Leveraged SMA200 Strategy Back-tested 1929 - 2019

[RayKeynes]

Dear all,

I have made a backtest using the following data:

UPRO simulated 1929-2019 from HPC 3.82alpha
SSO simulated 1929-2019 from HPC 3.82alpha
SPY w/o dividends 1929-2019 from HPC 3.82alpha
TMF simulated 1986 - 2019 from HedgeFundie


Approach: Hold 100% UPRO when SPY w/o dividends trends above SMA200 (of SPY as benchmark). If trends below SMA200, sell all UPRO and hold 100% in Cash.

Results: I've loaded all data simulated in Portfolio Visualizer to apply the strategy. As Portfolio Visualizer is only able to load data from 1985, I've tricked the visualizer by saying that data from 1929 and from 1950 is also since 1985, but that is just so that I can run the analysis.

3'000 investing monthly
Hold and/or invest 100% in UPRO / Cash; depending on SMA200 of SPY
not inflation adjusted (as no inflation data available for 1929 - 1985 in Portfolio Visualizer)

Please see the following screenshots:

1985 - 2019


1950 - 1982 (tricked)


1929 - 1962 (tricked)


1985 - 2019 (using UPRO when SPY over SMA200, otherwise invest all in TMF)



Conclusion
I was very surprised by these results. Please be aware, that the benchmark used (SPY) is without dividends. So actual results when investing will differ as SPY generated higher dividends from 1929 - 1985 than todays 2-3%.

However, the results are quite overwhelming to me as I've used rather conservative approaches when simulating UPRO and SSO. I've also compared simulated SSO to actual 2x SP500 (UPLIX) (see below screenshot for check):

Simulated SSO vs UPLIX - 1985 - 2019


Verdict
Please critically analze the above results and destroy them. Otherwise I would not see any reason why not to go for at least 50% of my portfolio into that strategy (100% UPRO / 100% CASH).

[Forester]

One should bear in mind the fortunate pathway of trend following since the mid 1980s, stocks starting at low valuations then reaching nosebleed levels. The trend signal flips off in 2000, then there are bond market profits as the Fed cuts rates whilst the stock market tracks sideways. Re-enter around 2003, then a vicious V-shape bear market in 2008, providing nice clean exit & re-entry points for equity trend following. Then a bullmarket up to rich valuations in late 2017.

This picture is a dream scenario for any flavour of trend strategy. I must confess I am considering starting such a strategy next year, but without using bonds. There's every chance bonds will stink in the 2020s which means leveraged bonds will stink even more, on top of that if US megacap stocks go sideways for a few years in a choppy market UPRO could be underwhelming. I'm surprised no one is investigating ex-US leveraged strategies.

[AlohaJoe]

Can you please edit your post to use proper capitalization? Using all caps is not only again forum policies on this specific site but also against general etiquette everywhere else on the internet.

We also require that you be considerate of our readers and:

use meaningful titles on new topics
refrain from posting naked links - all links should include an explanation or excerpt unless its meaning is clear from the context
avoid profanities, obscenities, lewd and otherwise offensive words and remarks
avoid posting in ALL CAPITAL LETTERS or otherwise using distracting formatting

[RayKeynes]

One should bear in mind the fortunate pathway of trend following since the mid 1980s, stocks starting at low valuations then reaching nosebleed levels. The trend signal flips off in 2000, then there are bond market profits as the Fed cuts rates whilst the stock market tracks sideways. Re-enter around 2003, then a vicious V-shape bear market in 2008, providing nice clean exit & re-entry points for equity trend following. Then a bullmarket up to rich valuations in late 2017.

Sir, maybe you haven't fully read my first post, but I've indeed backtested the application of this strategy from 1929 - 2019 and not just since the 80s. The 200SMA strategy outperformed a buy & hold strategy in each singular scenario.

[HEDGEFUNDIE]

My 55/45 strategy had 18.4% MWRR between 1987-2019, only 2% less than your timing portfolio, with a lot less work.

[RayKeynes]

My 55/45 strategy had 18.4% MWRR between 1987-2019, only 2% less than your timing portfolio, with a lot less work.

Yes Sir, the only difference is that your portfolio consists of holding TMF and UPRO together without selling them and relying on a -1 correlation.

