications one would have more confidence in the conclusion regarding shorts.
Read the book on the commute yesterday and really enjoyed it.
Any suggestion where you can get adjusted equity prices without survivorship bias for retail investors – most of the market data providers are focused towards corporate clients with prices to match
Hi Paul, if you google “PDU Alpha testing” you will find a source of data which includes historical index constituency and delisted stocks. I use the Australian data, but it seems they have the US data available too on a trial basis if you are subscribed to the normal EOD data service.
Had a look at the Premium Data feed, seems quite cheap e.g. yearly sub to US EOD is $297 and to go back to 1985 including delisted is $307; to go back to 1950 is an additional $400
Only issue really is geo-coverage, just US and Australia and I want to get the UK and Europe.
Andreas, I assume the next book is about corporate bonds and then we can have a nice asset allocation strategy between equities, fixed income and commodities – we will all have portfolios >$1m by then of course 🙂
For the book, I used QuantQuote data. If I recall, I think it cost about 1k up front and 100 a month for updates. Minute data resolution, delisted data, adjustment factors, mapping data, the works. Compared to the usual institutional sources, it’s a steal.
I’m also eagerly waiting for the solution promised by Premium Data. They’ve been working on a similar package for a year now, which they plan to sell at a much lower price still. I haven’t seen that yet, so I can’t vouch for quality of course. If I do test it, I’ll report back.
Yeah QuantQuote seems to be $895 for S&P 100 and $1870 for S&P 500 – $100 pm per 1000 symbols
They seem quite a curious company e.g. mentions of US company fundamentals data (2012) and expansion into Asia (2010), that never seem to have happened.
As a company, they are certainly curious. I got the feeling that they really had no interest at all in selling to me. In fact, it took me a while just to get them on the phone… My impression was that selling data might be a side business that they don’t prioritize, but I don’t know.
Still, I found the data quality to be great and at a reasonable price point. Too bad they don’t have global coverage.
Just received your book in kindle format yesterday. So far i read it up to now very interesting (like your first book) but when i reached the first charts and pictures i could not read anything in the graphics.
I used a kindle fire, kindle touch and even on my 27″ Monitor the resolution of the pictures is far to low to recognize the text in the pictures. Evevn zooming in doesn`t help.
Could you put the pictures on a special website in a higher resolution (at least 150 dpi), so i can print them out to have a look at it while reading the kindle version ?
Greetings from germany
Has the printed version of the book a higher resolution of the pictures ? Can anybody read the text in the pictures of the printed version ?
Thank you in advance
Sorry to hear that the images aren’t coming out well. I’ll post them all on this site as soon as I can. I’ll also talk to Kindle support and see what can be done to enhance them. I submitted them all in 600 dpi and it seemed fine on my own preview copy.
The printed version is readable, thought I’m planning on making the text larger there as well just in case.
Feel free to email me directly as well if anything is unclear, of you’d like a copy of all the images. Address is email@example.com. I’ll add back the contact form on this site as well… I took it down a few months ago when I was being overrun by spammers.
thank you for your help to send the pictures. I have send you an email.
Just a quick question if I may, the 200 day moving average of the index – should this be the plain index or the index adjusted for earnings (in my case I can pick from the $XAO All Ordinaries or the $XAOA All Ordinaries Accumulation data)?
I have the system coded in Amibroker now, to the point where I can generate the required table to follow the strategy, but I’m still working out how to get rotational trading working with rebalances…
I used the price index, but strictly speaking the total return index would make sense. It won’t make any real difference though, since we use it for such a crude purpose. It doesn’t really matter if you use 180 days or 220 days, so the difference between the price and tr index doesn’t really matter either on this scale. TR is more correct, but in the end it’s a rounding error.
Glad to see another Amibroker user.
Did you manage to implement the system in Ami?
Do you know of any web resources that would be helpful?
Does rotational trading mode allow staying in cash during bear market?
Ok, as a first step, I’ve put all images from the book in high res online.
Next I’ll make sure the source files are improved for the Kindle version as well.
Great Book Andreas! Downloaded and read in one evening. Left a review, and it is well deserved. Your stuff is going to change the investing world. By making this stuff available to any interested reader, the BS in the industry just stands out as trash even more strongly. Thanks for posting higher resolution charts. You could get the gist of the chart content from the text, except that pesky excel regression logic screen grab was laughably fuzzy for understanding.
Great read so far, about half way through. A quick question about using other indexes rather than the S&P500. In the real world rather than the book simulation spcae, would you likely add in a liquidity filter to ensure you’re not cornered in thinly traded positions? Probably less of a problem with the 500 but with the mid or small cap indexes more of an issue? If yes then would you use absolute volume thresholds or a $ traded per day limits (i.e. volume X price>$5m)?
thanks and will contribute Amazon review as soon as finished. Really enjoy the way you’re shaking up the boat challenging conventional (hobbiest) trading wisdoms re quotes, stop losses etc.
