HEDGEFUNDIE's excellent adventure Part II: The next journey

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
cos
Posts: 235
Joined: Fri Aug 23, 2019 7:34 pm
Location: Boston
Contact:

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by cos »

randyharris wrote: Mon Jul 13, 2020 11:05 pm
glenmalan wrote: Mon Jul 13, 2020 9:34 pmFor the Adaptive model, how many timing periods are you using? I use 1.
Volatility Trading Days? I use 21
Allocation weight method? I use inverse volatility
Trading Frequency? I use monthly
This is the setup, for me, with either TQQQ/TMF, or QLD/UBT.

https://www.portfoliovisualizer.com/te ... odWeight=0
How did you arrive at these parameters? Why a single performance period? Why is the performance period 1 month? Why is the volatility period 21 days?

Did you use something along the lines of the following logic to choose minimum variance over inverse volatility?
Image
(see: viewtopic.php?t=277661)
User avatar
cos
Posts: 235
Joined: Fri Aug 23, 2019 7:34 pm
Location: Boston
Contact:

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by cos »

markcloutier1212 wrote: Mon Jul 13, 2020 11:05 am With hundreds of pages its kind of hard to tell -- but is there any new consensus on the ideal rebalancing schedule for Hedgefundie's 45/55 TMF/UPRO approach? Is it 10% bands, quarterly, or something else?
The "consensus" seems to be quarterly if you're following the vanilla adventure. Alternatives have been explored in depth by Hydromod over here: viewtopic.php?f=10&t=284955

Reading through that thread should answer most of your questions.
BogleBobby
Posts: 38
Joined: Tue Jul 30, 2019 9:56 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by BogleBobby »

I'm going to throw some money in on the adaptive allocation approaches. Unless my math is wrong, It appears that the adaptive allocation approach outperforms a fixed allocation approach fairly consistently from 1987 to now (using the daily simulated dataset). If you break up the last ~30 years of data into five-year periods, an adaptive approach performs better 5 out of 6 of the five-year periods, so it doesn't appear to be just a recent fluke.

I'm a little more unsure about the UPRO vs TQQQ split. I have more confidence in future growth of Nasdaq companies, but it also seems like a bad time to get into QQQ due to the recent run-up. Also, I'm not huge on overweighting a segment of the market. But, this is all 'fun money' in my accounts that i can afford to lose, so why not take a risk. I think I'm going to allocate 50% of my money to UPRO and 50% to TQQQ to hedge my bets.
snowsnow
Posts: 1
Joined: Thu Jul 16, 2020 9:20 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by snowsnow »

Has there been any discussion here about the 3x consumer staples ETF $NEED?

Its less volatile than S&P 500 giving you less drag, but does it just suffer too much from the lack of diversity?
Mickelous
Posts: 134
Joined: Mon Apr 20, 2020 11:24 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Mickelous »

Did my first rebalancing today, and after my first quarter I am up 15%. Was as high as 20+ at one point but that's part of the ride. Still very impressed as that's greater than a whole year for most folk. Rebalanced to 45TMF/30UPRO/25TQQQ. Would have been 20% had I not messed around with leveraged gold, real estate, and utilities for a week or two before cutting my losses.
User avatar
danyboy7
Posts: 198
Joined: Tue Jul 14, 2020 11:35 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by danyboy7 »

Mickelous wrote: Thu Jul 16, 2020 10:05 am Did my first rebalancing today, and after my first quarter I am up 15%. Was as high as 20+ at one point but that's part of the ride. Still very impressed as that's greater than a whole year for most folk. Rebalanced to 45TMF/30UPRO/25TQQQ. Would have been 20% had I not messed around with leveraged gold, real estate, and utilities for a week or two before cutting my losses.
What do you mean by by messing with leveraged gold,utilities and reit ?
I have seen the light
Mickelous
Posts: 134
Joined: Mon Apr 20, 2020 11:24 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Mickelous »

danyboy7 wrote: Thu Jul 16, 2020 10:49 am
Mickelous wrote: Thu Jul 16, 2020 10:05 am Did my first rebalancing today, and after my first quarter I am up 15%. Was as high as 20+ at one point but that's part of the ride. Still very impressed as that's greater than a whole year for most folk. Rebalanced to 45TMF/30UPRO/25TQQQ. Would have been 20% had I not messed around with leveraged gold, real estate, and utilities for a week or two before cutting my losses.
What do you mean by by messing with leveraged gold,utilities and reit ?
I added them to my portfolio then changed my mind. Had all three at a combined 20% of the adventure at one point.
User avatar
danyboy7
Posts: 198
Joined: Tue Jul 14, 2020 11:35 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by danyboy7 »

Mickelous wrote: Thu Jul 16, 2020 11:55 am
danyboy7 wrote: Thu Jul 16, 2020 10:49 am
Mickelous wrote: Thu Jul 16, 2020 10:05 am Did my first rebalancing today, and after my first quarter I am up 15%. Was as high as 20+ at one point but that's part of the ride. Still very impressed as that's greater than a whole year for most folk. Rebalanced to 45TMF/30UPRO/25TQQQ. Would have been 20% had I not messed around with leveraged gold, real estate, and utilities for a week or two before cutting my losses.
What do you mean by by messing with leveraged gold,utilities and reit ?
I added them to my portfolio then changed my mind. Had all three at a combined 20% of the adventure at one point.
What do you think about leveraged x3 all weather with utilities instead of commodities ? https://www.portfoliovisualizer.com/bac ... sisResults
And why choosing nasdaq x3 leveraged,ins't it perhaps too risky ?
I have seen the light
Mickelous
Posts: 134
Joined: Mon Apr 20, 2020 11:24 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Mickelous »

danyboy7 wrote: Thu Jul 16, 2020 12:07 pm
Mickelous wrote: Thu Jul 16, 2020 11:55 am
danyboy7 wrote: Thu Jul 16, 2020 10:49 am
Mickelous wrote: Thu Jul 16, 2020 10:05 am Did my first rebalancing today, and after my first quarter I am up 15%. Was as high as 20+ at one point but that's part of the ride. Still very impressed as that's greater than a whole year for most folk. Rebalanced to 45TMF/30UPRO/25TQQQ. Would have been 20% had I not messed around with leveraged gold, real estate, and utilities for a week or two before cutting my losses.
What do you mean by by messing with leveraged gold,utilities and reit ?
I added them to my portfolio then changed my mind. Had all three at a combined 20% of the adventure at one point.
What do you think about leveraged x3 all weather with utilities instead of commodities ? https://www.portfoliovisualizer.com/bac ... sisResults
And why choosing nasdaq x3 leveraged,ins't it perhaps too risky ?
I personally like a majority of the companies in the nasdaq currently. I think you get the most overall movement from s&p/nasdaq/treasuries. There isn't enough movement from utilities and gold in the LETFs from what I was looking at to maximize gains.
Micronite
Posts: 6
Joined: Tue Apr 07, 2020 1:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Micronite »

