Timing the market using moving averages - an experiment

Wolverine · October 13, 2021, 6:54pm

The fund used for Frankly’s Extreme 95 Index solution, topic of this thread, is getting down slowly and not anywhere near hitting a trigger (we’d need several big drops a few days appart for it to happen as of now).

I’ve given a shot at testing this strategy with my taxable account, with a fund tracking the swiss SPI. That one hit the exit trigger some days ago and hit the entry trigger today, resulting on a net loss for me.

My reasoning was:

“I need as much returns as I can as quickly as I can and should take more risks in order to get a chance to get them. The market is going up pretty strongly, so there are returns to catch if it keeps going that way.”
It’s a small amount, so not holding new investments for 6 months shouldn’t create a tax problem since way less than my regular income is at stake.
If it does end up creating a tax problem, I’ll at least get to have an example of what can make the tax office qualify you as a professional trader, which would have been somewhat fun in and of itself (I don’t take loosing money very seriously at this point, missing out on heavier gains is a much higher risk for me, right now, as I need alternatives to my current employment to preserve my mental health and getting a few 10’000 CHF on the side would give me enough of a cushion to take a shot at a self employed venture that I can’t launch without a decent amount of months of savings backing me up).
Better to experiment with this all with low amounts before hitting a net worth where I’d put aside a “recklessly agressive” part in my allocation, meant to be invested based on momentum strategies on cryptos and leveraged ETFs.

I’ve mentioned it in answer to @user137 writing that it would be exciting to see Murphy’s law materialize, which it did, though not with the specific experiment followed by this thread.

I’ll post graphs when the fund followed in this topic hits a trigger, so as to make clear what is happening. Sorry for the confusing side story.

Wolverine · October 30, 2021, 2:17pm

In our serie “learning while it’s cheap”, this month featured the lesson: “trying to predict the market is a vain endeavour” (a shocker for most here, I know ). I was thinking that the Evergrande troubles would have disrupted the markets globally, it seems whatever part it has played in the September drop was of “business as usual” magnitude and didn’t hit a trigger.

It’s nice to see that the “staying invested” bias has worked and kept me invested through the drop and the ongoing recovery. The SPICHA adventure in my posts above is a different story and a good reminder that false positives are to be expected on the way.

Triggers

As the fund started dropping, so did the 20 days SMA, giving some leeway for the following days even as down days were registered. It seems that a serie of bigger swings is required for the exit trigger to activate, which is fine with me: the longer the drawdown, the more chances there are that some bigger volatility days happen together in a cluster. I’m all in all happy with the behavior of the selected triggers so far.

Time-weighted returns

Real data

Mood: No reason to think we’re out of trouble yet but no reason not to think we’re going up moving forward either. Starting to digest what it means not knowing what the market will do in the future. Starting to give thought that, maybe, even following the market is a waste of time and the most efficient life strategy is to just drop money in and focus our attention on our life.

October 2021 returns (all strategies): 3.10%
October 2021 max intra-month drawdown (all strategies): -1.01%

All time compound monthly returns (all strategies): 1.7% (annualized: 22.07%)
!New! All time max drawdown (all strategies): -5.18% (September 8 to October 5)

YTD total returns (all strategies): 17.13%
YTD max drawdown (all strategies): -5.18% (September 8 to October 5)

The good news is that, at the end of November, I’ll have a full year of data (minus one day but I’m willing to cut that tiny corner) and we’ll be done with these silly monthly returns datapoints (I’ll do “per annum” for the most part).

Hush · November 2, 2021, 12:37am

Hello, thank you for sharing this experiment.
If you want to make more than an annual report, feel free to do some per quarter.

Wolverine · November 2, 2021, 5:56pm

Sorry, I’ve expressed myself poorly. Updates will remain monthly, the only thing that will change is the “all time compound returns”, which will be p.a. and not monthly anymore. For some reason, this appeared to me like “a lot of data”…

Enjoy the ride!

Grog · November 4, 2021, 9:50pm

This 1000%
Stopped looking at the market in 2015. Just drop money whenever you have it and keep an eye on assett allocation

Wolverine · November 30, 2021, 8:41pm

And one year in.

Triggers
Moving to a one year display length:

We’re getting close but we’ve been there before. Could go down, could bounce back up and rise some more.

