SIF Folio: Rethinking my approach to stop loss rules

For me, the absolute cut off is 20%. From a mathematical perspective, a loss of 20% will require a 25% return to recover from the losses. At 25% this widens to a little over 33%. The maths starts to work against you the larger the loss you accept.  For growth shares, the stop the absolute stop is tighter at 15%.

When thinking about stop losses, it is very important to have portfolio management rules in place. Even with a 20-25% loss accepted on one company, you can still materially lower the risk of loss on the entire portfolio to something around 15%.  A 15% loss on the portfolio can be recovered within a few years of average returns. By making sure you have taken steps to minimise losses, you naturally improve your returns long term.

My book "The Equity Edge" outlines the use of stops and a set of simple, robust portfolio rules that protect the portfolio in downturns. What is unique about these management rules is that they naturally underweight equity exposure in the portfolio in downturns and get back to fully invested quickly, at the right time.

There  is also a section on using a dynamic trend following system for setting good entry and exit points for investments.

Hope this helps.

Mark.

Thank you Roland,
Great article. I read everything I can about stop losses! I`ve been investing for just 3 years now and find stop losses the single biggest conundrum. Articles/books/blogs seem about 50:50 for and against for the medium to long term investor[`like driving without brakes` v `should be called `make losses``] and I change my mind about them on a roughly weekly basis. I have no problem with cutting losses and have been in the habit of cutting anything which is down 20% with an automated stop loss. The problem, which cannot be anticipated, is a big market fall which would potentially stop you out of everything - and then when to buy back in? Having said that, I rather wish I had been stopped out of my reasonably large positions in DART and TEG in the recent downturn as I wouldn`t then have to look at horrendous drawdowns and would have plenty of cash to buy them back at what I think[and hope] are bargain basement levels! I am left with the nagging suspicion that stop losses are both good and bad and which one depends on your investment philosophy, time frame, which strategy the market is favouring, market volatility etc etc - so many variables that you will only know how useful or otherwise they have been in retrospect!

Thanks for your article Roland. Thought you may like this article on stop losses that I was sent not that long ago (can't recall where from!):

http://www.quant-investing.com/blogs/general/2015/02/16/truths-about-stop-losses-that-nobody-wants-to-believe

Summary of findings:

• When applied to a 54 year period a simple stop-loss strategy provided higher returns while at the same time lowering losses substantially
• A trailing stop loss is better than a traditional (loss from purchase price) stop-loss strategy
• The best trailing stop-loss percentage to use is either 15% or 20%
• If you use a pure momentum strategy a stop loss strategy can help you to completely avoid market crashes, and even earn you a small profit while the market loses 50%
• Stop-loss strategies lowers wild down movements in the value of your portfolio, substantially increasing your risk adjusted returns

Haven't applied these findings to my own portfolio (yet) as a test bed, so can't comment from personal experience on efficacy.

Hi Roland

Thank you for taking the time to do the article. I had put an question on Stop Loss's myself in the discussion section back in June when I had a whole reflection of I think there is going to be a second downward lurch and the need to protect capital, and slowly learning more about my own emotion, trading and holding strategy (still evolving). I now realize that the majority of losses across three portfolios was actually in two positions Card Factory (LON:CARD) (your note on profit warnings and just exiting and swallowing pride) and Reach (LON:RCH) (an income holding which cancelled its dividend), and three poorly executed trades with oversize positions and too much going on in my head, all three would have made gains if I had held for ~30 days more.

As a result I set up stops on the majority of the portfolios. There were a number of the stops executed, they were automated with Halifax. Looking back now I found they trigger in early morning trade and you get lower than your designated stop value. The majority happened on two days where the overall market lurched downward (e.g. bad evening in the US triggers London selling at 08:00).

I had the whole was I a genius and exited correctly, or was I a numpty and should have just stuck with the courage of my original conviction debate, and still in part I am now! So I created a shadow portfolio which I deleted at the weekend as it was just a way to either pat myself on the back or beat my head against the wall (again). In conclusion of the ~10 positions, three had bounced back to the original purchase or better Unilever (LON:ULVR) Sage (LON:SGE) and Fevertree Drinks (LON:FEVR) the others Aviva (LON:AV.) and Legal & General (LON:LGEN) were somewhere between the stop loss sell point and purchase and others like Bp (LON:BP.) and Revolution Bars (LON:RBG) had continued down. So really the conclusion was ambiguous at best...