This worked great from 1980 - 2019 but not before that time period (refer to "Long term debt cycle" as indicated by Ray Dalio, please refer to Federal Funds Rate development - since 1980 the rate only declined). My strategy works for the whole 1929 - 2019 period. Check it out. Sometimes holding Cash is better than staying in the market at all.

To all: I would appreciate having data to simulate TMF before the 80s!!! Anyone out here can help me?

[RayKeynes]

My 55/45 strategy had 18.4% MWRR between 1987-2019, only 2% less than your timing portfolio, with a lot less work.

I do not care pressing the sell and buy bottom if it increases my CAGR by only 0.5%.
UPRO data used is very conservative as well, so the 2% difference would be even greater tough.

[Stef]

Could you please update your data so that dividends (and maybe even inflation) are included?

[MotoTrojan]

Remind me again why there isn’t already a fund that does this for you?

[Stef]

Remind me again why there isn’t already a fund that does this for you?

So you think this shouldn't work?

[RayKeynes]

Remind me again why there isn’t already a fund that does this for you?

Well we should create one? Just because there is no ETF applying this strategy does not mean it was not successful in the past.

Or is Portfolio Visualizer calculating wrongly?

[Forester]

One should bear in mind the fortunate pathway of trend following since the mid 1980s, stocks starting at low valuations then reaching nosebleed levels. The trend signal flips off in 2000, then there are bond market profits as the Fed cuts rates whilst the stock market tracks sideways. Re-enter around 2003, then a vicious V-shape bear market in 2008, providing nice clean exit & re-entry points for equity trend following. Then a bullmarket up to rich valuations in late 2017.

Sir, maybe you haven't fully read my first post, but I've indeed backtested the application of this strategy from 1929 - 2019 and not just since the 80s. The 200SMA strategy outperformed a buy & hold strategy in each singular scenario.

Compare the buy & hold returns of SPY vs ULPIX (2xS&P500)
https://www.portfoliovisualizer.com/tes ... odWeight=0

https://www.portfoliovisualizer.com/tes ... odWeight=0

It's possible that to "get a tune" out of leveraged ETFs, one needs a bond bullmarket or timing luck. If the opportunity set isn't there, levering up can't polish the t***. It's likely that had leveraged funds existed in 1950 they wouldn't have outperformed for three decades, no one has that much time or patience.

[RayKeynes]

One should bear in mind the fortunate pathway of trend following since the mid 1980s, stocks starting at low valuations then reaching nosebleed levels. The trend signal flips off in 2000, then there are bond market profits as the Fed cuts rates whilst the stock market tracks sideways. Re-enter around 2003, then a vicious V-shape bear market in 2008, providing nice clean exit & re-entry points for equity trend following. Then a bullmarket up to rich valuations in late 2017.

Sir, maybe you haven't fully read my first post, but I've indeed backtested the application of this strategy from 1929 - 2019 and not just since the 80s. The 200SMA strategy outperformed a buy & hold strategy in each singular scenario.

Compare the buy & hold returns of SPY vs ULPIX (2xS&P500)
https://www.portfoliovisualizer.com/tes ... odWeight=0

https://www.portfoliovisualizer.com/tes ... odWeight=0

It's possible that to "get a tune" out of leveraged ETFs, one needs a bond bullmarket or timing luck. If the opportunity set isn't there, levering up can't polish the t***. It's likely that had leveraged funds existed in 1950 they wouldn't have outperformed for three decades, no one has that much time or patience.

Wrong. 3'000 monthly contribution to the strategy. See below (with ULPIX)


Same with UPRO - even more wrong. 3'000 monthly contribution to the strategy. See below (with UPRO)

[RayKeynes]

Going back to 1990, the strategy even outperforms stronger than the standard SP500 investor:

[JoMoney]

FYI ... PortfolioVisualizer is not limited to starting at 1985
PV is limited by the data you use on it, so if you're using a data set that only goes back to 1985 that's your problem.
When I upload and use data going further back, I can use those dates.

[Stef]

FYI ... PortfolioVisualizer is not limited to starting at 1985
PV is limited by the data you use on it, so if you're using a data set that only goes back to 1985 that's your problem.
When I upload and use data going further back, I can use those dates.