Time-weighted returns

Real data

Mood: excited by the smell of opportunity.

November 2021 returns (all strategies): -0.74%
November 2021 max drawdown (all strategies): -3.95%

All time compound returns p.a. (all strategies): 19.2%
All time max drawdown (all strategies): -5.18% (September 8 to October 5 2021)

YTD total returns (all strategies): 17.87%
YTD max drawdown (all strategies): -5.18% (September 8 to October 5)

Wolverine · January 1, 2022, 3:17pm

Volatility is back, baby!

December has been a good month, and 2021 has been a very good year. Nothing hitting a trigger though there have been 2 closer calls, so not much to say for this strategy for the time being. I’m not complaining.

Frankly’s Extreme 95 Active solution is nearing the 200 days of history and I’ll be switching to it during January, mainly because it seems to behave “passively” enough (in that it does follow the main market movements) and for the 2 days delay to buy/sell shares vs 3 for the passive solution. I’ll create a post for that in about 2 weeks so don’t be surprised when it comes.

Triggers

Time-weighted returns

Real data

Mood: nonplussed, we’ll see what next year brings.

December 2021 returns (all strategies): 2.57%
December 2021 max drawdown (all strategies): -2.35% (9.12 to 21.12)

There has been an intermonths drawdown of -5.06% between November and December (17.11 to 06.12)

All time compound returns p.a. (all strategies): 20.4%
All time max drawdown (all strategies): -5.18% (September 8 to October 5 2021)

2021 returns (all strategies): 19.25%
2021 max drawdown (all strategies): -5.18% (September 8 to October 5)

Wolverine · January 14, 2022, 6:14pm

Alright, I have 200 days of data on the Extreme 95 active solution from Frankly. A comparison of the Active vs Passive versions of the Extreme 95 solution during that timeframe below:

As stated previously, the active version behaves “passively” enough, in that it follows the mains ups and downs of the passive version with similar shapes. My guess is that they’re using the same funds but weighting them based on their expectations. The outperformance is absolutely not guaranteed but it incurs less fees (no buying or selling fees, small though they are for the passive strategy) and, above all, allows for quicker buying and selling of shares (2 days vs 3).

I’ve initiated the change of strategy tonight and will be out of the market the coming days as the shares are sold, then bought.

Wolverine · January 24, 2022, 5:10pm

Well folks, this is it!

The exit curve of the 20D SMA -4 strategy has been crossed.

On the right of the black dashed line is when I’ve started tracking the Extreme 95 Active solution rather than the Index one. Moving averages have been adapted accordingly (the curves under the Extreme 95 Active curve derive from Extreme 95 Active data).

September felt like it was it but we rebounced. November was more of a meh in my head, it seems that going flat with volatility finally caught the more reactive strategy (20D SMA -4, as opposed to the slower to react 200D SMA -10).

We’ll see if stocks keep dropping or if we have a false positive with a rebound. I’m a bit anxious I’ll not manage the orders to get the funds into cash properly, we’ll see how that goes. For better or worse, we’ll now see a difference between one of the strategies and the benchmark.

Mood: pretty nonplussed, actually, I’m becoming somewhat numb to this.

Edit: Hey, buying and selling shares manually is actually very easy and intuitive. Kudos to those who designed that user interface and validation process!

Wolverine · January 31, 2022, 5:37pm

Well, if it keeps going up, I’ll have timed the bottom perfectly!

Triggers

Time-weighted returns

With a spawn of 1 month for better view of the different outcomes (as of now):

Real data

Mood: pretty amused. Watching things too closely makes you focus on some events over others, picturing yourself as the worst market timer. Clouds, stocks, same thing.

		Jan 2022		All times (p. a.)		YTD
Extreme 95 (Bench.)	Returns	-5.1%		13.6%		-5.1%
	Max DD	-6.9%	04.01 - 26.01	-7.4%	17.11.2021 - 26.01.2022	-6.9%
20D SMA -4	Returns	-6.4%		12.2%		-6.4%
	Max DD	-6.9%	04.01 - 31.01	-7.4%	17.11.2021 - 31.01.2022	-6.9%
200D SMA -10	Returns	-5.1%		13.6%		-5.1%
	Max DD	-6.9%	04.01 - 26.01	-7.4%	17.11.2021 - 26.01.2022	-6.9%

I’m switching to a table display for ease of use. Let me know if you like the text version more (or would like the table as a picture with negative numbers in red. It’d be pretty red, then. ^^)

dbu · January 31, 2022, 5:52pm

Try shortening text strings wherever you can, table gets hard to read on (my) mobile.
(e.g. January → Jan or 1; drawdown → DD etc.)