I think that the best conclusion I have, and this is probably taken from one of @Michael Taylor books, is you have to be able to articulate why are you going to sell a share when you first buy it - so why are you buying it in the first place and what are you hoping to achieve from the share. Many of us articulate this is a buy and forget or hold forever, but the reality is we all live a finite time so we are going to sell so is it a situation like @Graham Neary and buying a house, retirement or other major life event. Is it taking a better opportunity, so like your 9 month time-frame, does the stock still meet the rules of your original investment criteria if not exit. Or is there also an element of the stock is not doing what you thought it would e.g. increase in value or pay a dividend and do you continue to hold and think maybe it will turn out ok but remembering that you need a 100% gain to recover a 50% loss... but this does bring in the whole exit at a Y% loss as a blanket, but potentially miss a bounce back (and beyond) or keep hoping and watching the X% become 5xY% if I knew the answer I would only be here as a hobby as I would have long ago had the retirement event, which I haven't.

Tommy

I once read a book a long time ago (over 10yrs) which laid out the 3 golden rules of investing. They were printed on different pages, to add tension I think.

There were:-

No.1. Cut your losses

.

.

No 2. CUT YOUR LOSSES!

.

.

No.3. Run your profits

.

.

My stop loss system is as follows

I have a very strict stop loss system.

I make running stop-loss between 10% and 20%, on almost every purchase.

When I buy a share I look it up on Sharescope and bring the 1-year chart up. I then run a 10% stop loss line under the share price. If it does not fit, I try up to 20% and to get the best fit.

In my spreadsheet, I make a column that checks if the share price moves up, if it does, then I move the stop up too. (running stops)

This works for me. As it takes all the agony out of, "shall I sell or not".

It is a bit more complicated than that, but that the basics.

I am a midterm investor, and on average my profitable trades last 62 days, and the nonprofitable trades last 27 days.

Hi Roland,

I’m sure you are aware that I am something of a nerd, with a big database!

Hence, I could not resist running the SIF folio through the ringer of one of my stop-loss testers.

There were 13 positions where data quality issues meant that I could not test them. Of the remaining 80 , 26 (33%) would have been culled prematurely by a 25% stop loss.

Of these 26 triggered stops 15 (58%) produced an adverse result. In aggregate the SIF folio sold* these 26 positions for £1.108m; with a stop loss in place the result was £1.042m - £1.087m (I will explain further down why I quote a range) , that is -2% to -6% realised from these positions.

This looks very similar to the position I have seen when running similar models on portfolios like NAPS.

A Tale of two “halves”

But there is actually more to it. Perhaps unsurprisingly half of the stops triggered happened this calendar year. Thus far, these do not look successful. These 13 positions in SIF would be valued at (or have been sold at) £639k, whereas the stop losses would have yielded £554k - £581k with 8 of 13 stops disappointing.

The counter-point to this, however, is that prior to this year the stops have done a reasonable job, in more “normal” market times. Sure, 7 out of 13 of the stops produced adverse results, but overall holding to term yielded £469k whereas the stops returned £488k -£507k (a 4-8% improvement).

So on the one hand it looks like there might be some value in the stops in “normal” market conditions. On the other hand three years of good work were wiped out in a short period of market volatility. Could you judge when to turn the strategy on or off? I doubt it.

Outsized Returns.

It is also worth noting that nearly 250% of the Stop Loss improvement pre-2020 was delivered through two positions.

Bilby (LON:BILB) stopped out in September 2018 at a price in the range 91-95p vs SIF eventually selling at 40p.

Staffline (LON:STAF) stopped out in January 2019 at 945p, compared with the eventual SIF close @ 324p.

Did either of these issue a profits warning, that would have made you sell even without a Stop-Loss?

The Cost of Friction.

I should explain that I have modelled this using the daily market opening price (as this is when the gaps occur).

I have quoted two returns for the stop losses, one being the stop target and the other being the actual opening price. In reality the higher target level is unattainable, even if you were using Spread Bet Guaranteed stops, you pay a premium for the privilege.

It is also worth noting, that might prices therefore take no account of the spread, whereas the sales recorded in the fantasy folio do. This is potentially enough to wipe out most of the gains delivered prior to this year.