Mickelous wrote: Thu Jul 16, 2020 10:05 am Did my first rebalancing today, and after my first quarter I am up 15%. Was as high as 20+ at one point but that's part of the ride. Still very impressed as that's greater than a whole year for most folk. Rebalanced to 45TMF/30UPRO/25TQQQ. Would have been 20% had I not messed around with leveraged gold, real estate, and utilities for a week or two before cutting my losses.
How much of your overall portfolio are you dedicating to this strategy? And are you doing it taxable or IRA?
typical.investor
Posts: 2326
Joined: Mon Jun 11, 2018 3:17 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by typical.investor »

Dr. Long wrote: Sun Jul 12, 2020 10:52 pm
Micronite wrote: Sun Jul 12, 2020 2:04 am Is it a good idea to use the Excellent Adventure outside of a tax-advantaged account?
No
I am currently holding some UPRO outside tax sheltered. I’ll to keep rebalancing within tax sheltered though where I hold both TMF and UPRO.
Mickelous
Posts: 134
Joined: Mon Apr 20, 2020 11:24 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Mickelous »

Micronite wrote: Fri Jul 17, 2020 1:48 am
Mickelous wrote: Thu Jul 16, 2020 10:05 am Did my first rebalancing today, and after my first quarter I am up 15%. Was as high as 20+ at one point but that's part of the ride. Still very impressed as that's greater than a whole year for most folk. Rebalanced to 45TMF/30UPRO/25TQQQ. Would have been 20% had I not messed around with leveraged gold, real estate, and utilities for a week or two before cutting my losses.
How much of your overall portfolio are you dedicating to this strategy? And are you doing it taxable or IRA?
About 90 percent of my 401k and 100 percent of my Roth IRA are in. What can I say I believe it works. My 401k would be 100 percent but it's a plan limitation. I'm still in the growth and accumulation phase I'm my investing lifecycle. 20 percent year over year actually gets me excited to save.
sf1988
Posts: 5
Joined: Tue Jul 14, 2020 1:41 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by sf1988 »

Micronite wrote: Fri Jul 17, 2020 1:48 am
Mickelous wrote: Thu Jul 16, 2020 10:05 am

How much of your overall portfolio are you dedicating to this strategy? And are you doing it taxable or IRA?
About 90 percent of my 401k and 100 percent of my Roth IRA are in. What can I say I believe it works. My 401k would be 100 percent but it's a plan limitation. I'm still in the growth and accumulation phase I'm my investing lifecycle. 20 percent year over year actually gets me excited to save.
Mind sharing your age and total balance?
User avatar
danyboy7
Posts: 198
Joined: Tue Jul 14, 2020 11:35 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by danyboy7 »

Hi Bogleheads,I would mind to share with you the leveraged AllWeather x3 version using utilities from the optimizing blog,which fascinating me a lot
Unfortunately backtesting the x3 AW leveraged x3 with utilities version and its risk parity version goes until 1992,not furtherback,even using negative CashX on portfoliovisualizer:
https://imgur.com/73K2VyG
Here the asset allocation for the 2 portfolios:
AW x3 with utilities AW x3 with utilites true risk-parity version
TMF – 40% 14%
UPRO – 30% 14%
TYD – 15% 39%
UTSL – 7.5% 12%
UGL – 7.5% 21%

Which would be the oldest dataset I could get for the AW x3 ? It would be awesome if you would manage to help me to get the oldest backtest possible. I have also wrote to siamond hoping he would help me (I've seen his awesome excel sheet of all assets backtest)
I have seen the light
Mickelous
Posts: 134
Joined: Mon Apr 20, 2020 11:24 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Mickelous »

sf1988 wrote: Fri Jul 17, 2020 11:15 am
Micronite wrote: Fri Jul 17, 2020 1:48 am
Mickelous wrote: Thu Jul 16, 2020 10:05 am

How much of your overall portfolio are you dedicating to this strategy? And are you doing it taxable or IRA?
About 90 percent of my 401k and 100 percent of my Roth IRA are in. What can I say I believe it works. My 401k would be 100 percent but it's a plan limitation. I'm still in the growth and accumulation phase I'm my investing lifecycle. 20 percent year over year actually gets me excited to save.
Mind sharing your age and total balance?
Age:31
Balance a bit over 20k. Currently playing a bit of catch up have about half a years salary (not expenses) invested but saving 29 percent in 401k including company match and putting in my Roth when I can.
locallyoptimal
Posts: 6
Joined: Mon Oct 28, 2019 2:43 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by locallyoptimal »

Hi friends,

Would anyone please point me to prior discussions regarding the "beta slippage" risks of the leveraged ETF strategy?

Here's what I mean: today UPRO stands at ~50, with SPY at 321.7
On 12/23/19, SPY closed at approximately the same level, 321.2 and UPRO closed at 69.8

i.e. SPY has recovered to ~December levels whereas UPRO is lagging by ~30%.

I see that HF and others refer to this as "decay", quoting the original post:
$6.80 is not 3X $2.76. That's the "decay", and it's nothing more than simple math. This dynamic can also work in your favor: if you check the performance of UPRO against the S&P 500 since inception, you will see that UPRO has delivered 5x the returns of the index. If an index tends to go up over time (i.e. exhibits positive momentum), a 3x leveraged ETF will tend to perform better than 3x the index over the long term.
but I'd like to read more recent discussions on the topic if any, particularly relating to the current macro-environment.

I've been using this interesting strategy for about a year and was heavily down in March, but averaged down on UPRO to ~27 during the bottom, with a bit of TQQQ purchased near the bottom (though just around 10%). I've been slowly shifting into TMF but I'm considering terminating my leveraged positions for now, as we seem to be in bubble territory propped up by the Fed; It's not clear to me that with interest rates so low, treasuries are in the same category that they've been in decades.
It seems to me that reality will catch up in the medium term, and hedging appears expensive. But of course, the market can remain weird longer than anyone can stay solvent :-)

That too I'm sure has been discussed--I'm reading through all of your interesting posts, but it would be really nice to have an index feature!