So IIUC your model triggered the sale (but not the rebuy yet)?

Wolverine · January 31, 2022, 8:06pm

Better? (I’ve edited it.)

Also, yes, sales triggered, no re-entry for the time being. The strategy is biased toward reinvesting quickly rather than staying on the sidelines, we’ll see if that was a good idea. ^^

If we’re going that way, I’d say the “problem” is diverging from simply taking the market’s returns and being done with it: you never know what you’re going to get. It could have protected my downside, and it still can, so there is that. I’ve been told the hard part was sticking with the strategy false positive after false positive, I’m willing to believe it.

It’s still a low amount of money, and it’s still probably not worth it vs keeping buying on the way down and up, at cheaper prices. The later seems also better for my peace of mind (not very worried but I’m probably giving this more focus than it’s worth).

Edit: for outside watchers (I’m sure @TeaCup knows this): this is expected behavior; volatility on the way up and down is normal, markets rarely go down or up in a straight line.

Wolverine · February 3, 2022, 7:17pm

And back in we go!

Frankly provides data only for the day before (so, today, it’s the price from Wednesday, February 2nd that is available), so I expect a down day tomorrow. It’s a bit disconcerting for me to act while expecting the downward trend to resume but I have to recall that one of the purposes of this design was to avoid any kind of need for predictions and/or thinking, so I’m biting the bullet, taking the loss and following the signals as they come.

That’s a 3.9% loss so far on the 20D-4 strategy when compared to buying and holding.

Also, some weird thing happened and there is now a slight difference in value for the benchmark strategy and the 200D-10. No idea what has happened but I’ve had display glitches with Frankly before so it may sort itself out tomorrow.

Mood: strangely enough, it’s kind of reassuring to see this kind of volatility while companies post earnings (Damn you, Alphabet! I would have gotten away with this too if it wasn’t for meddling children and a search engine!). I take it as meaning that price discovery is still active and we are not in a full on insanity market. I’m taking solace in that.

ItalianEngineer · February 18, 2022, 10:05pm

anything interesting is happening in the.movong averages in these days of 4Q results and (almost) war in Ukraine?

Wolverine · February 19, 2022, 2:47am

It’s actually interesting: we’re in slow bleed mode. It’s starting to look probable that we’ll cross the exit point of the slower to react strategy before the quicker one.

Another thing I’m starting to realize more thoroughly is that the quick to react strategy actually doesn’t offer reliable downside protection: it’s actually lower than buy and hold right now, and can trigger more false positives in the future, meaning that the market can actually go slightly up, and my strategy go down to zero. I had targeted something like a max drawdown of 8-15% when designing the quick reacting strategy, it was overly optimistic. We could easily go to a drawdown of 14% now, and it could get lower from there.

What this shows to me is that it’s one thing to plan and prepare for something and another one entirely to live through it, which is also part of why I’m conducting this experiment: to have better insights on my own psychology in regards to investing. It’s not taking a toll for the moment, I’m pretty much nonplussed, but I’ve also lost the apetite for random triggers just for the sake of it.

We’ll have to see what happens going forward but at this point, it seems like another way to have a risk/rewards profile potentially similar to buy and hold (meaning the downside protection and potential for more, or less, upside depends a good deal on luck and random patterns) with the added stress of having to look at the markets daily (or automate it but I wouldn’t feel good with automated trades). I’m starting to see the real power of buying and holding, which is the set and forget part of it. More energy for the meaningful things in life while still reaching financial success in due time.

Edit: putting an added thought here: I’m just back from a week in the Jura where we’ve done orienteering in the snow. One thing I’ve noticed is that I had a tendency to, at times, want a path to be the one to take and would then experience confirmation bia, seeing only the clues leading to that path being the right one, while neglecting those showing that “hey man, we’re not in a carebears world, yes, the path leading up the slope is the one leading to your target point”.

