Timing the market using moving averages - an experiment

[Wolverine]

So, I was wondering today why I hadn’t received a notification for the divesting of the funds yet and it appears that I have failed you, my friends…

Changing the investing mode from automatic to manual takes a few minutes and I forgot to check back in after that to actually order the sale of the assets (because without the actual order, entering manual mode doesn’t do me any good).

I’ll add a curve for the theoretical place where the strategy should be, though this is also an example of the kinds of risks I am exposed to with these strategies. I guess one could automate the sales and buys and link it to the triggers but I don’t have much trust in automated trading.

[Wolverine]

I guess it would be a good idea to make the June update before an entry trigger gets hit. Sorry for the delay.

Triggers

It seems people in June have decided that inflation was a bigger deal than expected (shocker, I know) and the passive strategy took a big hit around the 10th-13th, which took us into bear market territory. The 200D-10 has been in cash all month long. The 20D-4 has started the month invested, then hit an exit trigger on which I failed to act immediately (I took a 2 days delay). Both the 200D-10 and 20D-4 are currently sitting in cash.

Time-weighted returns

The dashed yellow line is where the 20D-4 would be sitting if I hadn’t missed the exit trigger, the full line is where I am actually sitting. Since stocks can go up bigly in a short timeframe, it’s not possible yet to assess if these exit moves where beneficial or not when compared to the passive strategy.

Real data

New contributions tend to offset the losses. I’ll call where I stand “flat”. Exiting with the 200D-10 provides downside protection but could also cost some gains upon re-entry.

Mood: still pretty detached, as displayed by me missing a trigger and this late update.

	20D-4	Th. 20D-4	200D-10	Bench.
TW Returns
CAGR	-1.5%	-0.4%	6.0%	2.1%
June 22	-8.1%	-6.4%	0.0%	-7.1%
YTD	-20.2%	-18.7%	-10.3%	-15.5%
Max DD
All time	-22.3%	-20.4%	-13.0%	-18.6%
June 22	-10.1%	-8.0%	0.0%	-10.1%
YTD	-21.9%	-20.0%	-12.5%	-18.2%

I’ve added the “Th. 20D-4” column for where the 20D-4 would have stood if applied adequately on each trigger. Both it and the actual 20D-4 results have entered negative CAGR, though I’m sure it’ll get positive again when stocks will have recovered.

[Wolverine]

Well, just in time. Entry trigger on the 20D-4 hit:

[Wolverine]

I am very late for the July update. July’s been an up month, and so has been August until last week (though we are now trending down so things may happen in the near future).

Quarterly fees have been collected by Frankly, so even out of the market strategies have had a small decrease.

Situation at the start of the month
The 20D SMA -4 strategy started off out of the market.
The 200D SMA -10 strategy started and ended the month out of the market. It is still uninvested as of now.

Triggers

The 20D-4 strategy hit a trigger on the 8th of July and got invested by the 13th.
The 200D-10 strategy stayed uninvested all month long (and still is).

Time-weighted returns

The 20D-4 exit and re-entry has been at a loss of roughly 2.5% vs blindly staying invested. This was helped by me forgetting to exit on time in June but I consider that part of the game and am registering the full loss as a result of the attempt. It has now triggered 4 times and only the 3rd has been successful. The total loss vs the market is currently at -8.3%.

Nothing can be said about the 200D-10 strategy’s behavior until it re-enters the market, which may take some time.

Real data

Back to all time highs with the 200D-10 strategy and the benchmark by the power of new contributions at the start of the investing career, when accrued assets are still low.

Mood: pretty detached (hence the late updates). The bottom could have been reached or this may just be a bear rally, waiting for late autumn to fall lower.

	20D-4	Th. 20D-4	200D-10	Bench.
TW Returns
CAGR	0.3%	1.3%	5.3%	4.4%
July 22	3.0%	3.0%	-0.1%	4.4%
YTD	-17.8%	-16.3%	-10.4%	-11.8%
Max DD
All time	-22.4%	-20.9%	-13.0%	-18.6%
July 22	-1.7%	-1.7%	-0.1%	-2.3%
YTD	-22.0%	-20.5%	-12.5%	-18.2%

Back in (slightly) positive CAGR territory for the 20D-4 strategy. The max drawdown of the 20D-4 strategy has hit a new low at -22.3%.

For the curious among you, this is where we stand today, on the 23rd of August:

There is room to go for both strategies but if we keep going down the way we have these last two days, the 20D-4 exit trigger may be hit shortly.

[Wolverine]

August update. You’ll probably hear more from me very soon since the 20D-4 trigger is close to get hit.