Would welcome any insights or corrections--thank you!
rockstar
Posts: 896
Joined: Mon Feb 03, 2020 6:51 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by rockstar »

locallyoptimal wrote: Sat Jul 18, 2020 2:50 pm Hi friends,

Would anyone please point me to prior discussions regarding the "beta slippage" risks of the leveraged ETF strategy?

Here's what I mean: today UPRO stands at ~50, with SPY at 321.7
On 12/23/19, SPY closed at approximately the same level, 321.2 and UPRO closed at 69.8

i.e. SPY has recovered to ~December levels whereas UPRO is lagging by ~30%.

I see that HF and others refer to this as "decay", quoting the original post:
$6.80 is not 3X $2.76. That's the "decay", and it's nothing more than simple math. This dynamic can also work in your favor: if you check the performance of UPRO against the S&P 500 since inception, you will see that UPRO has delivered 5x the returns of the index. If an index tends to go up over time (i.e. exhibits positive momentum), a 3x leveraged ETF will tend to perform better than 3x the index over the long term.
but I'd like to read more recent discussions on the topic if any, particularly relating to the current macro-environment.

I've been using this interesting strategy for about a year and was heavily down in March, but averaged down on UPRO to ~27 during the bottom, with a bit of TQQQ purchased near the bottom (though just around 10%). I've been slowly shifting into TMF but I'm considering terminating my leveraged positions for now, as we seem to be in bubble territory propped up by the Fed; It's not clear to me that with interest rates so low, treasuries are in the same category that they've been in decades.
It seems to me that reality will catch up in the medium term, and hedging appears expensive. But of course, the market can remain weird longer than anyone can stay solvent :-)

That too I'm sure has been discussed--I'm reading through all of your interesting posts, but it would be really nice to have an index feature!

Would welcome any insights or corrections--thank you!
I've been trading TQQQ monthly and making about 10% returns each month on 5% of my portfolio. I try to minimize my days holding as much as possible, but I hold to at least to settlement, and I don't buy again until I have settled cash. I buy on the big dip days, where it drops at least 3 points or more. My latest purchase is only up 5%, but I expect to sell some time next week. I have no clue if this is a bubble or not. But I'll cut my losses at the 20 day moving average for TQQQ. The rest of my portfolio I look to sell at the 300 day moving average.

I'll explore the strategy of shifting back and forth with TMF when the VXN gets back into the 20 territory, where the back tests look good for long term holding.

So far this has been a good experience, but it sure took some time to get my stomach used to the big daily moves.
User avatar
coingaroo
Posts: 93
Joined: Fri Apr 26, 2019 11:31 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by coingaroo »

privatefarmer wrote: Sun May 03, 2020 6:48 pm I keep hearing that “well what if the fed doesn’t want negative rates”. Guess what? It’s not up to the fed. We just had negative rates a few weeks ago! It’s up to the market. The fed can only influence. If the rest of the world goes into negative rates, money will flood over here pushing our rates negative as well. It’s market driven like everything else.

If institutional investors crave a safe haven so badly that they’re willing to intentionally give up some principle then rates certainly can be negative. Especially on the short end of curve.

Bond funds make money several ways, one being a steep yield curve. The reality is that VUSTX over the last 33 years has had a relatively stable 5% CAGR (real) despite rates starting high and ending low. If all these assumptions were true about LTT yields then you would’ve expected VUSTX to have had a much higher return over the first 15ish years vs the latter.
Hahaha no. The Fed owns 22% of US govt debt. When rates are about to reach the negative, the Fed will simply start selling govt debt, increasing yields.

The mere thought of this is enough to prevent yields from reaching in the negative territory. Don't bet against the Fed.

----

I saw a few other commentators here mention 85% UPRO and 15% VIXY, and it does seem to be quite an attractive alternative with zero bond exposure. I arrived at this figure by calculating the most efficient (highest historical sharpe) portfolio with SPY and VIXY, and arrived at 95% SPY and 5% VIXY.

Such a portfolio has 87% of the returns of SPY, however at 50% of the max drawdown.

Seek a 3x leverage and you have 85% UPRO and 15% VIXY.

Anyone running this strategy, potentially in combination with the classic long term bond yield parity? I'm considering 60/40 TQQQ/TMF and 85/15 UPRO/VIXY.
User avatar
danyboy7
Posts: 198
Joined: Tue Jul 14, 2020 11:35 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by danyboy7 »

coingaroo wrote: Sun Jul 19, 2020 5:08 am
privatefarmer wrote: Sun May 03, 2020 6:48 pm I keep hearing that “well what if the fed doesn’t want negative rates”. Guess what? It’s not up to the fed. We just had negative rates a few weeks ago! It’s up to the market. The fed can only influence. If the rest of the world goes into negative rates, money will flood over here pushing our rates negative as well. It’s market driven like everything else.

If institutional investors crave a safe haven so badly that they’re willing to intentionally give up some principle then rates certainly can be negative. Especially on the short end of curve.

Bond funds make money several ways, one being a steep yield curve. The reality is that VUSTX over the last 33 years has had a relatively stable 5% CAGR (real) despite rates starting high and ending low. If all these assumptions were true about LTT yields then you would’ve expected VUSTX to have had a much higher return over the first 15ish years vs the latter.
Hahaha no. The Fed owns 22% of US govt debt. When rates are about to reach the negative, the Fed will simply start selling govt debt, increasing yields.

The mere thought of this is enough to prevent yields from reaching in the negative territory. Don't bet against the Fed.

----

I saw a few other commentators here mention 85% UPRO and 15% VIXY, and it does seem to be quite an attractive alternative with zero bond exposure. I arrived at this figure by calculating the most efficient (highest historical sharpe) portfolio with SPY and VIXY, and arrived at 95% SPY and 5% VIXY.
Such a portfolio has 87% of the returns of SPY, however at 50% of the max drawdown.

Seek a 3x leverage and you have 85% UPRO and 15% VIXY.

Anyone running this strategy, potentially in combination with the classic long term bond yield parity? I'm considering 60/40 TQQQ/TMF and 85/15 UPRO/VIXY.
Bakctests with Vixy goes only until 2012,it' not enough imho
I have seen the light
KSActuary
Posts: 611
Joined: Fri Jan 13, 2012 10:53 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by KSActuary »

coingaroo wrote: Sun Jul 19, 2020 5:08 am
privatefarmer wrote: Sun May 03, 2020 6:48 pm I keep hearing that “well what if the fed doesn’t want negative rates”. Guess what? It’s not up to the fed. We just had negative rates a few weeks ago! It’s up to the market. The fed can only influence. If the rest of the world goes into negative rates, money will flood over here pushing our rates negative as well. It’s market driven like everything else.