Overall, therefore, it is hard for me to conclude that stop losses are going to improve your performance; but perhaps you can consider them a relatively inexpensive insurance policy?

Here’s the data

A thorny question Roland.  Like many contributors to this thread I also struggle with stop losses.  I find it very difficult to sell on price movement alone.  Robbie Burns has very successfully mastered the art of killing losses.  At one point I tried to emulate him, but it didn't work.  The stocks I sold at a 10% to 15% loss bounced back significantly.

As often said prices tend to over shoot on the downside and upside.  As a consequence I sold very little during the pandemic melt down.  I continue to hold a few  leisure orientated stocks to my detriment, because I feel that the price does not reflect the intrinsic value of the business, a view that has not changed, although I have been very tempted to hit the sell button.  The principal culprits are Itv (LON:ITV) and £FSTA.  I suspect I may be guilty of a little confirmation bias.  If and when these stocks recover depends very much on how things pan out over the next few years.

Thank you the thought provoking narrative.

Stop loss. 

Agree find it very hard to stick to a system.  I set support and resistance lines and study it hard if breaks either.   But Galliford Try Holdings (LON:GFRD) has fallen through several of my 'support lines' and could not bring myself to take the loss. StockRank and Quality says it's not 'junk' so have taken the long view and ride it out, but only as cannot bring myself to take the hit.  Cash flow could run out however, but hopefully survive to 2021. Trouble with that is when it does rise then everyone and their dog will have got bored and sell off.   So perhaps a system to get out earlier would have been more sensible to protect my capital.   

I think that stop losses don't make much sense from a value investing perspective. If you bought a stock it must be because you have estimated that its intrinsic value is much lower than the current price. If the price falls even further, you want to buy MORE shares, not sell what you have already bought. Markets are hysterical, which means that they will make the prices fluctuate in a chaotic manner. A good trading strategy cannot involve any automatic trigger that unloads an entire position based on such erratic behaviour. Most of the stocks in my portfolio have experienced a crazy downside of even 10-15% over the course of a few days, just to recover immediately afterwards and often even starting a rally. If I had used a stop-loss approach, I would have systematically sold each and everyone of them at a loss.

I agree that applying stop loss rules are a good thing to do in normal times. But this simply begs the question, 'What are normal times?'.

Some say the normal rules of thumb for stop losses are designed for Volatility (VIX practically) under 30. Beyond this threshold the market behaviour becomes un-investable using conventional stop losses. Volatility (as measured by VIX) is now well below 30, so you might think that normal times are here again? That may depend on what you are investing in and when you are invested, Specifically Small-caps in the low volume weeks of high summer: beware!

For your existing positions you could try applying Chandelier Exits based on a volatility threshold (as expressed by a multiple of Average True Range linked to a Sliding Window anchor). Alternatively you could draw trend lines which when violated trigger an alert to Sell.

https://school.stockcharts.com...

Roland, I've just begun to set Stocko's Alerts on my high-flying shares so that when an alert is received I can run the ruler over the share in question to make an informed decision. There's no doubt one must carry forward one's convictions, I wanted to sell all my shares last January, warts and all, and couldn't bring myself to do it.  It was a no-brainer really.  As an example, an alert set to a fall of 17% on Mysale (LON:MYSL) triggered earlier this month, I sold, being nervous about the Australian economy and redirected the funds into a couple of gold miners.  I've since been able to set up alerts on these. and may use your 25% figure for them.  So there are benefits to be had in having some kind of prompter.

Thanks for your article!

Hi Roland,  thank you for your article.  You raise an interesting question to which, I feel, there is neither a quick nor absolute answer.  There are many technical and personal factors at play.  One matter that I feel is definitely worth consideration is position and portfolio size rather than a fixed percentage from the buy/entry point.  On this basis you might argue therefore that stops should be triggered at say a loss of perhaps 1-3% of your portfolio.  After all, you can always buy back.  

Knowing when to exit a losing position is always difficult. Rather than setting a 25% or other stop loss on all holdings, is it not better to decide an appropriate loss based on the volatility of each stock at the time of making the investment/ As you know, many stocks are unlikely to drop more than 25%  in normal (if there is a normal) and so a stop at 25% would keep one in the position. If there is a likely volatility of say 50% and that risk is too high for comfort, the don't invest. I have frequently been caught out by placing too short a stop loss, all stocks need space to breathe and have had more success tailoring my investment sizes and stop losses after carefully looking at the volatility before investing. 