Depends very much on your choice of instrument universe and your trade sizes. If you’re dealing with the S&P 500 stocks, it’s obviously not a concern. Even if you’re dealing with the S&P 600 Small Cap Index, you’d have to be rocking a pretty sizable portfolio for liquidity to be a concern.
If you move to local country indices or larger universes with smaller stocks, you would need some sort of liquidity filter of course. Risks are higher with small companies, so your return targets must be higher.
The design of the liquidity filter would have to be adapted to your portfolio size, trading style and the kind of stocks you’re dealing with. I prefer to deal only in very liquid issues, so it’s not much of a concern for me.
Thanks again for another great book. I know it might be a bit more technical than your current two books, but I would love a book on automated trading systems for options on futures. Have you done worked in this area?
I’m glad you like the book! You’re right that options on futures would be a too narrow subject to write a book about. It would limit potential readers to a few hundred. Also, I don’t actually use these instruments so I wouldn’t be the right person to explain them.
Just got my copy today. I am hands on already! is looking great for now…good job!
Thanks, Javier! Hope you’ll like it.
Excellent read, Andreas. The hard copy could really do with a proof-read to iron out a number of errors but that’s a small point. We are running some cross-sectional momentum strategies, quite successfully, but I am very keen to take some of the flavours of your work and conduct some additional filtering. Big thanks for writing and I wish you success with this book.
I found a couple of typos to fixed, but please feel free to email me with any errors or typos. I’m planning on publishing an errata page with any errors, as well as fix the new copies being sold. An advantage with self publishing is that I can make changes any time and it takes 1 to 2 days for it to be incorporated in all new sold copies.
Hallo mr Clenow
Thank you for writing about rebelancing because that is something completly new for me (and also other I think)
The new book contains more information about trading than my total home libary of goeroe’s.
Thanks, Bart! I appreciate the feedback and I’m glad you liked the book! It would be great if you want to write a review on Amazon.
This book is a bit of an experiment in self publishing for me. If it all goes as planned, I might continue with more books of this style later on.
on page 91 I got really upset, hence please clarify:
Do you seriously use ATR(20) in your 16 year long backtest for evaluating performance?
If so how exactly do you adjust data for splits and dividends?
by “evaluating performance ” I mean your historical results represent the usage of ATR(20) for position sizing.
I used ATR with 20 day lookback to size positions, yes. Odd thing to get upset about. What am I missing?
The split and dividend adjustments were done the usual way. I.e. when a corporate action occurs, an adjustment factor is calculated. In a 2:1 spit, the factor would be 0.5. All prices back in time are adjusted by this factor, so that the corporate action does not show up as an actual price move, which is really is not. Dividend logic is very much the same. An adjustment factor is calculated and prices are back adjusted.
Naturally the price you see in such charts 16 years ago is very different from what people saw at the time. That really doesn’t matter for our simulation. The ATR will of course have a very different absolute value, but this again doesn’t matter. The relation of the ATR to the price does matter, and remember that both are adjusted the same way so that relation stands.
If the simulation says that we bought 1000 shares of IBM in 2003, odds are that the actual number would have been different if we had traded it in reality back then. You’d have to check the corporate actions adjustment ratio and the dividend adjustment ratio to calculate the actual traded shares. The PnL however is the same, or at the very least close enough.
I have notice a lot of confusion around these adjustments in the past and perhaps I should have written a proper chapter just on this. Well, I guess I’ll do an article on this site instead.
Thanks for the clarification Andreas. I beg to differ.
No question, the method of the strategy is sound, the principle as well, the 10 year MAR is decent. It’s a “larger account” strategy. Risk adjusted returns increase up to about 50 positions, CAGR has a max at about 25 open positions. Either way, please:
There is a great chance that your position sizing or volatility adjustments in terms of sizing are incorrect, or random at best.
Here is why: If you ratio adjust, for splits and dividends, that is: adjust ratio splits, adjust ratio dividends – there is only ONE relationship that is maintained: ratio or in other words percentage relations.
That’s a huge point.
Quote:”The relation of the ATR to the price does matter” yes, that’s exactly the problem. %-ATR must be used for sizing and re-balancing.
So: ratio adjusted data: %ATR *only* makes sense from an analytical point of view. ROC makes sense. Point based ATR is useless , same goes for momentum or other point based indicators. Some info may still be in wrongly calculated point-ATR, but why take chances?
That error keeps cascading. By that, I mean really cascading because there are stocks that split more often than others, distorting the actual price and 100% guaranteed the point based ATR. The effect is worse the more you go back in time. Instead %ATR must be used. There is no doubt from an analytical point of view that the results will represent trading reality better, hence I’m wondering why it hasn’t been used.
By the same token, for futures using %based stops or %-indicators on point-based back-adjusted data will give errors. That would deem the results as useless.
That begs another question with the commissions: US stock commissions are in cents per share. After all the adjustments in ratio terms, the actual price must be known to calculate the correct commissions. The shortcut is to use a “guessed” %of trade volume (#of shares x price) in terms of commissions which would allow a possible extension of the backtest a decade or so further. Next thing that hasn’t been mentioned is the fact that US stocks did trade in fractions prior to 2001. that adds another source of error. Long term quant models can be tricky.
