Some StockRank Portfolio Strategy Returns - 2013 To 2017 | Edward Croft

20th Jan '17 by Edward Croft

34 comments 22813 reads

We’re working on some internal data research tools at Stockopedia which will contribute to a lot of editorial insights, and we hope, some really great product launches in future. Here’s a few insights from my own recent research into the performance of UK StockRanks strategies in the last few years.

(Disclaimer: Past performance not an indicator of future results etc).

1. Sector Diversification - does it reduce returns ?

How did a NAPS portfolio process (picking the top 2 stocks from each sector) stack up against selecting the top 20 ranked stocks without any additional diversification ? Here’s a chart of the NAPS based approach…

Meanwhile the top 20 by StockRank without additional diversification:

The performance was very similar, but the sector diversified portfolio performed much more smoothly. The “Top Ranked” portfolio was only invested in an average 3 sectors throughout the whole period, whereas the NAPS invested in 10 sectors consistently. Essentially, the NAPS process led to stronger risk-adjusted-returns.

The observant will know that my New Year’s NAPS portfolio (written up here) returned only just over 5% in 2017. The reason it didn’t keep up with the first chart above was because I meddled with my original rules ! Meddling with rules hasn’t paid.

2. Just 10 stocks - do concentrated portfolios outperform ?

Does extra breadth in a portfolio really add any performance ? I’ve long argued that if you are a systematic investor it makes sense to broadly diversify. If you are genuinely a talented stock picker (which you probably aren’t) then you should concentrate. Here’s a chart of a Top 10 only NAPS portfolio.

A nice performance, but it shows a significantly reduced return versus a more broadly diversified 20 stock portfolio and a lot more volatility. Concentration hasn’t paid.

3. Microcaps - is there an edge in hard to trade shares?

All the above tests have a minimum £50m market cap cutoff - i.e. they are excluding the microcaps. If we drop that to a £10m cutoff to let in the top ranking microcap shares how would it have done ?

It’s squeezed a bit more performance, but I’d argue that the additional transaction costs and volatility may not…

Unlock with Email

Unlock with Google

Already have an account?
Login here

Stock

kelly

gain

sharpe

CTH

3.44

1.09

0.0339

CAML

5.41

1.22

0.0649

FCRM

7.49

1.39

0.0984

IQE

10.5

1.47

0.126

DRX

11.8

1.48

0.135

QRT

13.1

1.42

0.134

MCLS

1.37

0.132

INDV

13.2

1.4

0.131

III

12.9

1.4

0.13

REC

13.6

1.37

0.13

EXI

1.35

0.124

MNDI

12.8

1.33

0.119

TEP

12.4

1.31

0.113

GAW

12.6

1.28

0.111

PFC

11.3

1.26

0.102

MANX

10.6

1.23

0.0917

FIF

10.1

1.2

0.0853

CCC

8.82

1.17

0.0743

MGNS

7.91

1.15

0.0664

WIZZ

6.74

1.14

0.0583

unwise2

20th Jan '17

Hi Ed

Interesting stuff as always. Can I just clarify the rules used to construct the performance of the first graph? Other than top 2 in each sector and Mcap over £50m was there anything else such as maximum bid/offer spread?

Regards

Unwise

underscored

I very much want to see an internationally diversified NAPS performance. Any further smoothing of performance?

Edward Croft

No I ignored the spread rule in all these charts... it's a point of further testing. The £50m Market Cap should ensure the spreads are fairly tight . I'll try and post some more detail later.

Jon1958

21st Jan '17

Thanks Ed. I am running a small real portfolio along the lines of your randomised portfolio above. I can't bring myself to go entirely random but inject a discipline that I can only buy a stock if it is ranked 90 or above. will let you know how I get on.

Nick Ray

Very interesting results. I have a suspicion that the optimum size of portfolio might be very dependent on the kind of rules used to build the portfolios.

It seems clear that there should be an ideal size of portfolio because below a certain size you will have insufficient diversity and hence too much volatility; but conversely if the portfolio has too many stocks then you will approach the index return and volatility and you have effectively built yourself a mini tracker fund. But what is that optimum size and how sensitive is it to other choices? For example, my own experiments seem to suggest that ten stocks is enough provided that you select from high Q stocks, but you need more if you select from a wider base, both of which tend to move the portfolio closer to the index return and volatility.

I have ranked the naps2017 selection by the estimated 256-day sharpe ratio of the combined portfolio as I add more stocks (So the table below is cumulative. Each row describes the effect of adding the stock on that row to all the stocks on the previous rows.) This suggests that you might only need ten stocks (the ten in blue) to get similar performance. So this can act as a hostage to fortune in 2017 if it manages to omit all the best performers going forwards!