August was an up month until Jerome Powell talked and made it very clear that the FED was not going to shift to an easing policy after just one month of not negative US CPI evolution. As far as I’m concerned, the market is now making sense again and despite the recent quick down trend, I find there is a good amount of comfort in that thought.

Situation at the start of the month
The 20D SMA -4 strategy was invested.
The 200D SMA-10 strategy was out of the market.

Triggers

No trigger has been hit. The 200D-10 strategy needs a lot to move and the June-August uptrend wasn’t quite enough for that. The 20D-4 strategy is poised to cross the exit trigger again soon.

Time-weighted returns

The buy and hold benchmark strategy has crossed over and back under the 200D-10 strategy, though there’s no conclusion that can be drawn for now, given how insensitive the 200D-10 strategy is.
The 20D-4 strategy is loosing for now.

Real data

New contributions still have a big impact, lessening the effect of market fluctuations. There’s no really significant difference in the outcome of the 3 strategies for now.

Mood: interested again, though in a positive way (the market reactions seem to align with my outlook of the economic situation, so things are getting more into what I feel is my control zone again). My bet is that we’ll have a general downtrend until roughly March next year, as the winter and the effects of the tightening policies of central banks start to affect investors confidence and available assets. That could come with temporary and powerful uptrends, though. The market indeed can remain irrational longer than I can remain solvent, afterall.

	20D-4	Th. 20D-4	200D-10	Bench.
TW Returns
CAGR	-0.8%	0.2%	5.1%	3.1%
August 22	-2.0%	-2.0%	0.0%	-2.0%
YTD	-19.4%	-17.9%	-10.4%	-13.5%
Max DD
All time	-22.4%	-20.9%	-13.0%	-18.6%
August 22	-4.7%	-4.7%	0.0%	-4.7%
YTD	-22.0%	-20.5%	-12.5%	-18.2%

A down month overall.

[expatriate]

Thanks for keeping up posting!

[Wolverine]

It fought well and hard to the point of really baffling me but the 20D-4 Strategy finally surrendered and breached its exit point.

It may have had a bit more resiliency built in it than I would have liked in these circumstances. xD

[ItalianEngineer]

This trial is getting more and more interesting.

Did you also studied other numerical methods before going for moving averages?

[Wolverine]

I’ve restrained myself to moving averages, though I’ve tested simple and several kinds of exponential moving averages, with signals either at intersections with the main curve, a smoothed version of it, a translated version of it or several moving averages.

These are the two strategies I have the most confidence in, with more confidence in the 20D-4 SMA than in the 200D-10. One of the assumption was that results would only present any sort of significance over a long period of time, going through several kinds of market conditions. It is fortuitous that we are going through this this early.

[ItalianEngineer]

Now that you have all these data, whtly are you not doing a numerical exercise of implementing the other strategies you mentioned?
I am pretty sure for you, looking the good way you manage numbers, it should not be a big problem to implement it in excel or any coding language

I hope you will take this suggestion,being sure that I would not be the only one interested in this here

Thanks in advance Wolverine

[Wolverine]

I’m currently using my focus on trying to understand the macro economic factors of the situation we are in, how market participants react to them and how their psychology and mine evolve as time passes by and things happen.

After that, I’d like to devote some time to designing a risk partiy portfolio centered on a somewhat risk averse swiss investor with a home country bia.

I’d also like to take some time and get into Julianek’s series on valuing a business.

I don’t have much time nor interest for other technical models at the time being. I know @kraphael is making use of MACD indicators, though it seems to suit him well and he may not be looking for developping other tools.

Also, and most of all, these models/strategies function mainly in pair with our own psychology. Am I able to follow it? Does it give me more leeway to try more risky strategies on the side? Is it eating all of my time and not efficient on a maximization of enjoyment of life basis? You don’t get those answers with paper tracking, you have to have real money invested.

I don’t see myself pursuing those other strategies in the future, though this is only my current forecast and time may, or may not, change my mind.

[ItalianEngineer]

You have a point. The time invested in this is definitely non negligible. That’s why I was mentioning to do it a posteriori now that you have all data pretty well organized. Just as a numerical exercise.

Is there anyone else in this forum, on top of what you mentioned already, using this kind of analysis?

