1Yr·

The thing with the baking tests

@Paketknecht highlighted the advantage of boring stocks over tech stocks (same return, less stress) in his article, which is well worth reading, and backed up this assertion with a backtest. https://getqu.in/VKv17a/

But backtests are a tricky thing. As with statistics, you should never trust a backtest that you haven't optimized for your own opinion ;-)

So I tried to do the backtest myself and examined the robustness of the result shown. For the long tests, I replaced CSCO with AMGN because it was also a (bio)tech stock in the late 90s and the data goes back to 1988. Here are my findings:

1) The result "boring stocks achieve the same as tech stocks" applies to the 1999-2024 period shown. But only to this one! 1988-2024 the tech portfolio outperformed the boring portfolio by a factor of 7(!) (18.8%pa vs. 12.6%pa). 2009-2024 the factor is 5 (23.5%pa vs. 10.7%pa). The selection of the time window for backtests is therefore crucial.

2) An argument against boring stocks? Not at all! If you replace just one stock in the boring portfolio with a tech stock, e.g. Unilever with Microsoft, the 1988-2024 return increases from 12.6%pa to 18.1%pa. The selection of individual stocks for the backtest is therefore also decisive (caution: survivership bias!)

3) Since the question about the savings plan came up: Assuming a starting capital of $10,000 and a monthly savings installment of $1,000 from 1999 onwards, the final result is as follows. Long-term portfolio: 1,307,000$, tech stock portfolio: 1,924,000$. Now the tech portfolio is ahead precisely because of the high volatility!

4) If you run the stocks through the portfolio optimizer and calculate the allocation with the highest return at 20% vola pa. in the period 1999-24, you get 14.9%pa. With which allocation? McDonalds 40%; Procter&Gamble 20%; Microsoft 25%; AMD 15%. Who would have thought it? An (almost) even split between boring stocks and tech stocks would have been the best option.

So: always diversify across assets and strategies! That's the best way to go.

(That's the result of this test, but as I said: it's best to check for yourself ;-))

Here are the links to test for yourself:

Boring stocks vs. tech stocks: https://www.portfoliovisualizer.com/backtest-portfolio?s=y&sl=7AIueJOT6WGykW8hNQ7cOYAs a savings plan: https://www.portfoliovisualizer.com/backtest-portfolio?s=y&sl=2QmAGEL9NIkQw2DMSQlqypOptimal portfolio: https://www.portfoliovisualizer.com/optimize-portfolio?s=y&sl=3Gh9m1BJnHHClsbNzp2nDW

app.getquin.comEchtzeit Portfolio Tracker

24 Comments

Nostradonkey💩@DonkeyInvestor

1Yr

That goes straight to the #gqevergreens

•

Yassin@RisingAktie

1Yr

@DonkeyInvestor It's almost like a knighthood from the Queen herself😂🤭

••

Epi@Epi

1Yr

@RisingAktie I'm already feeling quite chivalrous myself.

••

Nostradonkey💩@DonkeyInvestor

1Yr

@Epi I thought like the Queen

••

Nostradonkey💩@DonkeyInvestor

1Yr

@RisingAktie so it is

••

Gamla S.@gamlasvensk

1Yr

Great example! Thank you

•

Sparfuchs@Reinecke

1Yr

@ccf even if it no longer exists.

•

Gamla S.@gamlasvensk

1Yr

@Reinecke I didn't even notice that this was discontinued 🙈

••

Aktienmasseur@Aktienmasseur

1Yr

I see little added value in such backtesting. Past performance provides zero insight into future performance. Furthermore, stock picking was also carried out here and a one-off case was built up. It would be more interesting to see whether there are stocks and sectors that perform just as well in the future.

•

Epi@Epi

1Yr

@Aktienmasseur It's not as simple as saying that past performance provides zero insight into future performance 1.
only the data from the past to assess future prospects. 2. momentum is an anomaly that has already been confirmed hundreds of times. 3. past performance does reveal the character of a company: is it growing or managing its portfolio, is it cyclical or in decline?

•

erT@erT

1Yr

Is it even possible to compare a Tec share from 1988 with one from 2024 and 2054? The backtest is in all honor and I am a friend of such valid statistics, but we also have to ask ourselves what we can actually extrapolate from it into the future and what value it has for our future investments. After all, the whole market is subject to constant change.

•

Epi@Epi

1Yr

@erT I agree, of course. However, there are sectors that are easy to predict. People needed toothpaste and toilet paper 20 years ago and will still need them in 20 years' time. Smartphones are a different story. If glasses, a neural projection or whatever comes along - nobody knows. This risk exists and you can pocket the premium for it. In the end, you know what has prevailed. Of course, the backtest has a bias: AOL, Yahoo, General Electric - all the stars of the late 90s are not included, nobody could have known that at the time.

••

JM@jm_finance

1Yr

Perfect info, thank you :-)

•

Eric@Weisswurstsepp

1Yr

Thanks for the contribution! I feel confirmed in my intuitive thoughts and the knowledge and experience I have built up 😌. Although these are actually the basic rules of investing: Spread broadly across as many asset classes as possible and diversify again somewhat within these. A one-sided strategy requires a lot of effort and time, is often risky and is usually only temporarily successful (see your backtest) 😌

•

Mister_ultra@Mister_ultra

1Yr

The statement that a healthy mix over boring and tech is the best made me happier than I thought 😂

•

M47@Jiyanask

1Yr

Backtesting is for shit.... You can set the parameters to the past so that you make 100% return on average but you can't predict the future and or mirror from the past to the future

I once played through in 2021 from 2018 to 2021 I had 100% average return....2022 and 2023 tested and didn't even beat MSCI. why? I had set the parameters for 2018 to 2021 so that I had the perfect setting but no longer for 2022 and 2023....I can do the game again today from 2021-2023 but still can't predict 2024++....

•

Tim@investorsapiens

1Yr

Backtests are useless because they are only designed for a very specific situation that will never occur again. In this case, the shares and the time periods were chosen precisely so that the story works.

••

Epi@Epi

1Yr

@investorsapiens I wouldn't say that in general. Backtests can show, for example, how an allocation reacts to crises or performs in a general boom. History does not repeat itself, but it does rhyme.

••

Tim@investorsapiens

1Yr

@Epi The story in the quoted article is very artificial. I don't want to know how long it took to find suitable stocks that fit the picture and confirm the story. Why no Google? Why IBM and Cisco? Why AMD and not NVIDIA? Why no Apple or Amazon? I can tell you, otherwise the story simply wouldn't have made sense. But you already wrote that in your test. I've been in the business for 16 years and every strategy and story, no matter how good it looked and worked in the backtest, never worked in practice. I'm not saying that consumer stocks are bad. I'm just talking about the story that's being told.

•

Epi@Epi

1Yr

@investorsapiens You're basically right, of course. However, backtests are not as pointless as you think. Because every time we look into the future, we necessarily look at the past first - that's all we have. Backtests try to rationalize and quantify this view. This has its limits, of course, but it also has possibilities and, in my opinion, we should make use of them.

••

Deleted User

1Yr

Comment was deleted

Epi@Epi

1Yr

@PowerWordChill I assume that boring stocks meant stocks with a steady but low performance. The problem with this is that some of the stocks that have been boring in the last 25 years were not necessarily boring in the 25 years before that. So you only know afterwards which stock belongs where. Will Apple be the boring stock of the next 25 years? Could be. Or not...

••

Paketknecht@Paketknecht

1Yr

Comment was deleted

Epi@Epi

1Yr

@Paketknecht With pleasure! The topic is complex and perhaps we can establish an interesting discussion here?

••