DeepState
Multi-Strategy, Quant and Fundamental
- Joined
- 30 March 2014
- Posts
- 1,615
- Reactions
- 81
Can I please give you a hug? Damn, where have you been all this time????
Understand dollar cost averaging fully and have the disposition to implement it consistently, then GFC type volatility is actually the sort of event you hope for.
Monte Carlo in excel? Ew....
</tangent>
You are welcome. Thankyou for your somewhat more reasoned tone.
You are pointing out a valid scenario which, as per BRTY's single scenario, led to the underperformance of re-bal vs buy-hold. In the extensive simulation, I mentioned that fully 40% failed. Amongst these will be exactly the scenarios you and BRTY have encountered, experienced or imagine might occur.
Re-bal does not claim to be pure arbitrage. I have not claimed it is. The equations, if you were to translate them, don't indicate that either. There is risk of failure as with any strategy you could name other than printing money.
So, given that we have risk of failure, why do it? Because it has positive expectancy and the edge is not trivial even after tax and comm.
Each of the scenarios you have mentioned are real. They can happen. As are a range of great scenarios that favour re-bal. What is missing from this debate is the assignment of probability to these scenarios. I know it sounds vague and all, but perhaps I can explain this via a casino.
Welcome to Cynic-ville Sentosa, a bloody awesome marble walled casino with lights and glamour. You, in your hood with digits running over your face are the Roulette dealer. You have odds stacked in your favour. Now, the kinds of scenarios you are mentioning are like drawing 5x 32s in a row with some punter re-staking that slot each time. The Casino loses. It's a really bad day. But the chances of that are the same as drawing 5x 0 in a row. Or some mix of all the numbers. They are all equally likely. Bad and good scenarios. But when you add it all up, the odds in your favour (re-bal profits) make the casino money through time. You need to assign probability to each scenario and weight it accordingly. At the moment, what you are doing is pointing out the possibility of a bad outcome. No argument from me. They exist. But go on to weight that and all the other ones and you come up with positive expectancy. It's just unnatural to think it terms of these distributions, so I appreciate this is going to be a bit tough. People more naturally think in terms of scenarios as you have done. And, by internal wiring from our ancestors from caveman time, focus more on bad scenarios than good.
There is a note/thesis to So_Cynical (oh, the irony) in the thread "Does Rebalancing Work?" which explains how you can prove it to yourself. It just requires a spreadsheet and a tiny bit of work.
It doesn't need a rocket scientist to see the massive flaw in rebal. They wouldn't see it because they understand this stuff. In fact, I have a friend who actually is a rocket scientist managing portfolio this way. They are in Australia, manage $8bn and are growing like a weed. It just needs a spreadsheet that a high school student could put together in an hour to see that this stuff works in the real world. I hope you have achieved a HS level education and can pull it off. It's no skin of my nose if you don't do this and can't find the truth of it or can't be convinced of it by display of facts, theory or verbal/written obfuscation. Actually, why bother even raising this if you are self-declared closed minded and nailed shut?
Chillax dude.
Check out Accounting Standard AAS25 for how it's done in the real world with real money with real people who actually understand this stuff
Now do it 9,999 times more and let's see what happens
RY,
let's walk forward with any you care to name.
Market Vectors Indices have constructed an equal weight index.
View attachment 57773
I’m not sure of the construction of the two indexes but it looks like Australian total gross equal weight index has been having a tough time of it compared to the total market over the historical 10year chart.
They have also launched an equal weight ETF in March (Interestingly, RY joined ASF in March– you don’t have anything to disclose do you RY?) So your comparison can easily be tracked on a go forward basis. ASX code MVW.
There is also a white paper on the subject on the Market Vectors site
http://www.marketvectors-australia.com/library/education/
RY has pretty much covered most of it but the References are always useful if anybody wants to read further.
First...I hope you are kidding about being an employee of MV and joining at roughly the time of this ETF launch. For avoidance of doubt I am not, and do not endorse or support any of MV's products, its officers, associates or subsidiaries.
This is interesting. I think it compares the MV ETF you described against the ASX 200 or something. The MV ETF construction universe is different by design and this is found in:
http://www.marketvectorsindices.com...utions-licenses-australia-equal-weight-index/
Please note that the universe is different to that of the ASX 200. Secondly, please note that the index contains only 77 names - as if to highlight that further. Whilst it is valid to compare this against any index you want, if it is to be taken as a demonstration of rebal, then the stock universes of the two should be identical. It is not.
RY,
1. I don't know how some-one types so fast in so many different threads.
2. Thankyou for quoting superannuation standards, I have been talking about trading for ones own account, nothing to do with super, but I take it as another parameter you have added.
3. A person bothers to go through the work on a random portfolio of traded stocks, and in an attempt to maintain your position of how rebal works, just totally fail to acknowledge the simple fact that it will not always work. Plus the fact that there are times where it can be detrimental to a portfolio.
4. I will clearly acknowledge there are times and stocks when averaging down from your winners does work but NOT ALWAYS as you have been suggesting.
5. rebal or dollar cost averaging or whatever you want to call it, is not a panacea, is so often found on forums.
6. How about you start a thread on a portfolio of your selected stocks with the clear easy rules at the beginning, where we walk forward with actual quarterly rebal and see how it pans out. This forum has been around for years. Instead of theory, getinto the real world as some others have done. That way we can see over time how much actual outperformance your strategy delivers.
Are you up to it?
A simple 10 stock selection, quarterly rebal, compared to buy and hold, $50,000/stock start. You didn't like my selected universe so let's walk forward with any you care to name.
1.Just probing a coincidence. No offence meant.
2. As I said I don't know too much about the make up of the indices. the only conclusion I can draw from the historical comparison is that the equally weighted ETF (as its been structured) would not have beaten a whole of market ETF based on market capitalisation.