[FI_yes_RE_maybe]

@ItalianEngineer : i’m actually using similar models to make my investment decisions…

I’m basically combining moving averages (a shorter one and a longer one to get a better feeling in what direction the market is moving) which then gives me entry and exit point.
As moving averages tend to be somehow slow to react in stress situation, i’v developped a pretty good “low-point-indicator”… those low points are only relevant in a phase where you are not invested and tend to appear all couple of years - but only when there is too much stress in the market (did not work in the financial crisis for example, but well in Eurocrisis, as well as during the sell off at the end of 2018). The last 2 ones where:

• in March (Market Low at the time before the rebound) and
• during the covid-low in 2020

The model is such that there are only very few transactions over time… this year would have been:

• out 1st of February
• in March low (due to low-point indicator) for 3 weeks as there was no “in” signal coming from the MA and then out again
• still sitting on the sidelide…

So i’m pretty fond of such models as it is empirically proven that momentum has an edge in investing…but i won’t elaborate much more on it as we are looking at a specific experiment here.

[ItalianEngineer]

Eventually in a different post or in private, I would. Interested in knowing more about this low point indicators you developed.
Literature background? Or fully developed by you?

Are you doing all these manually or through programmed trading sw?

[Wolverine]

@ItalianEngineer

Are you doing all these manually or through programmed trading sw?

I’ve been assuming the question was directed toward @FI_yes_RE_maybe but just in case, everything is manual for me. Both because I like it that way and because I don’t have the desire to develop the skillset required to automating it while still understanding what’s going on behind the scenes.

Alright, folks, it’s update time and boy are these times, well, interesting. September was a down month, big time, as I assume US market participants finally came to terms with the idea that the FED was going to tighten for real. That is, until the first days of October but they seem to have settled again.

Situation at the start of the month
The 20D SMA -4 strategy was invested.
The 200D SMA -10 strategy was out of the market.

Triggers

The 20D-4 strategy proved excrutiatingly resilient to triggering its exit and only got divested on the 27th. It just barely stayed so at the start of this week.

The 200D-10 strategy didn’t get close to triggering its entry. As expected, it is way less subject to volatility, allowing for potential greater relative losses on a miss but also greater peace of mind by not being subject to the daily swings of a market like the high volatility one we’re having now.

Time-weighted returns

It’s still too early to draw conclusions on that front. The 20D-4 strategy is loosing for now and will have some work to do if it is to get ahead out of this specific downturn.

Real data

Not much to say. Looking from this perspective, being out of the market and having new contributions actually taking my assets higher looks good, though I’m a bit detached from it these days.

Mood: not gonna lie, I’m not sleeping much with the kind of gyriations we are having right now. Not really because of this experiment, nor because of the buy and hold part of my portfolio but because I’m trying to grow a small betting position I have on the discrepency between where the market is and my lecture of it. Trying to read the market is excrutiating.

	20D-4	Th. 20D-4	200D-10	Bench.
TW Returns
CAGR	-4.6%	-3.7%	4.9%	-1.2%
September 22	-7.3%	-7.3%	0.0%	-7.6%
YTD	-25.3%	-23.9%	-10.4%	-20.1%
Max DD
All time	-26.0%	-24.7%	-11.3%	-20.9%
September 22	-8.4%	-8.4%	0.0%	-8.7%
YTD	-25.6%	-24.3%	-20.5%	-20.5%

Red all around and new lows. Let’s see if we are heading into Downtober or Octuper! As of right now, all bets are off.

[Wolverine]

Well, the 20D-4 entry trigger just got crossed. This is not a 2008-like crisis. (I had used 2008 for my backtesting).

[ItalianEngineer]

When you say that it is not like the 2008, what do you mean in practical terms, from the moving average perspective?

Can you please share your back calculations on the 2008 and eventually highlight what you mean?

Thanks a lot for the great effort you are doing for this community. Many of us read your post, but we tend to forget to thank you for sharing such good, extensive and continuous job. Thanks!

[Wolverine]

That data is on an “old” hard drive on which I will have to get my hand again. I’ll try to answer within a week but that might take a bit longer.

[Wolverine]

Howdy folks? Long time no see!

Things have happened that got in the way of updating this thread, mostly life but also a software update that did outclass my smartphone and made me unable to access my account data (but I still had access to the benchmark fund, so all is good).

No trigger was hit during that time, until today. The entry point of the 200D‐10 strategy has been hit and all funds will be invested going further, until the next exit trigger gets hit.

I’ll do more reporting this weekend but that’s where things stand for now.

[Wolverine]

I have to come clean and announce that this experiment has been brought to an end. I had lost track of it when I changed computer early this year, then dealt with the perturbations coming with being fired, searching for a new job and starting it.

My plan was to get back to shaping my tracking file and posting updates as things got better, then I found and bought a house that was at the border of my affordability threshold and, after some debating, have decided to move all my 3a assets to my financing bank in order to “gain” leverage when I’ll come back to them to negociate a bigger mortgage to fund the renovations I am planning to do.

A big thanks to you for having been following me and my apologies for not coming through. May your investing ventures be fruitful and your life fulfilling.