I’m curious about your feedback.
A stock is is trading at around 100 and tends to move by around a percent a day. It has an ATR of exactly 1 right now. We have a portfolio size of 100,000 and we use a risk factor of 0.1%.
100,000 * 0.001 / 1 = 100 => we buy 100 shares@100, spending 10,000.
A month later there’s a 2:1 split. All data back in time are updated in the simulation. The adjusted price was now 50 and the adjusted ATR was 0.5. Now we suddenly have 200 shares at an entry price of 50.
100,000 * 0.001 / 0.5 = 200 => we buy 200 shares@50, spending 10,000.
All points are adjusted by the same ratio. Try running the same simulation using ATR% and you’ll see that the results are nearly identical.
You are of course right in that simulations are more unreliable the further back that they go. It’s a general problem with simulations. Of course, all simulations are unreliable to some extent and reality will never play out the way it was simulated. Still, it does give you some indications as to how your models might behave.
Absolutely agree on percentage based stops for back adjusted futures. That just doesn’t work.
As to commissions, in the US markets it’s easy to get ticket fees instead of percentage of notional which is common in the rest of the world. Fractions vs decimals doesn’t have a meaningful impact for this type of long term model.
I have enjoyed and learned a great deal from both of your books. I read most of Stocks On the Move last night and finished it this afternoon. Very easy to read. I am particularly interested in your ranking methodology in chapter 7. I run a similar strategy however I use a different method for calculating momentum/relative strength. I want to perform the momentum calculations described in your book to see how your method compares to what I am currently using. My method does not directly take into consideration volatility like the method you describe and that is something I want to explore. I consider volatility in my shorter term strategies but not in the intermediate/long-term momentum strategy,
Background: To rank stocks by momentum/relative strength I take the performance for multiple time periods and compare to the S&P 500 over the same periods. I then rank rank by level of outperformance/underperformance. My approach is intermediate to long-term and I assign weights to each time period with the longer term periods weighted more heavily. I have exit criteria in the event that a stock plunges while showing a high momentum score. I currently use this method for stocks and ETF’s.
I am not a statistician and I struggled a bit with some of the technical aspects of the math in chapter 7, particularly how to calculate the exponential regression slope. I understand the concept and it makes a great deal of sense to me but I cannot figure out how to put all the pieces together. I use NinjaTrader and have the data for linear regression slope and R-Squared. I’m not sure how to take this information and obtain the exponential regression slope. Any suggestions or clarifications as to what I am missing are greatly appreciated.
Congratulations on a very good book!
Let me try a brief explanation here, and I’ll return with a full article on the topic as soon as I find the time.
There’s an easy solution, which is shown as Excel logic here: http://www.stocksonthemove.net/book-images/#prettyPhoto%5Bgallery%5D/15/
Briefly, get the natlogs of the prices, run a standard linear regression on that, Exp it back again and annualize it.
The problem with linear regression is that you can’t compare it across stocks with very different prices. With expreg, you get a percentage and that can be compared.
Thanks for the reply. Appreciate the info. I’ll give it a try this afternoon and see how it goes.
I looked at the “Regression Logic in Excel” illustration and I have a couple of questions.
I spent some time reviewing this in-depth this afternoon and I think I have a handle on it now. Will you tell me if what I have below is a correct understanding of what you did in chapter 7?
We need to find the Adjusted Slope to measure momentum that incorporates volatility. This is a 2-step process.
First, find the linear regression slope and annualize it. This annualized number is the Exponential Regression Slope. The NinjaTrader platform gives me the linear regression slope but it appears as though I need to use Excel to annualize this number.
Second, once I have the annualized number I need to multiply by R-squared (obtained from NinjaTrader) to obtain the Adjusted Slope. Adjusted Slope is the number we are looking for in order to rank the stocks by momentum WHILE incorporating volatility.
Do I have the 2 steps outlined above correct?
Thanks for your patience. I appreciate the help.
I made a sample spreadsheet to demonstrate the logic. You can download it here: http://www.followingthetrend.com/mdocs-posts/stocks-on-the-move-ranking-logic/
I hope that sheet should explain better.
Thanks for the detailed reply. I was hoping that I would be able to download data from my charting package, calculate the missing data (Ln for example) and run the additional calculations without having to download so much data. It appears at first glance that is not possible. I’ll need to put more thought into this and see what I come up with. Thanks again for your time and for writing books that challenge me and help me grow as a trader/investor.
You should be able to do all those calculations directly in the charting application. I wouldn’t do it in Excel, that’s just for demonstrating the logic.
You could build your own function in the charting application that replicates the logic from my spreadsheet.
Good book Andreas . Good job !!
Only one question.
Why didn’t you include the comissions in the backtesting ?
In my opinion, estrategies with many trades the commissions are an important factor.
Thanks, Ricardo! Glad you liked it!
Commissions were included, but I failed to explain it. I used a fixed ticket fee.
I’d appreciate if you share the amount of the commissions used in the backtests. The commissions are one of the most important variables in the strategies.