Not the case for me. Can you please post a screenshot?

[JoMoney]

FYI ... PortfolioVisualizer is not limited to starting at 1985
PV is limited by the data you use on it, so if you're using a data set that only goes back to 1985 that's your problem.
When I upload and use data going further back, I can use those dates.

Not the case for me. Can you please post a screenshot?

Uh.. Ok

[Hydromod]

I found that when backtests with UPRO and TMF align rebalances with the end of month or the end of a quarter, the results tend be anomalously better than when backtests align rebalances with other dates or other durations (e.g., weekly, biweekly). The quarterly backtesting performed to decide on Hedgefundie's strategy showed the anomalous bump for quarterly rebalance intervals when rebalances are near the beginning/end of the month.

Portfolio visualizer aligns with the end of the month by design.

You can see a dependency on timing frequency in

https://www.portfoliovisualizer.com/tes ... odWeight=0

by setting the timing period to weekly or setting the timing to "on signal". In this test, monthly and bimonthly timing do much better than weekly or quarterly.

Unfortunately portfolio visualizer doesn't give the option for creating a tell-tale chart. If you see a few discrete jumps causing the outperformance, it was likely due to timing luck; if steady gains over time, it may be exploitable.

You might want to backtest with modified datasets that are based on subtracting one, two, or three weeks from the same sequence, or several days, to get a feel for how much might be timing luck.

So one might wonder whether (i) there will be exploitable behavior associated with the end of month going forward (in which case you should always make the investing decision at the end of the month) or (ii) any exploitable behavior will be arbitraged out.

There is some evidence that dividends tend to be clustered, which in turn affects prices; perhaps that provides a systematic effect.

[Uncorrelated]

According to my calculations 100% S&P has an expected CAGR of around 5% + risk free rate and a 50:50 mix of UPRO has a CAGR of around 5.6%+ risk free rate. Simply put, even if your timing strategy switched between UPRO and S&P at random, you would still expect higher performance than 100% S&P.

The performance differences you are seeing likely has nothing to do with your timing strategy but everything with taking more risk. This is clearly visible in the screenshot because you outperform the S&P 500 while taking twice the amount of risk. Higher risk results in higher reward (as long as you don't pass the kelly criterion), who would have thought?

[Stef]

Uh.. Ok

How do you do that? When uploading data I get the message, that it's limited to 40 years.

Which tool are you exactly using? Not backtest portfolio?

[RayKeynes]

According to my calculations 100% S&P has an expected CAGR of around 5% + risk free rate and a 50:50 mix of UPRO has a CAGR of around 5.6%+ risk free rate. Simply put, even if your timing strategy switched between UPRO and S&P at random, you would still expect higher performance than 100% S&P.

The performance differences you are seeing likely has nothing to do with your timing strategy but everything with taking more risk. This is clearly visible in the screenshot because you outperform the S&P 500 while taking twice the amount of risk. Higher risk results in higher reward (as long as you don't pass the kelly criterion), who would have thought?

I understand your view. But when analyzing the results in the past one can see that taking more risk to achieve higher returns was very favourable in a lot of scenarios. There were only a few scenarios (especially 1960 - 1985), when the timing portfolio did not outperform the market portfolio.

On the other hand, one is not invested in the market during crazy market crashes, which is also a nice thing. But as I've said, I'd never invest 100% of my portfolio in such a strategy.

[rascott]

I've been running such a strategy using TQQQ for a handful of months now. Your backtests are fairly similar to what I came up with. Trend following systems "work" in that they obviously keep you out of really bad bear markets. The tradeoff for this insurance has been you underperform pure buy and hold. However using leverage seems to flip this script (or at least it has in the past).

The biggest risk I see in this system is a choppy flat market, where you are getting whipsawed around. But every system has drawbacks (including plain old buy and hold)

I don't find it a very cumbersome strategy..... one could go months/years without making a single trade. It's pretty simple to watch/ check MAs.

There are a variety of offshoots you could use, as to where you set your trigger points, what you rotate into (bonds, cash, even just non- levered equities).

I wouldn't bet the house here..... but a 10% slice could turn into something. The TQQQ I'm holding is up roughly 40% in the last 4 months or so. Good luck.