If institutional investors crave a safe haven so badly that they’re willing to intentionally give up some principle then rates certainly can be negative. Especially on the short end of curve.

Bond funds make money several ways, one being a steep yield curve. The reality is that VUSTX over the last 33 years has had a relatively stable 5% CAGR (real) despite rates starting high and ending low. If all these assumptions were true about LTT yields then you would’ve expected VUSTX to have had a much higher return over the first 15ish years vs the latter.
Hahaha no. The Fed owns 22% of US govt debt. When rates are about to reach the negative, the Fed will simply start selling govt debt, increasing yields.

The mere thought of this is enough to prevent yields from reaching in the negative territory. Don't bet against the Fed.

----

I saw a few other commentators here mention 85% UPRO and 15% VIXY, and it does seem to be quite an attractive alternative with zero bond exposure. I arrived at this figure by calculating the most efficient (highest historical sharpe) portfolio with SPY and VIXY, and arrived at 95% SPY and 5% VIXY.

Such a portfolio has 87% of the returns of SPY, however at 50% of the max drawdown.

Seek a 3x leverage and you have 85% UPRO and 15% VIXY.

Anyone running this strategy, potentially in combination with the classic long term bond yield parity? I'm considering 60/40 TQQQ/TMF and 85/15 UPRO/VIXY.
If your implying that the Feds will begin reducing their balance sheet then I wouldn't want to be in UPRO or TMF.
User avatar
privatefarmer
Posts: 710
Joined: Mon Sep 08, 2014 2:45 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by privatefarmer »

It’s a global economy. When US yields rise, investors from all the over the world rush in to buy up treasuries bringing the yields back down. To think that the fed can solely control yields is foolish. When so many trillions of dollars worth of foreign debt is negative yielding, it is not unreasonable to think the US will follow.
Verto
Posts: 3
Joined: Wed Jun 24, 2020 9:10 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Verto »

KSActuary wrote: Sun Jul 19, 2020 10:39 am
coingaroo wrote: Sun Jul 19, 2020 5:08 am
privatefarmer wrote: Sun May 03, 2020 6:48 pm I keep hearing that “well what if the fed doesn’t want negative rates”. Guess what? It’s not up to the fed. We just had negative rates a few weeks ago! It’s up to the market. The fed can only influence. If the rest of the world goes into negative rates, money will flood over here pushing our rates negative as well. It’s market driven like everything else.

If institutional investors crave a safe haven so badly that they’re willing to intentionally give up some principle then rates certainly can be negative. Especially on the short end of curve.

Bond funds make money several ways, one being a steep yield curve. The reality is that VUSTX over the last 33 years has had a relatively stable 5% CAGR (real) despite rates starting high and ending low. If all these assumptions were true about LTT yields then you would’ve expected VUSTX to have had a much higher return over the first 15ish years vs the latter.
Hahaha no. The Fed owns 22% of US govt debt. When rates are about to reach the negative, the Fed will simply start selling govt debt, increasing yields.

The mere thought of this is enough to prevent yields from reaching in the negative territory. Don't bet against the Fed.

----

I saw a few other commentators here mention 85% UPRO and 15% VIXY, and it does seem to be quite an attractive alternative with zero bond exposure. I arrived at this figure by calculating the most efficient (highest historical sharpe) portfolio with SPY and VIXY, and arrived at 95% SPY and 5% VIXY.

Such a portfolio has 87% of the returns of SPY, however at 50% of the max drawdown.

Seek a 3x leverage and you have 85% UPRO and 15% VIXY.

Anyone running this strategy, potentially in combination with the classic long term bond yield parity? I'm considering 60/40 TQQQ/TMF and 85/15 UPRO/VIXY.
If your implying that the Feds will begin reducing their balance sheet then I wouldn't want to be in UPRO or TMF.
Massively up year to date, but last I checked they have already started to reduce their balance sheet.
Semantics
Posts: 168
Joined: Tue Mar 10, 2020 1:42 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Semantics »

coingaroo wrote: Sun Jul 19, 2020 5:08 am
privatefarmer wrote: Sun May 03, 2020 6:48 pm I keep hearing that “well what if the fed doesn’t want negative rates”. Guess what? It’s not up to the fed. We just had negative rates a few weeks ago! It’s up to the market. The fed can only influence. If the rest of the world goes into negative rates, money will flood over here pushing our rates negative as well. It’s market driven like everything else.

If institutional investors crave a safe haven so badly that they’re willing to intentionally give up some principle then rates certainly can be negative. Especially on the short end of curve.

Bond funds make money several ways, one being a steep yield curve. The reality is that VUSTX over the last 33 years has had a relatively stable 5% CAGR (real) despite rates starting high and ending low. If all these assumptions were true about LTT yields then you would’ve expected VUSTX to have had a much higher return over the first 15ish years vs the latter.
Hahaha no. The Fed owns 22% of US govt debt. When rates are about to reach the negative, the Fed will simply start selling govt debt, increasing yields.

The mere thought of this is enough to prevent yields from reaching in the negative territory. Don't bet against the Fed.

----

I saw a few other commentators here mention 85% UPRO and 15% VIXY, and it does seem to be quite an attractive alternative with zero bond exposure. I arrived at this figure by calculating the most efficient (highest historical sharpe) portfolio with SPY and VIXY, and arrived at 95% SPY and 5% VIXY.

Such a portfolio has 87% of the returns of SPY, however at 50% of the max drawdown.

Seek a 3x leverage and you have 85% UPRO and 15% VIXY.

Anyone running this strategy, potentially in combination with the classic long term bond yield parity? I'm considering 60/40 TQQQ/TMF and 85/15 UPRO/VIXY.
I was playing with a mini-adventure with 85% TQQQ and 15% VIXY briefly, but backed out of it when short-term VIX futures dropped below mid-term futures recently - the mean reverting nature starting when it's already high combined with curve slippage just seems to make it too expensive a hedge. XVZ might be an interesting alternative, it dynamically shorts a portion of short-term futures to combat slippage and can adjust daily, but it doesn't look like it would have done much in 2018, and can still incur slippage at the far end of the curve. So it seems like more of a tail risk hedge (similar with the VXTH strategy) than a useful asset during a prolonged bear market.
User avatar
coingaroo
Posts: 93
Joined: Fri Apr 26, 2019 11:31 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by coingaroo »

I purchased some synthetic VIXY backtested data. You are right: it is ineffective in slow market declines like the GFC, providing ineffective protection.