I’m trying to develop the strategy in my platform and I’d like approximate the most possible to your test.
I had forgotten and had to go check… 🙂 I used $20 ticket fee and a 100k portfolio.
It’s very easy to get the amount of transactions down for the model in the book. I wanted to avoid optimizing for anything, including that. I wanted to keep it as a theoretical demo model.
If you do want to reduce trading, try simulating with less frequent rebalancing and with a higher threshold for changing position size. You won’t see much of a performance impact, but you can reduce trading a lot.
Hi Andreas, Excellent book. Finally i see a trading system which is robust and still feasible to follow for retail investor like me. I have already started using it.
One question. Do you use in your manager accounts and funds only the strategy detailed out in the book for stocks or do you use counter trend strategy similar to plunger indicator in futures to smooth out drawdowns.
As I usually say, it would be great if you could write a review on Amazon. I know many authors cheat in various ways to get hundreds of five star reviews for their books. I refuse to use these tricks, but real reviews from real people are always very valuable.
Your question: The strategy in the book doesn’t make for a very attractive institutional product by itself. Mostly because of the flat periods. It would be near impossible to sell a strategy where you can potentially get paid for doing nothing for years, even if doing nothing would beat almost everyone in such years. To make a professional product out of this, you’d need to combine it with something else.
For our own equity momentum fund, we combine a core momentum strategy with a long term cross asset ETF model that’s designed to smooth volatility and to provide income during prolonged bear markets.
Qualified investors, feel free to contact me for more info. With the tightening regulations in Europe, everyone needs to be careful about not doing public marketing, which this comment is clearly not… 🙂
Thanks Andreas for quick response and that makes sense. I did my review on Amazon.
Thank you for the review, Govind! Very appreciated
Hello – I would like to buy the book, but being new to this style of trading I have a question with regards to building the model myself. In order to get the most out of the book, what software or back testing tools do I need to have access to. I am assuming excel, but are readers able to use something like Portfolio123 to build the trading rules in and backrest it? Sorry if this question seems basic, appreciate the help!
Backtesting anything in Excel is near impossible, unless the strategy is amazingly simple. It’s not a good platform for that type of work.
Assuming you’re talking about my second book, Stocks on the Move, I used RightEdge for simulations with data from QuantQuote. I used a MySql server for data storage and made my own adapters to transform data as required.
Most consumer grade simulation platforms are utterly useless. The people making them don’t understand better and can’t be bothered, since they’re not actually using the software. If you’re serious about systematic trading, you should get a proper simulation environment and learn what’s needed to operate it.
Thanks so much for the reply. Your site is great and I have learned a lot by going through. Thanks for the suggestion on RightEdge. I actually reviewed that article you posted about this software and will download it tonight to try it out. That said, I am not a programmer and the only limited programming I do is html, which is certainly different. That said, I am motivated to learn.
I am speaking about Stocks on the Move – I bought it and will devour it tonight. I guess my question was really, if I wanted to utilize the strategy you talk about in the book to momentum trade stocks, then what platforms will I need to have to run the “screen”. Perhaps you get into in the book, but thought I would try to get some clarity before getting too far down the path.
Thank you Andreas!
Hello Andreas. Love both your books! I have a question on strategy rules from Stocks on the Move. When re-balancing, if volatility has decreased for each of your positions (unlikely, but possible), rules would have you increase size of each position, which might not work if you have limited free cash left. What would rules be in such a situation? It is definitely possible that I am missing something. Many thanks for all your work and your willingness to share knowledge with the trading community. Greg
Running real portfolios, you need to think through these kind of details and decide how you want to handle it. It won’t make much of a difference in this case, but it’s good to have clear rules. The simulation from the book tried would in such situation submit all orders and when the cash is out, the remaining open orders would simply be cancelled.
Hi Andreas, thank you for the dialogue. You mentioned in the book that the strategy worked despite the large cap nature of the S&P 500, not because of it, which makes sense. You hinted (I think) that strategy would likely yield better long-term results on smaller cap indexes if you are willing to live with the inevitable higher volatility. Just wondering is you ran strategy on S&P mid-cap 400 and/or S&P small-cap 600. If so, could you comment on your findings?
If you use S&P 400 or S&P 600, you’re likely to see higher returns but also higher vola. You might also want to research using broad, international indexes, such as MSCI World. Of course, you’ll add some complexity that you need to handle. You’d get a currency problem that you didn’t have before, and you may need some constraints on sector and geography, to avoid ending up with too extreme corner portfolios.
These topics can get a bit complex and would make for quite a boring book that only a few industry professionals would like. I wanted to keep the book focused on simplicity and something that most readers are able to replicate and implement.
I really love your book and I bought it on Apple Store. I’m from south africa and Im going to say this is the best book ever written in the world ex the bible. lol. Every other book out there is trash seriously. I have been in this game for so long and this is the only book that has impacted me in such resounding ways.
Firstly: I read your book following the trend too and you have stated the strategy in your book wouldnt work with small capital. I completely agree and understand your point but what should small/retail investors do then? Not bother to trade futures at all? If they still can what would strategies would be the alternative?