[Stef]

I've been running such a strategy using TQQQ for a handful of months now. Your backtests are fairly similar to what I came up with. Trend following systems "work" in that they obviously keep you out of really bad bear markets. The tradeoff for this insurance has been you underperform pure buy and hold. However using leverage seems to flip this script (or at least it has in the past).

That's a very interesting observation.

Are you using SMA200?

[RayKeynes]

1985 - 2019



1950 - 1983

A timed strategy using SMA200 and a portfolio consisting of 50% VFINX and 50% UPRO delivered by far superior annual returns and at the same time did not result in big drawdanws while always staying invested in the market 100% of the time.

If SMA200 is crossed, one does simply sell the 50% UPRO and buy VFINX instead.

[Forester]

1985 - 2019



1950 - 1983

A timed strategy using SMA200 and a portfolio consisting of 50% VFINX and 50% UPRO delivered by far superior annual returns and at the same time did not result in big drawdanws while always staying invested in the market 100% of the time.

If SMA200 is crossed, one does simply sell the 50% UPRO and buy VFINX instead.

I wonder if USMV could work even better than VFINX. Min vol tends to hold up higher when there's market stress.

[RayKeynes]

Backtest of 67 periods à 25 years

SMA 225 Strategy (green and light green) - only difference is whether you hold 100% Cash oder 100% SP500 when Price below SMA225
red = UPRO (100%)
yellow = SP500 Total Return (100%)


Can someone please backtest that as well?

[RayKeynes]

Done a backtest with Portfolio Visualizer implementing the following strategies for the period 1929 - 2019:

- SMA252 with 50% UPRO and 50% SP500 Total Return - in case of signal - invest all in Cash
- SMA252 with 50% UPRO and 50% SP500 Total Return - in case of signal - invest all in SP500 Total Return
- UPRO 100% (as comparison)
- SP500 Total Return Index (as Benchmark)

Here are the results:


The SMA252 portfolio consisting of 50% UPRO and 50% SP500 total return and in case of signal - invest in Cash - has outperformed pure SP500 Total Return investment in all of the 67 periods tested. Not in a single period, this strategy has been weaker.

However, the additional return needs to be "paid" with slightly higher volatility as can be seen in the table.

[Forester]

In other words, buy-and-hold minimum volatility index will match leveraged S&P 500-switching into cash, but with half the annual volatility and 10% of the headache.

[RayKeynes]

In other words, buy-and-hold minimum volatility index will match leveraged S&P 500-switching into cash, but with half the annual volatility and 10% of the headache.

I do not get it, what do you mean? Did you read the numbers?

[garlandwhizzer]

I am a bit skeptical of assuming that backtesting results, particularly when using indexes in which trading is free and frictionless instead of real funds with frictions, costs, and management expenses. The fact that dividends were not included in the analysis adds to my basic skepticism about the reliability of the performance results going forward. Also the tax implications versus buy and hold indexes are important if these investments are not in tax sheltered account. This strategy purports to reliably produce significant risk adjusted outperformance relative to buy and hold in essentially all market cycles. This is the holy grail of investing and I can't help but wonder why with so many brilliant and knowledgeable full time professionals studying the market every day, looking for exactly that, no one has until now discovered it.

I looked at the prospectus of UPRO and found the following quote in the first line under Principal Risks.

Principal Risks
An investor in the Fund could potentially lose the full principal value of his/her investment within a single day.

Apparently the management of UPRO believes there is very serious if extremely rare potential risk which has not yet shown up on backtesting as far as I can see. I agree that this is an interesting strategy and that the backtesting results are very strong but with the limitations above. Portfolio turnover is 15%, expense ratio is .92%, and that does not necessarily include all trading and leverage costs. Regardless of backtesting results, I do not have unquestioning faith that future returns of any investment asset or plan will reliably be a replay of the past. The strategy uses 3X leverage and makes market timing decisions on a regular basis. Leverage and market timing can magnify results if properly done (which can be aligned perfectly on backtesting) but in general both err more often than they succeed whether decisions are made by formula of emotions. For those who choose to follow this interesting strategy, I wish them luck as I do for all investors. I will not however be one of them. I've been investing regularly for 30 years and during that time I have seen a lot of investing secret sauce that looked fool-proof and compelling when presented go wrong over time. It's hard for me to get over that.