I guess stocks and bond it is. Sad, with bonds at near zero interest rates.

Image

Any other strategies (including algorithmically traded ones) negatively correlated to stocks?
Micronite
Posts: 6
Joined: Tue Apr 07, 2020 1:57 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Micronite »

I am also concerned about inflation caused by the stimulus to overcome the pandemic. We have put so much currency into circulation that inflation will spike once velocity returns to normal.

I am looking for ways to hedge this strategy against inflation and rising rates. I am considering 40% UPRO / 30% TMF / 30% LTPZ (long tips) or 40% UPRO / 60% TYD (3x intermediate). How would these perform in a rapidly rising interest rate environment? I do not have enough data to perform a backtest prior to 2010.

https://www.portfoliovisualizer.com/bac ... tion4_2=60
taojaxx
Posts: 111
Joined: Wed Jul 18, 2012 8:25 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by taojaxx »

Micronite wrote: Mon Jul 20, 2020 7:00 am I am also concerned about inflation caused by the stimulus to overcome the pandemic. We have put so much currency into circulation that inflation will spike once velocity returns to normal.

I am looking for ways to hedge this strategy against inflation and rising rates. I am considering 40% UPRO / 30% TMF / 30% LTPZ (long tips) or 40% UPRO / 60% TYD (3x intermediate). How would these perform in a rapidly rising interest rate environment? I do not have enough data to perform a backtest prior to 2010.

https://www.portfoliovisualizer.com/bac ... tion4_2=60
Unleveraged LTPZ will be a constant drag and a costly insurance. 55% UPRO, 22% TMF, 67% LTPZ (LT TIPS from PIMCO) and -44% Cash is what I do, works so far. Need a margin account for this obviously.
User avatar
Meaty
Posts: 796
Joined: Mon Jul 22, 2013 7:35 pm
Location: Texas

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Meaty »

I know TQQQ has been discussed but has anyone thought about 45 UPRO/45TMF/10 TQQQ? Per PV, it adds nearly 2 CAGR, lowers max drawdown by 5%, and improves sharpe ratio modestly by .02
"Discipline equals Freedom" - Jocko Willink
langlands
Posts: 612
Joined: Wed Apr 03, 2019 10:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by langlands »

Meaty wrote: Mon Jul 20, 2020 1:28 pm I know TQQQ has been discussed but has anyone thought about 45 UPRO/45TMF/10 TQQQ? Per PV, it adds nearly 2 CAGR, lowers max drawdown by 5%, and improves sharpe ratio modestly by .02
I like the idea of adding TQQQ because I like tech. But the back-testing statistics you give is all just overfit. As a general rule, posters on this site and in this thread in particular rely way too much on back testing. Back testing is good for very general trends (i.e. stocks tend to go up in the long run and outperform bonds, but bonds have lower volatility) and should be used sparingly. Tweaking allocations and trying to find the one that leads to the best back-tested results is completely overfitting and useless.
User avatar
Meaty
Posts: 796
Joined: Mon Jul 22, 2013 7:35 pm
Location: Texas

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Meaty »

langlands wrote: Mon Jul 20, 2020 3:43 pm
Meaty wrote: Mon Jul 20, 2020 1:28 pm I know TQQQ has been discussed but has anyone thought about 45 UPRO/45TMF/10 TQQQ? Per PV, it adds nearly 2 CAGR, lowers max drawdown by 5%, and improves sharpe ratio modestly by .02
I like the idea of adding TQQQ because I like tech. But the back-testing statistics you give is all just overfit. As a general rule, posters on this site and in this thread in particular rely way too much on back testing. Back testing is good for very general trends (i.e. stocks tend to go up in the long run and outperform bonds, but bonds have lower volatility) and should be used sparingly. Tweaking allocations and trying to find the one that leads to the best back-tested results is completely overfitting and useless.
Thanks for your reply. Is there other data or rational that would suggest adding TQQQ as I’ve specified is sub optimal compared to 55 UPRO / 45 TMF?
"Discipline equals Freedom" - Jocko Willink
langlands
Posts: 612
Joined: Wed Apr 03, 2019 10:05 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by langlands »

Meaty wrote: Mon Jul 20, 2020 3:49 pm
langlands wrote: Mon Jul 20, 2020 3:43 pm
Meaty wrote: Mon Jul 20, 2020 1:28 pm I know TQQQ has been discussed but has anyone thought about 45 UPRO/45TMF/10 TQQQ? Per PV, it adds nearly 2 CAGR, lowers max drawdown by 5%, and improves sharpe ratio modestly by .02
I like the idea of adding TQQQ because I like tech. But the back-testing statistics you give is all just overfit. As a general rule, posters on this site and in this thread in particular rely way too much on back testing. Back testing is good for very general trends (i.e. stocks tend to go up in the long run and outperform bonds, but bonds have lower volatility) and should be used sparingly. Tweaking allocations and trying to find the one that leads to the best back-tested results is completely overfitting and useless.
Thanks for your reply. Is there other data or rational that would suggest adding TQQQ as I’ve specified is sub optimal compared to 55 UPRO / 45 TMF?
I don't think it's suboptimal. The following is my general framework for thinking about this:

S&P 500, QQQ, and bonds have been around a long time. In my opinion, the data required are your best estimate of the return of these 3 assets, the volatility of these 3 assets, and the correlations between these 3 assets. That's 9 numbers. Volatility can be measured very reliably, so just take an estimate of the volatility over the past few months (the fact that volatility can be measured so accurately is why some people are doing target volatility adjustments). Correlation is less reliable, but still very robust (you could use say last 3 years or so of data). Returns are of course the holy grail and highly unpredictable. Use your best judgement. Probably the average return of US stock market over the entire past century is a good estimate for S&P 500 return. I'd add a percentage point or so for QQQ (because I like tech) but you could subtract 1 or leave it the same. If you believe the CAPE Shiller stuff, you could also adjust for our currently high P/E. Bonds...yeah I have no idea. I think to do this strategy, you need an opinion on bonds, and because I don't, I don't do this adventure. It seems the standard thing to do is to just take the current yield as the expected return going forward.