Secondly: Can a variant of your equity model work for other asset class. Momentum has been researched in other asset class and found to be prevalent too. So this should also work trading futures too right? Basically all I’m saying is, I dont really like trend following because I cant psychologically deal with stop losses as Im a long term trader and I have found stop losses can detract from long term trading. So can I use this for futures too?
Futures aren’t really made for smaller traders. The problem is that the contract sizes are far too large. For diversified futures strategies, you need to be able to have 20-30 positions open, and even if you take a single contract that would generate far too high risk.
For retail, CFDs offer some potential. Just be careful with your counter party risks.
Equity model on other asset classes: Sure, but the key word being variants. Equities are a bit special. You have a huge amount to chose from. You have indexes. Massive beta to the indexes etc. The rules in the book take this into account, so if you want to apply the principles to other asset classes, you’d need to adapt for that.
The regression logic should be fine to use. Rebalancing should always be used for any long term strategy. You can pick and chose what works for the asset class you’re applying it to.
I’d suggest to do proper simulation work before going live of course.
Hi. Bit of a newbie question as I’ve not traded before although have been interested in it for a while and am keen to start after reading your book SotM. So apologies if I’m not understanding things correctly!
I’m planning on developing my own system, probably in RightEdge as I want to look at trading in the FTSE as I’m in the UK.
I have done some coding in the past (mostly in VB) but not in C#. My job does not involve programming now but I manage people that do so getting into C# is not going to be a bad idea!
So to my first question, I’m just trying to work out, do you reinvest profits here or does the money invested when buying stocks always stay at 100K?
Watch out for the stamp duties in the UK! They can quickly kill systematic strategies in that market.
The portfolio is the portfolio, and no money was moved in or out of it during the simulation. The portfolio grows, or shrinks, with trading results.
Thanks Andreas, I’ve just been reading up on that. I may start small so transactions will be under the duty threshold- I know you’ve said a 10K account is the smallest you’ve simulated but I want to try simulations with a smaller account to see what happens!
Otherwise maybe I’ll stick to the US indices after all ……
OK, so just to make sure I’ve understood that, if you’ve liquidated everything into cash due to market conditions and you now have $120,000 and you start buying again, you put all of that 120K back in?
What happens if you’ve liquidated and are in a drawdown situation? do you bump it back up to the original 100K when you start buying again? i.e. you have to put more money in?
Thanks again. I’ve just left a review on Amazon btw as your book has been a revelation for me !
Think of it as a portfolio. If you sell something, your portfolio will contain a larger cash position. It’s not like the money leaves. So it’s not a matter of taking money out and putting it back in. It’s just a matter of whether the money is in cash or stocks.
If you reach 120k and the rules tell you to sell all positions, that cash remains in the portfolio. When it’s time to buy again, position sizes are based on the 120k. Same if you lost money, and you’re down to 80k before the rules take you out of the market. When you buy again, 80k is your base.
In reality, you should never leave large sums of cash on a broker account, but that’s a whole different story. If your cash amount is lower than your government depositor guarantee, you should be fine. But leave a million in cash at your own peril. If your brokers goes belly up, your cash goes poof.
Thanks for the review, Tim!
Just finished reading your book. Interesting strategy. Can you provide the number of trades by year for the si mutation period? Also, what prices did the simulation use to determine return? End of day?
The simulation always buys or sells at market on open the day after a signal. Valuation of the portfolio is based on end of day prices.
Number of trades can be reduced dramatically if you have a concern regarding commissions, but for the US market, trading is really cheap these days.
Trades per year for the simulation:
2006 – 399
2007 – 430
2008 – 42
2009 – 445
2010 – 489
2011 – 413
2012 – 390
2013 – 402
2014 – 394
Hi Andreas –
Trust you are well.
I’ve been downloading and saving a truncated version of the stock ranking tables since I signed up for the Equity Momentum portfolio a few weeks ago. I managed to miss the stock ranking table of 29 Sept. I wonder if it (or any historical table, for that matter) is available to Equity Momentum subscribers?
Historical ranking tables, eh? Well, all the data needed for that is already in the database, so that shouldn’t be too hard to fix…
Hi Andreas, this is a great book… thank you… I got a paperback and I have just written a 5 star review on amazon where I said it was the best trading book I have ever read… which is true.
I wish I got it earlier… I realized I was making some errors in my backtesting… I wasted a lot of time. Your lesson is very valuable.
I’d like very much to backtest this strategy… but all the nuances you mentioned, like re-balancing, getting the right historical data, etc… made me realize I underestimated the amount of work needed… which makes it tempting for me to just go ahead and start trading it… need to think about it (anyway we are below EMA 200 on the stock index now, so I have some time to test 🙂
Do you think it will make a big difference if I rank the stocks on something simpler, such as price distance from EMA divided by ATR (to lower the rank of stocks that are highly volatile)?
Slippage… with a buy order at market on open, and a larger stocks universe incl. mid caps, even with small position sizes, I may get a high price on less liquid stocks… did you consider any slippage impact in testing? Or you assumed I would get a fill on the open price?