Garland Whizzer

[Stef]

The interesting thing is that SMA200/250 doesn't work with regular investments. You are selling and buying for the same price, sometimes buying slightly higher. You are basically avoiding big drawdowns, but getting the same or slightly worse return. And that seems to be crucial when it comes to leveraged ETFs. The drawdowns are the killer. Avoiding them might boost the performance significantly, like demonstrated in these backtests. Just one example with SMA200/250:



With no leverage you would slightly decrease your return when selling/buying where the lines are crossing. But with leverage you avoid a lot of volatility decay.

[RayKeynes]

The fact that dividends were not included in the analysis adds to my basic skepticism about the reliability of the performance results going forward. Also the tax implications versus buy and hold indexes are important if these investments are not in tax sheltered account. T

My model does include dividends as the benchmark is the SP500 total return index. Furthermore, UPRO has been estimated back to 1929 rather conservatively and its borrowing costs.

Regardless of backtesting results, I do not have unquestioning faith that future returns of any investment asset or plan will reliably be a replay of the past. The strategy uses 3X leverage and makes market timing decisions on a regular basis.

I totally agree with you that the past returns do not predict how the future will turn out. However, the strategy even outperformed in the 30s when the SPY was hit extremely and in all different scenarios as well.

I agree with you that these are the two biggest risks with this strategy:

1. A big 1-day stock fall above SMA252 (while in "bull run") - never happened in history. October 19 1987 happened below SMA252, so you'd be protected from that one.

2. A fall of more than 33% on a single day would wipe out your investment value completely in this strategy. However, if this happends within the first 5 years of your investment horizon, you might still start the strategy completeley new...

3. Markets that run flat for the next 25 years. That would - however, also kill a strategy investing in SP500.

[RayKeynes]

Please see above the whole backtesting now finished for all 25-year periods
Key comments to methodology: no inflation adjustments, 3'500$ monthly investment / 10'000$ initial, calculated values using PortfolioVisualizer

I've compared invevsting in SP500 total return with a strategy investing 50% in SP500 total reutrn and the other 50% in UPRO with the simple moving average 252 as an indicator whether to invest or stay in cash.

Please see the results above.
On median, the strategies resulted in the following end portfolio values:
SP500 TR = 6.5 Mio
SMA252 / Cash (50/50) = 12.3 Mio
SMA 252 / SP500 TR (50/50) = 9.3 Mio

[RayKeynes]

up for new discussion

[RayKeynes]

Some new updates for the ongoing bear market and therefore also decline in UPRO valuation.

I've analyzed all bear markets in the past 100 years and calculated via backtesting the lowest UPRO value taking into account the highest UPRO valuation of around 80$ as of early 2020.

Here are the results:

Great Depression 1929- Lowest Value of UPRO: 0.05$

Crysis 1968 - Lowest Value of UPRO: 17.86$

Crysis 1973 - Lowest Value of UPRO: 7.60$

Crysis 1980- Lowest Value of UPRO: 20.42$

Crysis 1987- Lowest Value of UPRO: 16.22$

Crysis 2001- Lowest Value of UPRO: 8.51$

Crysis 2008- Lowest Value of UPRO: 3.89$


Discussion
Most of the lowest UPRO value in past crysis ranged from 4$ (2008/2009) to 17$.

It is hard to predict the bottom value of the leveraged ETF UPRO. However. At the current valuation of above 20$ - one can definitely say that there is still room for UPRO to decay and loose more value.

[RayKeynes]

[Thread merged into here, see below. --admin LadyGeek]

Dear all,

I was thinking for a while how to savely profit via Leverage in the financial markets. Thus, I have backtested various strategies going back to 1928.

My idea
I have tried multiple strategies, mostly using Leveraged ETFs as it represnets a relatively comfortable financial instrument as you do not have to "actively manage" it like options or future contracts or other derivatives.