Once you have these numbers, you can find the optimal allocation. Depending on how comfortable you are with matrix algebra, the formula is given in section 1.2 ( The Efficient Frontier with a Risk-free Asset) of http://www.columbia.edu/~mh2078/Foundat ... e-CAPM.pdf. Essentially you invert the covariance matrix and multiply by the vector of returns (adjusted by risk free interest rate). That gives you the mix of the three assets you want.
User avatar
Meaty
Posts: 796
Joined: Mon Jul 22, 2013 7:35 pm
Location: Texas

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Meaty »

langlands wrote: Mon Jul 20, 2020 4:46 pm
Meaty wrote: Mon Jul 20, 2020 3:49 pm
langlands wrote: Mon Jul 20, 2020 3:43 pm
Meaty wrote: Mon Jul 20, 2020 1:28 pm I know TQQQ has been discussed but has anyone thought about 45 UPRO/45TMF/10 TQQQ? Per PV, it adds nearly 2 CAGR, lowers max drawdown by 5%, and improves sharpe ratio modestly by .02
I like the idea of adding TQQQ because I like tech. But the back-testing statistics you give is all just overfit. As a general rule, posters on this site and in this thread in particular rely way too much on back testing. Back testing is good for very general trends (i.e. stocks tend to go up in the long run and outperform bonds, but bonds have lower volatility) and should be used sparingly. Tweaking allocations and trying to find the one that leads to the best back-tested results is completely overfitting and useless.
Thanks for your reply. Is there other data or rational that would suggest adding TQQQ as I’ve specified is sub optimal compared to 55 UPRO / 45 TMF?
I don't think it's suboptimal. The following is my general framework for thinking about this:

S&P 500, QQQ, and bonds have been around a long time. In my opinion, the data required are your best estimate of the return of these 3 assets, the volatility of these 3 assets, and the correlations between these 3 assets. That's 9 numbers. Volatility can be measured very reliably, so just take an estimate of the volatility over the past few months (the fact that volatility can be measured so accurately is why some people are doing target volatility adjustments). Correlation is less reliable, but still very robust (you could use say last 3 years or so of data). Returns are of course the holy grail and highly unpredictable. Use your best judgement. Probably the average return of US stock market over the entire past century is a good estimate for S&P 500 return. I'd add a percentage point or so for QQQ (because I like tech) but you could subtract 1 or leave it the same. If you believe the CAPE Shiller stuff, you could also adjust for our currently high P/E. Bonds...yeah I have no idea. I think to do this strategy, you need an opinion on bonds, and because I don't, I don't do this adventure. It seems the standard thing to do is to just take the current yield as the expected return going forward.

Once you have these numbers, you can find the optimal allocation. Depending on how comfortable you are with matrix algebra, the formula is given in section 1.2 ( The Efficient Frontier with a Risk-free Asset) of http://www.columbia.edu/~mh2078/Foundat ... e-CAPM.pdf. Essentially you invert the covariance matrix and multiply by the vector of returns (adjusted by risk free interest rate). That gives you the mix of the three assets you want.
Thank you. This is very helpful. I’m already in the adventure for about a year. Up 30% but looking to add TQQQ if it makes sense
"Discipline equals Freedom" - Jocko Willink
Gufomel
Posts: 563
Joined: Sat Feb 14, 2015 9:52 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Gufomel »

On days (or longer periods) where UPRO and TMF are both up, where’s the money flowing out of? Cash I suppose? Even gold is up.
jarjarM
Posts: 305
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

Just food for thought, I decided to simulate several of strategies via python as a pet project described in the threads:

.........................cagr (%)......sharpe........max_drawdown (%, based on daily)
SP500x3 Baseline.......28.3.........0.7...............-76.8
60/40 Portfolio.........31.4.........0.9...............-61.6
VolatilityPortfolio......33.5.........1.3...............-43.3
MomentumPortfolio....30.7.........1.3...............-45.9
AdaptiveAssetAll........34.5.........1.3...............-43.4


Image

Note: Volatility is measured using 22 days, momentum is measured using 6 months and adaptive asset allocation is based on minimal variance from the link below.

https://www.investresolve.com/inc/uploa ... epaper.pdf
jarjarM
Posts: 305
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

cos wrote: Wed Jul 15, 2020 2:57 pm
randyharris wrote: Mon Jul 13, 2020 11:05 pm
glenmalan wrote: Mon Jul 13, 2020 9:34 pmFor the Adaptive model, how many timing periods are you using? I use 1.
Volatility Trading Days? I use 21
Allocation weight method? I use inverse volatility
Trading Frequency? I use monthly
This is the setup, for me, with either TQQQ/TMF, or QLD/UBT.

https://www.portfoliovisualizer.com/te ... odWeight=0
How did you arrive at these parameters? Why a single performance period? Why is the performance period 1 month? Why is the volatility period 21 days?

Did you use something along the lines of the following logic to choose minimum variance over inverse volatility?
Image
(see: viewtopic.php?t=277661)
Not randyharris, but my "excellent adventure" is of the similar vein. I use 22 trading days (equivalent to 1 month look back), also trading monthly (this is somewhat important if you do decide to go with adaptive asset allocation methodology) but I use minimal variance since the CAGR is a ~1% better from 2009 -2020.

My simulated result:
Image
Walkure
Posts: 290
Joined: Tue Apr 11, 2017 9:59 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Walkure »

Gufomel wrote: Tue Jul 21, 2020 11:17 am On days (or longer periods) where UPRO and TMF are both up, where’s the money flowing out of? Cash I suppose? Even gold is up.
UPRO probably also gets an boost from the ongoing herd movement out of small value and into large growth.
User avatar
cos
Posts: 235
Joined: Fri Aug 23, 2019 7:34 pm
Location: Boston
Contact:

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by cos »

jarjarM wrote: Tue Jul 21, 2020 7:56 pm Not randyharris, but my "excellent adventure" is of the similar vein. I use 22 trading days (equivalent to 1 month look back), also trading monthly (this is somewhat important if you do decide to go with adaptive asset allocation methodology) but I use minimal variance since the CAGR is a ~1% better from 2009 -2020.