Thank you for the great book and all the best!
Thanks for the review, Michal!
I tried to demonstrate concepts in the book that are not commonly seen in books, but very commonly used in the field. Rebalancing is one of these things of course.
You can rank on different factors are get similar results. Keep the big picture in mind though. The obvious problem you may get with distance to EMA is that it doesn’t say anything about how the price got there. You might get extreme gap situations, takeover plays etc. So in that case, you may need to do a manual filter at the end before trading to get rid of the odd ones. Nothing wrong with that, but hard to back test.
I always use realistic slippage in the simulations, based on realized in-house slippage. In retrospect, that may have been a poor choice for the book, since I don’t want to make those numbers public… I should have used some standard slippage numbers to avoid that. Oh well, doesn’t make a lot of difference for this type of market.
At this point I have ranking tables since 2 Sept. (excepting 29 Sept.). Of note is that of the five stocks below their 100-dma on that day, all are now positive. On that day, one would have to go to the 38th stock on the list to reach the 100% target weighting without going over (96.8%). On 2 Oct., one would have reached near-100% target weighting (99.8%) at the 32nd stock on the list, with no stocks in negative trend. Despite the fact that during this period the .SPX has been under its 200-dsma, are there some inferences that can be made regarding the market; namely, that at least the top tier momentum stocks are recovering?
Well, just speculating. But this leads to a couple of other suggestions: (1) Add an .SPX or SPY chart with a 200-dsma line to easily check on current market regime and (2) Add 100-dsma line to the individual stock charts?
Hi Andreas –
Not complaining, just bitching (it works the other way around too): That ranking table is now two days old! How do you expect a respectable, fee-paying member of SofM Club to get his game on with stale data? My spreadsheet is wilting.
Oh, must be computer glitch…
Thanks Ken. Odd, but the automation tool seems to have failed this morning. I’m running it again now.
And it’s only one day old. 🙂 Today’s Thursday, so data should be as of Wednesday, but it’s just showing as of Tuesday.
Give it ~30 mins from this comment, and all should be updated. I’m rerunning the site automation software.
Great books Andreas. I really enjoyed both the content and the style. I have a basic question concerning the stock momentum system and how you define the gaps that make you “nervous”. I interpret the percentage move as from the lowest close to the highest close in the defined time frame, is that correct? Or is it the gap between the current close and those lowest and highest previous closes? Are you perhaps referring to the lowest Low and highest High ? Thanks for any clarification.
Aha moment, the details don’t really mater. The principle of having a gap filter is what counts?
Andreas, I loved your trend following book and I spent most of today reading Stocks on the Move. Your books have been an amazing find for me. It’s so rare to find a finance book that clearly explains the concepts without getting lost in pages and pages of overly complicated formulas.
My only quibble is that the Kindle version of Stocks on the Move lacks a table of contents. I find myself re-reading sections of your books and a table of contents would help with the electronic version.
Hi Andreas. I’ve had the following code built for me based on what you said in the book for a popular UK based Investment programme. Can you have a quick glance and see if this is in-line? It ‘seems’ to work but when looking at DGE.LSE it seems to give 90 Day Adjusted slope of -0.0048 – yet the chart is clearly in a nice linear trend since August 15?
//@Description:Returns the volatility adjusted momentum
// Care has been taken in preparing this code but it is provided without guarantee.
// You are welcome to modify and extend it. Please add your name as a modifier if you distribute it.
var period = 90;
var maxGap = 15;
if (status == Loading || status == Editing)
period = storage.getAt(0);
maxGap = storage.getAt(1);
if (status == Adding || status == Editing)
dlg = new Dialog(“Settings”,180,55);
dlg.addNumEdit(“VAL2″,8,-1,-1,-1,””,”% Max Gap”,maxGap,0,100);
period = dlg.getValue(“VAL1”);
maxGap = dlg.getValue(“VAL2”);
setTitle(period+” Volatility Adj Momentum”);
var data = share.getPriceArray(period);
//eliminates shares with daily gaps >= maxGap
for (var i=1;i= 1+maxGap/100 || gap<=1-maxGap/100)
var logPrices = ;
//create a trend (linear regression) object and a moving average object
var trend1 = new Trend(period);
var ma1 = new MA(period);
//calculate the trend and average of the log prices
for (var i=0;i<data.length;i++)
logPrices[i] = Math.log(data[i].close);
var av = ma1.getNext(logPrices[i]);
//annualise the slope of the trend
var annSlope = Math.pow(Math.pow(Math.E,trend1.getSlope()),250)-1;
//begin calculating R2
var SStot = 0;
var SSres = 0;
for (var i=0;i<data.length;i++)
var trendVal = trend1.getValue()-(trend1.getSlope()*(period-i-1));
//variance of the average
SStot += Math.pow(av-logPrices[i],2);
//variance of the trend
SSres += Math.pow(trendVal-logPrices[i],2);
var R2 = 1 – (SSres/SStot);
var vam = R2 * annSlope;
It seems the code is using an MA on 'x' period (allows me to set n days). Would this be correct? Or does code make the MA a linear regression MA and not a normal MA?