Stock	kelly	gain	sharpe
CTH	3.44	1.09	0.0339
CAML	5.41	1.22	0.0649
FCRM	7.49	1.39	0.0984
IQE	10.5	1.47	0.126
DRX	11.8	1.48	0.135
QRT	13.1	1.42	0.134
MCLS	14	1.37	0.132
INDV	13.2	1.4	0.131
III	12.9	1.4	0.13
REC	13.6	1.37	0.13
EXI	13	1.35	0.124
MNDI	12.8	1.33	0.119
TEP	12.4	1.31	0.113
GAW	12.6	1.28	0.111
PFC	11.3	1.26	0.102
MANX	10.6	1.23	0.0917
FIF	10.1	1.2	0.0853
CCC	8.82	1.17	0.0743
MGNS	7.91	1.15	0.0664
WIZZ	6.74	1.14	0.0583

mpat89

I have a subscription to Stockopedia USA but I am not able to find performance comparable to the UK. Can anyone share their experience?

I really want some information on the US. I have a US subscription but I I get the impression that I need to be a stock picker to win in this market. Maybe because factors are so big now?

The Stock Rock performance is not clear like with the UK, see: http://www.stockopedia.com/stockranks/performance/?exchanges=NYQ%2CASQ%2CPCQ%2CNAQ%2CNMQ%2CNSQ&field=_QVMScore_lo?=50&period=1&buckets=deciles

Yes it's definitely a much more efficient market. We've found strong factor spreads in almost all international markets except the USA - the spread is much more moderate in the USA.

Why would this happen? The US has much more standardised reporting, instant data delivery from the SEC in XML format, and a much broader and more mature quantitative / algorithmic investing environment. I anticipate that all markets will eventually become more and more efficient and tend towards the USA.

In the meantime we're working on improving our ranks to be more regionally defined... I think we can squeeze more returns from US stocks. On top of improving the factor returns, there's quite a bit that can be done via alpha cloning - but we need to release SmartMoney for that.

Thanks Ed, that was a really useful reply.

So would you agree with the conclusion I am beginning to form, at least for the time being, that the US is better suited to stock picking. For example, a concentrated portfolio of good ideas using execution ideas from the book you previously recommended 'The Art of Execution'.

By Alpha Cloning are you referring to ideas such as following institutional money and common 'best ideas'?

Nick - thanks for the analysis. I'm planning on adding position size calculators in future, and include volatility weighting and mean variance weighting as options too. Just need time and/or more resources.

In the meantime I'm planning on exploring how returns could have differed using different weighting approaches. Watch this space.

bluetonto

Just joined Stockopedia and must say i am very impressed, still finding my way through lots of pages and information .Is there a page that shows your nap selection portfolio for 2017 on a daily /monthly performance ?
which would be very interesting to follow.

Ed - this sounds very promising indeed. I am finding mean/variance ideas to work very well as an alternative to the momentum component of the QVM triangle.

Incidentally, the (complete) naps2017 selection has an extremely low variance. Obviously history is not an exact guide to the future but the signs are that it could be a very stable performer.

gus 1065

Hi Blue,

Welcome to Stockopedia.

What I have done in the past is put together a shadow portfolio using the Folios feature above picking a selection of say 20 stocks from the Naps portfolio then buying a notional £1,000 of each (and depositing say £20k to cover the purchase). It will only take a few minutes to set up and the you can monitor all of the stocks daily and on a cumulative basis and also an overall P&L performance. It would also give you a live education on how to use this particular feature.

Best,

Gus.

stock_Hunter

22nd Jan '17

HI Ed,

No Alert in the chart? price and volume alert is very important ....

Howard Adams

Hi ED

Great article as usual, many thanks.

With regard to volatility weighting, you might like to look at the work done with TradeStops https://tradestops.com/ which centres around a VQ (Volatility Quiotient) per stock (or ETF, Investment Trust or even index).

Using this VQ as a basis, the site also offers support to examine buy size, position size within portfolio as well as asset allocation.

I appreciate you might regard TradeStops is a bit of competition, but Stockopedia is unquestionably far superior across the board than TradeStops. Some if its ideas however, do offer a useful augmentation particularly for risk/reward balancing within a portfolio, making visible volatilities between different stocks and quite sophisticated alerting based not just on price movements but also historical stock volatilities.

(disclosure I am a subscriber to TradeStops but have no other connection with it).

Regards
Howard

Some StockRank Portfolio Strategy Returns - 2013 to 2017

1. Sector Diversification - does it reduce returns ?

2. Just 10 stocks - do concentrated portfolios outperform ?

3. Microcaps - is there an edge in hard to trade shares?

Unlock the rest of this article with a 14 day trial

About the Author

Edward Croft

34 comments

Unlock the metrics for all these stocks and more...

Unlock the metrics for all these stocks and more...