My simulated result:
Image
If you're using a Python script to calculate allocation weights every month, would you mind sharing the code here?
jolmscheid
Posts: 78
Joined: Fri Feb 24, 2017 1:22 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jolmscheid »

jarjarM wrote: Tue Jul 21, 2020 7:46 pm Just food for thought, I decided to simulate several of strategies via python as a pet project described in the threads:

.........................cagr (%)......sharpe........max_drawdown (%, based on daily)
SP500x3 Baseline.......28.3.........0.7...............-76.8
60/40 Portfolio.........31.4.........0.9...............-61.6
VolatilityPortfolio......33.5.........1.3...............-43.3
MomentumPortfolio....30.7.........1.3...............-45.9
AdaptiveAssetAll........34.5.........1.3...............-43.4


Image

Note: Volatility is measured using 22 days, momentum is measured using 6 months and adaptive asset allocation is based on minimal variance from the link below.

https://www.investresolve.com/inc/uploa ... epaper.pdf

For adaptive asset allocation, you are just inputting UPRO and TMF into portfolio Visualizer and it "spits out" each month what the allocation should be for the next month?
jarjarM
Posts: 305
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

jolmscheid wrote: Tue Jul 21, 2020 8:44 pm
For adaptive asset allocation, you are just inputting UPRO and TMF into portfolio Visualizer and it "spits out" each month what the allocation should be for the next month?
No, in my chart it's based on solving the minimum variance using the critical-line algorithm (1 month vol and 6 month momentum look back and extracting minimum variance). I don't use portfoliovisualizer since this was a pet project to keep me sane during the shelter-in-place fun :oops: Plus I have a lot more control of the hyper parameters during tuning and better understanding how the solver is actually working. I assume if portfoliovisualizer is solving for minimum variance then the output is similar.

https://www.researchgate.net/publicatio ... timization
jarjarM
Posts: 305
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

cos wrote: Tue Jul 21, 2020 8:28 pm If you're using a Python script to calculate allocation weights every month, would you mind sharing the code here?
Give me a few days to clean up some stuff. Also, I pull my data from Tiingo so you'll either have to alter the code or create an account there.
djeayzonne
Posts: 101
Joined: Wed Dec 06, 2017 2:14 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by djeayzonne »

jarjarM wrote: Tue Jul 21, 2020 9:32 pm
cos wrote: Tue Jul 21, 2020 8:28 pm If you're using a Python script to calculate allocation weights every month, would you mind sharing the code here?
Give me a few days to clean up some stuff. Also, I pull my data from Tiingo so you'll either have to alter the code or create an account there.
I would also be very interested in this.
perfectuncertainty
Posts: 161
Joined: Sun Feb 04, 2018 7:44 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by perfectuncertainty »

jarjarM wrote: Tue Jul 21, 2020 9:32 pm
cos wrote: Tue Jul 21, 2020 8:28 pm If you're using a Python script to calculate allocation weights every month, would you mind sharing the code here?
Give me a few days to clean up some stuff. Also, I pull my data from Tiingo so you'll either have to alter the code or create an account there.
That's very interesting. Is all your data using only UPRO and TMF?
jarjarM
Posts: 305
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

glenmalan wrote: Tue Jul 21, 2020 10:55 pm
jarjarM wrote: Tue Jul 21, 2020 9:32 pm
cos wrote: Tue Jul 21, 2020 8:28 pm If you're using a Python script to calculate allocation weights every month, would you mind sharing the code here?
Give me a few days to clean up some stuff. Also, I pull my data from Tiingo so you'll either have to alter the code or create an account there.
That's very interesting. Is all your data using only UPRO and TMF?
Yes, but obviously this can be done on multiple positions. The original paper was looking at a 10 position universe and use momentum to select the top 5 and solve for minimal variance.
perfectuncertainty
Posts: 161
Joined: Sun Feb 04, 2018 7:44 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by perfectuncertainty »

jarjarM wrote: Tue Jul 21, 2020 11:00 pm
glenmalan wrote: Tue Jul 21, 2020 10:55 pm
jarjarM wrote: Tue Jul 21, 2020 9:32 pm
cos wrote: Tue Jul 21, 2020 8:28 pm If you're using a Python script to calculate allocation weights every month, would you mind sharing the code here?
Give me a few days to clean up some stuff. Also, I pull my data from Tiingo so you'll either have to alter the code or create an account there.
That's very interesting. Is all your data using only UPRO and TMF?
Yes, but obviously this can be done on multiple positions. The original paper was looking at a 10 position universe and use momentum to select the top 5 and solve for minimal variance.
I have noted from my own testing in PV that Minimal Variance works well with 2 symbols, but when using 3 (like TQQQ, UPRO and TMF) inverse volatility has quite a bit higher CAGR. Have you looked at that possibility?
Semantics
Posts: 168
Joined: Tue Mar 10, 2020 1:42 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Semantics »

glenmalan wrote: Tue Jul 21, 2020 11:15 pm
jarjarM wrote: Tue Jul 21, 2020 11:00 pm
glenmalan wrote: Tue Jul 21, 2020 10:55 pm
jarjarM wrote: Tue Jul 21, 2020 9:32 pm
cos wrote: Tue Jul 21, 2020 8:28 pm If you're using a Python script to calculate allocation weights every month, would you mind sharing the code here?
Give me a few days to clean up some stuff. Also, I pull my data from Tiingo so you'll either have to alter the code or create an account there.
That's very interesting. Is all your data using only UPRO and TMF?
Yes, but obviously this can be done on multiple positions. The original paper was looking at a 10 position universe and use momentum to select the top 5 and solve for minimal variance.
I have noted from my own testing in PV that Minimal Variance works well with 2 symbols, but when using 3 (like TQQQ, UPRO and TMF) inverse volatility has quite a bit higher CAGR. Have you looked at that possibility?
That's because inverse volatility treats the assets as though they are independent and is doubling up on equities, which is the higher performing asset class. I would imagine that you'll get higher std dev and lower Sharpe ratio though. Min variance will allocate less weight to TQQQ+UPRO because they are very strongly correlated. In general you should get better risk-adjusted returns with min variance (if you believe past correlations predict future ones).
jarjarM
Posts: 305
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