Hey Mr. W,
I can’t really debug the code, but your number doesn’t sound too odd to me. The code is written for some application that I don’t use, with libraries that I don’t use.
I tried a quick thing though. I just punched open a normal price chart of LSE and drew a linear reg line on it. For the past 90 trading days, I got a linear slope of about -1.07, or around -0.04%. My check is very crude and not meant to properly match the actual number, but it’s close enough.
Sure, it’s up since August but 90 trading days ago start in early July, and we had that failed rally to hurt the ranking number.
Thanks Mr C for the sense check. I wanted to check my SP500 stats with a recent one from your model. Any chance of posting an updated S&P chart or at least a few so that I can correlate and confidence check my code/programme ;-))))
On a related note; having completed a very good trading/portfolio management course with a very well know practitioner/antagonist – part of the strategy is at odds with your recommendation of never shorting stocks? The course is question advocates building spreads based on macro analysis. Top down, economy, country sector etc – using PMIs/ESI/Housing Starts etc. You then pick winning and losing sectors, then drill further to find individual stocks most likely to rise and fall based on your macro read. Then create a spread (long/short) and trade the pairs ratio. Shows promising results but I’m now dubious on the /short part of the spread trade based on your teachings? In the recent Aug sell off the spreads performed as expected and made very good % returns based on the short elements (beta hedging used in building the spread). I’m just after your thoughts on this type of trading, given your models don’t advocate selling short (although these are closely managed short term trades of <3months)!
Thanks in anticipation!
I wouldn’t say never short stocks, just that it’s a very different game with much higher risks and much lower success probabilities. Retail traders shouldn’t do it, generally speaking.
Spread positions is mostly an institutional game. There are some unique risks and problems with it, but of course profitable strategies as well. It’s easy to be fooled by how easy it is to short stocks in theory. In reality, you need to locate the shares, deal with high funding costs, keep track of restricted shares etc. You can’t just run a normal simulation and think that you can go short whatever you like, wherever you like.
In reality, you’ll find many “Great shorting opportunities” to be when a stock wasn’t possible to short anyhow. And the cost… oh, the cost of being short for extended periods of time can completely destroy a trading strategy.
But let’s assume you can deal with this. Well, now you need to leverage up your portfolio to get meaningful return potential. You’d have to take on very large positions. Cash equities aren’t futures, so this can be difficult. You also have to make assumptions about correlations, with may or may not hold up. Suddenly the stocks move in opposite directions, and you get hit on both legs.
I’m not in any way ruling out this type of trading. But this is a game for the professionals, and a very difficult one. For most traders, simply not shorting stocks makes more sense.
Thanks for the elucidation on these facts Andreas – very much in line with my thinking. Last time I tried this on a small live portfolio I got my short side profit ironed out because the stock went ex-divi. I had forgotten the basics of the short seller being liable for the divi payment. No wonder the stock looked like a good short on the chart – and as soon as it went ex-divi it started to bottom out and climb again! Over here the method uses CFDs rather than outright stocks. So easier, but susceptible to the CPP of the broker, albeit said tutor has managed to engineer a good deal with respected London broker, devoid of usual bucket shop tactics (wide bid/offer spreads, stop runs out of kilter with underlying etc). One could argue being long the macro picks and then beta hedging with some cheap OOTM options or index CFDs rather than building individual legs of the spread. Total up long exposure and hedge total long book rather than each leg etc? Plus tutor teaches no stops. Now, in a leveraged bet this makes me very uneasy. The panacea would be a broker allowing stops on the spread ratio rather than having to manually manage two separate legs. This would make risk management much easier.
CFDs can be great, but they’re not without their own unique problems too. Especially CFDs on cash instruments, like stocks.
Keep in mind that CFDs on stocks normally have a funding cost of around 8% p.a. Yep, that’s how much you’ll lose if you hold a CFD on a stock for a year, and the actual stock remained flat. These instruments are for short term trading, not for investing. On the positive side, you’ll get rid of the ridiculous stamp tax. Of course, you can do that by trading US stocks and avoiding the UK too…
CFDs on futures make more sense to me, since the cost is in the spread and there’s no ongoing funding cost.
Andreas, it has been a few months and I continue to think your book is best.
I am planning to implement this system once we are in a bull market again.
One more question if I may… you said you choose Wednesday because it has a 20% probability of being the best day 😉
OK, but… as I did some backtests, I saw significantly different results between weekdays.
Could you please run your program again and compare the results between scenarios of placing orders on Monday, Tue, Wed, Thu and Friday?
I would be great to check if results are in the same range.
Thanks a lot!
Basic question. I am currently reading your ‘Stocks on the Move’ book and I was wondering, do you then use RightEdge for calculating the regression slope and then ordering the results (like an screener) or any other software?
And if I wouldn’t like to do back testing and just apply this logic to a specific index (group of tickers), do you know if there’s any other software that could get the end of data prices from free sites like Yahoo or Google?