Great question Roland. There is probably more than one way to skin a cat re stop losses, but its not good to feel like a rabbit in the headlights when a serious market crash is unfolding. You need a plan that feels right for you, and suits your attitude to risk.

With  stop losses, I think that you need to decide if you are defending against losses in the overall market (or sector), or defending against a company that might be going to the wall.  

Re the market as a whole, if you believe that the market will rise in the long term, and you hold index funds (or companies that closely correlate the market), and you have spare cash, then you should invest after (or during) a crash. "Buy when there is blood in the streets" - to use Ray Dalio's term. If, later on, you find yourself enjoying spectacular returns (ie you feel that you are comfortably satisfying your investment objectives), you should shift some funds into cash (in preparation for a rainy day) - dare I suggest you go out and spend some of it? 

If large cash holdings aren't your thing, then diversification into  defensives and precious metals can prove useful insurance measures that help to smooth out excessive volatility and reduce those sleepless nights.

Regarding specific companies, getting out as soon as there is a profits warning is sensible. Unfortunately, retail traders are often the last to know. So in the absence of a profits warning, (or any other explained event) a sharp fall in share price relative to the company's sector is sometimes the only alert we have to help us. Here a blanket percentage figure, whilst simple to apply, might not appropriately factor a company's share price volatility, its sector, individual circumstances and overall long-term value. Setting a figure is tricky, but it is certainly sensible to have a level tailored to each company, at which you determine that the investment was a mistake and its time to get out.

Oh, and another thing. People don't think straight after market crashes. But if you can keep your head........etc, etc

Sounds sensible to me - and like you - I am holding onto £MGNS    One of the frustrations I have found is that having set a SL figure -  the price plummets far below, and by the time you get a chance to sell - its not 25%  down but much more.......

Hi Roland

As we all know selling is much harder than buying. Behavioural biases make it doubly difficult.

As HumourMe mentions 'The Art of Execution' by Lee Freeman-Shor is an interesting and relatively quick read and tackles this question.

He analysed the investing habits and returns of 45 top investors (over 1800 investments and over 30000 trades). Despite their experience and skill less than 50% of their investments made money. The key to the most successful was to cut their losses quickly and dispassionately and to top slice their winners as they rose in value rather than sell too early.

The most effective way to cut your losses was to have a stop-loss set somewhere between 20 and 33% (the range used by the best investors he analysed). To be clear this was not a soft stop-loss, each investor had his or her preference.

To avoid all behavioural biases this stop-loss was set at the time the share was bought. So a suggestion would be that for your SIF folio and to maintain full transparency you buy the share as you state and at the same time set a stated percentage stop-loss.

I suspect this would improve your returns but of course only time would tell :-)

Many thanks for your regular SIF folio updates.

Regards

Iron

With stops you need to answer "what are you guarding against?".

As crashes tend to cause all shares to correlate downwards (and continue) an alternative to consider (to lock in gains and sit in safety) is to sell everything when the portfolio is down X%. You need a rule to get back in though.

"The Art of Execution: How the world's best investors get it wrong and still make millions in the markets" makes for interesting reading on losses. If they come into play  the market is disagreeing with your assessment, as if you expected a fall to occur that breaks any threshold, surely you would have waited to buy? I can see for value investors (with a portfolio) stops make less sense. If a strategy relies on momentum as a component then lots of shares triggering stops is a sign that the portfolio isn't doing what was expected.

My conclusion is that if you are timing entry then they can be tighter. If they are for capital preservation they can be position based (i.e. per share) or portfolio based.  Or both. If position based 10%-25% trailing, based on volatility may be necessary. If portfolio based it could be a lot less. Portfolio based might also be easier to manage.

Hi Roland, 

I've really enjoyed all your articles this year and found it very useful to see you run the SIFfolio.  

As a 'long term investor' (typically holding stocks from 3-8 years). I find it difficult to apply stop losses on price action. If the investment thesis has not changed I would probably add to the position at the low price. I am more likely to sell out if and when the investment thesis is found to be incorrect or changed substantially.

I would agree with you though if one is using a rules-based systemic way to build the SIF folio then a stop loss should help performance rather then hinder it.

Is it possible to post a year to date view of the performance of the portfolio?

Many thanks

BSV