I had to fight it actively and ponder each time a decision had to be made whether it relied on all the information I had available or if I was using confimation bia to get the result I wanted. I think the same holds true for investing: it’s easy to get lured by ideas and see only the good points of them, then, when faced with reality, consider that reality is at fault for it. Keeping an open eye on the situation and staying honest with ourself requires active attention, market participants do what they do, and it’s very hard to predict. The markets can go up when all the signs we see point to down, it can seesaw for some time with participants apparently getting out and in and out and in and it is normal. Market participants will do what they want and that will make prediciting the market a vain endeavour. Expecting a certain set of results on the short term, even when backed by a lot of concurring indicators, is bound to get frustrating.

ItalianEngineer · February 19, 2022, 11:00pm

interesting situation from your graph. let’s see how it will evolve.

indeed, I guess that the major benefit of this experiment was for you the learning on how you truly are as a investor, and finally as a person.

if you continue with this experiment, please be sure that you at least a fan following you

Wolverine · February 28, 2022, 6:12pm

A quick reminder of my bias:

The 20D-4 strategy is meant to be quick to exit the market, and quick to re-enter. It is based on the assumption that “when in doubt, be invested”. It has triggered once so far and is currently underperforming the benchmark (buying and holding the same fund).

The 200D-10 strategy is meant to be slow to exit the market, but be decisive when it does. It hasn’t triggered so far, though for unknown reasons, my last investment in this strategy and the benchmark happened on a 1 day delay, so that there is a very slight difference in favour of the 200D-10 strategy as of yet (which is in no part due to the strategy itself).

That being said, things are starting to be interesting.

Triggers

We went real close to triggering an exit on the 20D-4 on Friday, but it didn’t by a hair and has now rebounced a bit. It could get triggered again any day now, or not. I’m not willing to take bets.

We still have a ways to go before triggering an exit on the 200D-10 strategy.

Time-weighted returns

The miss on the 20D-4 is visible.

Real data

The difference seems less impressive. I’m not too disapointed for now, this was bound to happen.

Mood: too much money sitting on the sidelines ready to get in or out on a whim. Can’t wait for central banks to actually tighten and reduce that amount, hoping it would lead to the market to make some kind of sense once again.

New display for the numerical data. Let me know if it’s better than the last. TW stands for time-weighted, CAGR is all times compound annual growth of returns, Max DD is max drawdown:

	20D-4	200D-10	Bench.
TW returns
CAGR	9.1%	11.0%	11.0%
February 22	-1.9%	-0.9%	-1.0%
YTD	-8.2%	-6.0%	-6.0%
Max DD
All times	-11.3%	-9.2%	-9.2%
February 22	-5.9%	-5.9%	-5.9%
YTD	-10.8%	-8.7%	-8.7%

The underperformance and failure as a downside protection in this specific case is apparent.

And that’s it for today folks! Have fun and don’t blow all your money on luck reliant strategies.

Wolverine · March 3, 2022, 6:43am

An additional thought: I find interesting to note that while the value of the shares of the fund have globally gone down, all 3 strategies are invested right now. My own sentiment was toward more downard pressure up until yesterday (and I’ve acted on it with other invested money I don’t want to loose). I am now considerably less convinced and think we may have hit a bottom, or not (in short, I don’t know, while I previously thought I might have an inkling, which might very well have been a mirage).

What I’d like to note is that even though my sentiment has changed, I am not willing to act on it (with my other funds, not in this experiment, which is set and will follow the rules I’ve stated at the start). I’m mainly resistant to change. I don’t mind too much being invested or not (for now, as the amounts considered are still low) but I do mind selling or buying with the lot.

Once out, getting back in is psychologically more difficult (for me), I imagine it gets worse the longer we stay out. Once in, getting out also becomes psychologically more difficult.

Wolverine · March 8, 2022, 4:56pm

Aaaaand… here we go again. We were over the bar by 3 cents yesterday, the exit trigger for the 20D-4 strategy has been crossed today.

Either we rebounce next, or we’ll cross the 200D-10 exit trigger too.

Mood: pretty nonplussed.

Wolverine · March 17, 2022, 6:14pm

Back in we go! It didn’t take long, I must say. It’s looking like another net loss for the 20D-4 strategy.

Mood: Confident in the future (my bet is the markets should move up and no trigger get hit in the coming weeks).