I calculate the regression and all else in RightEdge, and I use a custom class for storing and ranking the stocks based on the relevant ananlytic. This class sorts all the stocks based on the best ranking number and buys from the top of the list.
If you don’t want to simulate, but rather just implement the logic, it’s much easier. The easiest of course is to sign up for the automated ranking tables on this site! 🙂 http://www.followingthetrend.com/equity-momentum-report/
Yahoo and Google are ok but not great data quality. They’re probably close enough though, if you just want a ranking table and can live with a few errors here and there.
I was looking at the s&p 600 equity curve for the momentum portfolio and it seems like the strategy has been underperforming since 2006. Why do you think that is and does that raise concerns for applying it on s&p500 in the future?
Thank you… Honestly I am not entirely sure how I stumbled onto your first book, to then find your second book and read it first. Mind you this was only a week ago.
I owe you much thanks for helping me make the ever so need paradigm shift to professional trader thinking… e.g. “simple theories, statistically sounds modeling, back tested w/o being fitted” I do remember I was looking for an alternative to the typical trend following narrative. And I do remember looking for a trend following system that would work on stocks… Which I now believe to be a bit fool hardy. As many other retail traders my tail of woe is filled with snake oil salesman, jumping from idea to idea, failing at all of them, and finally learning there is truth out there… Only to realize I have only seen the tip of the iceberg with regards to real statistical truths about the market. So now with my new found enlightenment, my novice dues have been paid in capital, naivete, and person assessment. I can say I am ready to move forward with a rules based decision model, that fits my capital capacity, and personality.
I will start coding my first strategy this evening…
I really love your book Stocks on the Move, I too am a little bit confused on how to use the ranking methodology. I recall you used to have the spreadsheet available for download to show how you do it. It seems that the spreadsheet is no longer available – what happened?
Is there any way I can still get access to it?
Im wondering this too. I have done a calculation in excel which looks like this:
(LR Slope*250)/Share price*100= Exponential linear regression? then Exp Lin R*Lin R2=Adjusted slope.
Is this correct?
Damien and Mathias,
Sorry for the slow reply…
Try this: http://www.followingthetrend.com/data/AppleRegressionLogic.xlsx
Is it possible to calculate the adjusted slope if you only have:
1. Linear regression slope (120 days) =0,5785
2. R2 (120 days) =0,850
3. Share price (last close) =243,60
59,37*0,850= 50,46 (adjusted slope)
I have used this calculations to rank stocks on the OMX Stockholm exchange. Seems to work or is it completely wrong?
The stock I used in the calc is Boliden (BOLI.XSTO) if you want to look it up.
I am currently +15% this year (started 10 aug) using your momentum strategy with some slight changes.
Using a linear regression is a bad idea. In that case, you’re measuring the change in currency units, in your case Kronor. That means that a stock with a price of 500 will always show a larger value than a stock trading at 10. Do you really care if the price moves up 5 SEK, or are you interested in the percentage move?
That’s why I suggest to use exponential regression.
edit: Yes, I should read the whole post before replying… While your method is technically incorrect, it will give a close enough kind of number. I haven’t tested it, but it probably works ok.
It seems to work well enough and since it´s all I have available right now it´ll have to do. I have started to study C# so hopefully in the future I will be able to do this the proper way.
Tack för hjälpen!
It took me three months to build and test a hybrid version of Andreas’ StocksOnTheMove/Equity25/Equity50 systems with some home-brewed modifications including automated orders and only consuming public data sources Yahoo Finance and Google Finance. After countless backtests and minute checks and double-checks of my code I kick-started the monster on US-Election-Day: 8 November 2016 (any day is a good day to start isn’t it?). This turned out to be a good decision: Big jumps up in current high-momentum shares (e.g. Nvidea, NetApp, BestBuy, Chesapeake, UrbanOutfitters) combined with an solid upward USD-EUR produced a quick >12% profit in only 7 trading days. Part of this is sheer luck and of course it will not continue like this, but it’s great to see such initial jump in the equity curve.
A big thank you to Andreas for sharing his knowledge, experience and ideas with us and I can truly recommend his models and services to any serious and independent investor.
I’m glad to hear it, Peter!
Wow, going live with a new model on a critical day like that. Very brave! I wouldn’t have done it, but I’m glad it worked out.
Just keep in mind that if you had launched the exact same model a month earlier, you’d be down 5% and probably not as happy. A few years ago I launched a similar model on Asian stocks two weeks before the Japanese tsunami… That was a little painful in the short run.
It’s all about the long term though. There will always be shock events up and down and we just can’t control that.
Would be great if you wanted to leave a review on Amazon. These day’s that’s the main way that people find books.
I bought your book Stock on the Move recently and really enjoy reading it. I plan to try out your momentum strategy. But I’m new to the trading and still try to work out some basics. I have trouble to find how to calculate the max gap for the last 90 days mentioned in your book. I would appreciate it if you could show me how it is done.
I used a really simple method: It’s just the percentage difference, close to close, between any two consecutive days. I should perhaps have used a different word here, since this is not what most people would call gap…
That’s great! Thanks for your clarification.