I have backtested the results of UPRO (3x leverage on SP500) and SSO (2x leverage on SP500). While the results were quite good - I still saw that the risk of sharp drawdowns is omni-present. The higher the volatility (measured in VIX), the more your leveraged portfolio will suffer from volatility decay. As an increase in volatility is generally happening in Bear markets (e.g. in market downturns), we can turn this knowlege into a strategy.

As a result, I have tried to implement additional safety measures so that the leverage strategy does not suffer from large drawdowns (e.g. 1929 - 1940; 1973; 1982; 2001; 2008 and now 2020).

Strategic Approach
The strategy, which I've backtested using 92 years of financial data, works as follows:

a) Monitor SPY as a benchmark Index
b) Monitor the 235-day moving average of SPY for your buy and sell signal
c) Whenever SPY closes above the 235-day SMA, buy UPRO, if its below, sell UPRO and hold Cash.

One might think that this strategy is so simple and thus cannot work. However, I have fully backtested the above-mentioned portfolio and was very impressed with the results (refer to attached excel: https://gofile.io/?c=jNLR6v).

These are the rolling returns for the portfolio (non-inflation adjusted, no contribution); simply the growth of 10'000$ of capital:

Median Return for 25 years: 14.73% (End portfolio value: 310'705$ (not inflation adjusted)
Average Return for 25 years: 16.54% (End Portfolio value: 458'870$ (not inflation adjsuted)
Max Return for 25 years: 23.08%
Min Return for 25 years: 2.71%

I must add that the minimum return occured in the period from 1929 to 1954 (Global Depression) and one only had to wait 2 additional years (until 1956) until the return was at least 9.8%.

Risk associated with this Strategy
It is clear that this strategy will lead to a higher standard deviation and potentially smaller (Yes, smaller) drawdowns as you hold Cash during Crysis.

The highest risk is a large drawdown above the Simple Moving Average. For example, a single-day drawdown of the SP500 of 7% will lead to your portfolio crashing 20% (if it happens while you're not staying in cash, thus above the SMA).

This has happily not happened in the last 100 years as all large single-day crashes happened below the Simple Day Moving Average.

It is crucial that you understand the risks associated with this strategy.

One does not have to 100%-rely on this strategy but it is a very nice way to boost your average portfolio returns of an investment period of more than 20 years. While the average SP500 portfolio will generally lead to an average return of 10.0%, this strategy has a return of 14.8%.

I recommend not investing more than 30% of your portfolio in such a strategy.

Grahpic Explanation of the strategy
The below screenshots show how this strategy works in real-world financial markt conditions:

[tvubpwcisla]

Don't overthink your investment plan. Keep things simple. The biggest risk you have is between you and the keyboard. I have not seen one investment plan call for leveraged funds.

[RayKeynes]

Don't overthink your investment plan. Keep things simple. The biggest risk you have is between you and the keyboard. I have not seen one investment plan call for leveraged funds.

Just because you have never seen one - does not mean it doesn't work. I agree thus, there is no higher market return without "increased risk". However, risk is manageable.

[CardioMD]

Usually when I hear “strategy” I tune out. However, I decided I’d give your post a read. You did quite a bit of work and I’m sure the numbers look great in hindsight. I do think you’re trying to squeeze a few more pennies out of a dollar and the risk, to me, isn’t worth the work and the timing involved to make it work.

I prefer to buy American business and keep a proper asset allocation, but that’s just me. Best of luck and keep us updated.

[rascott]

I've been using this since last summer... but use TQQQ.

Isn't really a new idea... there have been some papers written on this, as well as a variety of articles you can online. All have shown the same basic conclusions.... volatility clusters..... and it occurs when markets are under their 200 day averages.

[rascott]

Usually when I hear “strategy” I tune out. However, I decided I’d give your post a read. You did quite a bit of work and I’m sure the numbers look great in hindsight. I do think you’re trying to squeeze a few more pennies out of a dollar and the risk, to me, isn’t worth the work and the timing involved to make it work.

I prefer to buy American business and keep a proper asset allocation, but that’s just me. Best of luck and keep us updated.

A few more pennies? No, not even close. 5% CAGR over long periods is exponentially more money.

[ARoseByAnyOtherName]

Don't overthink your investment plan. Keep things simple. The biggest risk you have is between you and the keyboard. I have not seen one investment plan call for leveraged funds.