Semantics wrote: Tue Jul 21, 2020 11:27 pm
glenmalan wrote: Tue Jul 21, 2020 11:15 pm I have noted from my own testing in PV that Minimal Variance works well with 2 symbols, but when using 3 (like TQQQ, UPRO and TMF) inverse volatility has quite a bit higher CAGR. Have you looked at that possibility?
That's because inverse volatility treats the assets as though they are independent and is doubling up on equities, which is the higher performing asset class. I would imagine that you'll get higher std dev and lower Sharpe ratio though. Min variance will allocate less weight to TQQQ+UPRO because they are very strongly correlated. In general you should get better risk-adjusted returns with min variance (if you believe past correlations predict future ones).
Semantics is right, volatility will strongly bias toward high equity ratio. The interesting thing is that momentum/volatility combo actually yields a better CAGR (+2%) with reasonable sharp ratio. I’ll post the results after I double check some work tomorrow.
BTW, does anyone know if the max drawdown from portfolio visualizer is based on monthly or daily? Thanks.
User avatar
danyboy7
Posts: 198
Joined: Tue Jul 14, 2020 11:35 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by danyboy7 »

jarjarM wrote: Wed Jul 22, 2020 2:03 am
Semantics wrote: Tue Jul 21, 2020 11:27 pm
glenmalan wrote: Tue Jul 21, 2020 11:15 pm I have noted from my own testing in PV that Minimal Variance works well with 2 symbols, but when using 3 (like TQQQ, UPRO and TMF) inverse volatility has quite a bit higher CAGR. Have you looked at that possibility?
That's because inverse volatility treats the assets as though they are independent and is doubling up on equities, which is the higher performing asset class. I would imagine that you'll get higher std dev and lower Sharpe ratio though. Min variance will allocate less weight to TQQQ+UPRO because they are very strongly correlated. In general you should get better risk-adjusted returns with min variance (if you believe past correlations predict future ones).
Semantics is right, volatility will strongly bias toward high equity ratio. The interesting thing is that momentum/volatility combo actually yields a better CAGR (+2%) with reasonable sharp ratio. I’ll post the results after I double check some work tomorrow.
BTW, does anyone know if the max drawdown from portfolio visualizer is based on monthly or daily? Thanks.
Thanks,waiting for that. Post also a PV link please :wink:
I have seen the light
perfectuncertainty
Posts: 161
Joined: Sun Feb 04, 2018 7:44 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by perfectuncertainty »

jarjarM wrote: Wed Jul 22, 2020 2:03 am
Semantics wrote: Tue Jul 21, 2020 11:27 pm
glenmalan wrote: Tue Jul 21, 2020 11:15 pm I have noted from my own testing in PV that Minimal Variance works well with 2 symbols, but when using 3 (like TQQQ, UPRO and TMF) inverse volatility has quite a bit higher CAGR. Have you looked at that possibility?
That's because inverse volatility treats the assets as though they are independent and is doubling up on equities, which is the higher performing asset class. I would imagine that you'll get higher std dev and lower Sharpe ratio though. Min variance will allocate less weight to TQQQ+UPRO because they are very strongly correlated. In general you should get better risk-adjusted returns with min variance (if you believe past correlations predict future ones).
Semantics is right, volatility will strongly bias toward high equity ratio. The interesting thing is that momentum/volatility combo actually yields a better CAGR (+2%) with reasonable sharp ratio. I’ll post the results after I double check some work tomorrow.
BTW, does anyone know if the max drawdown from portfolio visualizer is based on monthly or daily? Thanks.
I believe it is monthly. The Max Drawdown in your calculation seems to be considerably higher than PV.
Mickelous
Posts: 134
Joined: Mon Apr 20, 2020 11:24 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Mickelous »

Image
Image
roasted sweet potato

401k looking good. Just wanted to show some proof of the adventure in action. I've never been more convinced of a passive investing strategy.

I have more in other portfolios this is just the first one I switched over.
jarjarM
Posts: 305
Joined: Mon Jul 16, 2018 1:21 pm

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by jarjarM »

jarjarM wrote: Wed Jul 22, 2020 2:03 am
Semantics wrote: Tue Jul 21, 2020 11:27 pm
glenmalan wrote: Tue Jul 21, 2020 11:15 pm I have noted from my own testing in PV that Minimal Variance works well with 2 symbols, but when using 3 (like TQQQ, UPRO and TMF) inverse volatility has quite a bit higher CAGR. Have you looked at that possibility?
That's because inverse volatility treats the assets as though they are independent and is doubling up on equities, which is the higher performing asset class. I would imagine that you'll get higher std dev and lower Sharpe ratio though. Min variance will allocate less weight to TQQQ+UPRO because they are very strongly correlated. In general you should get better risk-adjusted returns with min variance (if you believe past correlations predict future ones).
Semantics is right, volatility will strongly bias toward high equity ratio. The interesting thing is that momentum/volatility combo actually yields a better CAGR (+2%) with reasonable sharp ratio. I’ll post the results after I double check some work tomorrow.
BTW, does anyone know if the max drawdown from portfolio visualizer is based on monthly or daily? Thanks.
This is why one should never post right before putting a naughty kid to bed. Here is the data after looking at data from 2010 to 2020. Best of course is just straight up TQQQ with ~50% CAGR but of course the drawdown is significant and sharpe ratio is on the low side. For the time period prescribed, TQQQ is so strong that it's easy just go with 60/40 weighted and balance monthly and you'll get very good result. 30/30/40 is UPRO/TQQQ/TMF balanced monthly and that's weaker due to UPRO underperform TQQQ by wide margin after 10 years of run. Volatility and Momentum and their combo basically will lean heavily on TQQQ and the performance shows it. Adaptive with minimal variance can't keep up since it place emphasis on low correlation and hence heavier weighting toward TMF. Of course, all of these are in the backdrop of the strongest decade for QQQ since in its inception (it was 1.70 at start and now at 117, a 68x gain). I won't bet on another decade of such outperformance but what do I know.

.........................cagr (%).....sharpe........max_drawdown (%)
TQQQ Baseline........50............1.0...............-69.9
60/40 TQQQ/TMF.....44...........1.4...............-48.7
30/30/40 Portfolio....38...........1.3..............-49.9
VolatilityPortfolio.....41...........1.4..............-48.6
MomentumPortfolio...40...........1.3..............-53.4
Combo..................42............1.4..............-48.6
AdaptiveAssetAll......36............1.4..............-43.4
Image
Chicken Little
Posts: 421
Joined: Fri Feb 22, 2019 5:03 am

Re: HEDGEFUNDIE's excellent adventure Part II: The next journey

Post by Chicken Little »

Mickelous wrote: Wed Jul 22, 2020 10:45 am Image
Image
roasted sweet potato

401k looking good. Just wanted to show some proof of the adventure in action. I've never been more convinced of a passive investing strategy.

I have more in other portfolios this is just the first one I switched over.
Numbers are worth thousands of words.

Thanks for posting.
Post Reply