Just because you have never seen one - does not mean it doesn't work. I agree thus, there is no higher market return without "increased risk". However, risk is manageable.

You’re obviously convinced your strategy will woke and aren’t interested in any feedback. So, just do it.

But come back here periodically and post your performance compared to the SPY benchmark. The proof is in the pudding after all.

Also you are far from the first person to come up with something like this. Check out the Excellemt Adventure threads for the latest and most prominent recent examples on this board.

[amp]

Why did you choose to use the 235 day moving average? Is it just the one that backtests the best? And how does it affect your returns if you choose a different MVA--200, 250, 300, etc?

[midareff]

For some reason I keep hearing that old failed hedge Fund tune about risking money you do have and do need to make more money you probably won't need. I'm more like CardioMD, and very content owning US businesses and their debt instruments.

[nisiprius]

1) Please say that a strategy "would have" returned unless you can show that you or someone else has actually followed the strategy. Even with Massachusetts Investors' Trust I think it's better to say "would have" because it's not clear that anyone has actually held shares of that fund since inception, but at least it is a real fund that used real money and we know what it really returned: 9.60% CAGR from 7/15/1924 through 4/24/2020..

2) The more leverage you use, the more dependent your results will be on the occurrence of rare events. The rarer an event is, the less reliable a backtest is as a prediction of the future.

The more leverage you use, the more you are relying on your personal financial strength in a crisis... and the less likely it is that you will actually be able to follow the strategy.

3) Results based on products that did not exist are suspect, because for whatever reason there is a history of performance decline when previously uninvestable assets are suddenly made available to an enthusiastic public in a new, investable form. A good example of this would be "commodities" (collateralized commodity futures). Circa the early 2000's backtested results looked great, but after they became available in mutual fund and ETF form and became accepted as legitimate assets for ordinary retirement savers, performance declined and then turned negative. It is widely believed that there was a change in behavior and that it was caused by "financialization."

[nisiprius]

...However, risk is manageable...

Everybody who takes risk believes it is manageable; otherwise they would not take it.

[RayKeynes]

Why did you choose to use the 235 day moving average? Is it just the one that backtests the best? And how does it affect your returns if you choose a different MVA--200, 250, 300, etc?

I have back tested using 100-400 Day moving averages.

The best results were achieved using SMA180-SMA270

SMA225 did result in the best returns over the last 100 years, however, it did also significantly increase the times you had to buy and sell. The larger the SMA used, the fewer you have to sell or buy. SMA235 only had a slightly minor return than SMA225 but with significantly fewer transactions thus making it more comfortable to ride.

Depending on the 25 year period under investigation, a 180SMA might outperform a 270SMA and Vice versa. However, the important note is that Leveraged SMA220 to SMA250 have always outperformed the Market in every single 25 year period.

[WoodSpinner]

OP,

You might find this thread helpful....

viewtopic.php?t=270035

WoodSpinner

[amp]

Why did you choose to use the 235 day moving average? Is it just the one that backtests the best? And how does it affect your returns if you choose a different MVA--200, 250, 300, etc?

I have back tested using 100-400 Day moving averages.

The best results were achieved using SMA180-SMA270

SMA225 did result in the best returns over the last 100 years, however, it did also significantly increase the times you had to buy and sell. The larger the SMA used, the fewer you have to sell or buy. SMA235 only had a slightly minor return than SMA225 but with significantly fewer transactions thus making it more comfortable to ride.

Depending on the 25 year period under investigation, a 180SMA might outperform a 270SMA and Vice versa. However, the important note is that Leveraged SMA220 to SMA250 have always outperformed the Market in every single 25 year period.

How many more transactions were there because the MA changed from 225 to 235? If it really was a lot, it speaks to a certain brittleness if a 10-day change can have that much of an effect.

Anyways, if you're really going to follow through on this strategy I do hope you will document your returns and market movements somewhere, whether it's in this thread or on your own blog. I do wish you luck, but it's been my experience that few mechanistic trend following strategies last very long in the real world. Even the big names, like Meb Faber, seem to regularly tinker with their approaches when they don't perform as well as expected.

[JoMoney]

You already had a thread on this a few months ago
viewtopic.php?f=10&t=297591