3. As for the go forward comparison that BRTY requested this seems an easy and transparent comparison of rebalancing - all that needs to be done is record the buy and hold return of the 77 odd companies and see how MVW performs against that over time. (a long time to allow expectancy to rise above randomness)
1. I have provided an analogy that accurately depicts a significant risk inherent to the rebalancing methodology. One doesn't need to be versed in mathematics, statistics, probability theory, or even casino games to be able to recognise, understand and appreciate said analogy!
2. As it happens the roullette wheel has been deliberately engineered to more precisely reflect the outcomes that might typically be expected according to probability theory.
Financial markets on the other hand, have not undergone such engineering (their intent, purpose and design are quite different!), hence the casino roullette wheel can hardly be considered a valid comparison. Furthermore, unlike the financial markets, none of the roulette numbers ever go into receivership following a prolonged downturn in fortunes!
3. Millions upon millions of repetitions are usually required before events can even be expected to conform to the dictates of probability theory, hence any expectancy suggested by such theory could not reasonably be depended upon for investment performance. (The average human life span would only afford a comparatively small event sample - the theoretical expectancy would be unlikely to eventuate!!)
4, Stocks, commodities, currencies real estate etc. do at times go into prolonged states of decline. History has shown this and continues to do so!
5. My rainbucket analogy gives ample warning of the perils of rebalancing in respect to such occurences, whereas the roullette wheel has precious little (if anything) to say on such matters!
1. Gut instinct would have me think a fair bit of rebalance performance over indexes has to do with initial equal weight effect.
2. Sometimes you just need a break - you may find that also in time..
3. What it takes to harvest one edge doesn't necessarily add to another - and in fact trying to do so can cause damage to the original edge. That's not a discussion for 12:30am but a mighty good topic for one rainy day and could probably go a long way to reconcile the different angles you and BRTY might be coming from.
4. Care to detail your data source(total return by quarters -nice) and back testing engines? They appear pretty impressive.
1. It is a small part, what other similarly simple strategies do you think give investors an edge?
2. This is fine if your focus is on total returns.
For people who invest to provide passive income to meet expenses, retire or follow other passions, the income component is more important.
Sorry TPI, I signed off too soon.
1. I think you are looking to low turnover strategies. If that is so, then Premium Based Strategies are suitable candidates. They are long term in focus an involve some notion of risk bearing that can't otherwise be absorbed by others in the market. Things like deep value which is screened for credit worthiness, buying small caps. You may have heard of Low Vol. That's achievable but more intense and needs valuation and earnings quality overrides. You can add trend stuff into the mix, which works very powerfully in Australia for some reason, but no where near as powerfully elsewhere. All that is simple.
The above is heavily mined so the basic strategies are not going to be super-profit makers any more. Unfortunately, simple stuff produces fairly simple returns in this type of scenario.
Otherwise, you get rich by not losing money. Watch for tax. Knowing your tax rate, you should either tilt to or away from dividend payers. Watch comm. There is massive excess turnover. Just doing nothing is often better than watching screens. Harvest the right parcels and know the consequences of harvesting a large CGT parcel requires a very powerful idea to justify it. This really adds up over time.
You sound like an adviser or planner. You would be aware of the horrible record of mistiming markets in retail (let along insto). So a huge part of this is simply sticking to the game plan. But, it's hard because you don't know if your game plan is busted. To me, you invest in a way that allows for really bad scenarios but not the extreme ones. When that's set, you find a way of protecting tail risks so you can hack it in the event of really awful outcomes and stick to the game plan and stay cool as the world melts around you.
That's the simple truth of it.
2. Actually, you can build a portfolio with an income focus and add re-bal on top. The concept works anyway. It would, for example, work on a portfolio of bonds.
Cheers
1. Totally agree. The scenario is fine. The analogy is very attractive and creative. My point was that that, if you are representing risk to the re-bal process, you also need to couple the scenario with a likelihood.
2. Yes, a casino game, if it isn't rigged like LIBOR, has knowable and stable odds in situations like Roulette. That's what makes them good starting points for investments and probability theory.
I just reached out to the first three texts that I have on probability and finance. After blowing the dust off them, they each use games of chance like tossing coins to introduce the idea. These texts deal with probability theory.
Title / Page of first mention of coin toss
a. DeGroot & Schervish, "Probability and Statistics", Fourth Edition / p27 of 890
b. Walpole & Myers, "Probability and Statistics for Engineers and Scientists", Fifth Edition / p11 of 765
c. Rudd & Classing, "Modern Portfolio Theory", Second Edition / p35 of 525.
If these texts alone, which are deep into Probability Theory and it's real world applications, see fit to utilize games of chance to introduce more complex issues, then for a guy who eschews all of it, it seemed reasonable to start there. Do you actually think I think the markets work like a Roulette wheel? No more than I think the markets actually operate like buckets of water where one has a leak.
Probability Theory, by the way, is perfectly applicable to games like Roulette to Poker and Black Jack. But you seem to want to go deeper.
Since you have espoused the term Probability Theory and are talking about sample size and parameter stability, you must be versed or making it up. I'll give you the benefit of the doubt. If you believe that a Roulette wheel is engineered (yes, it is) and thus is not a valid comparison with markets, then you will need to explain why the moment generating function of the sum of lognormal returns being security price returns, drawn from an unknown and flexible distribution, will not asymptote to the same description for higher moments as the one drawn from draws from a game of Roulette. But...it does. Cynic, this is the same stuff that is the bedrock of options pricing using binomial models. The entire options market disagrees with your statement.
As I've already mentioned, Casino games are simply not a valid comparison to the financial markets!As for Roulette numbers going into receivership. That's absolutely true. But they don't need to. The scenario of effective bankruptcy is where there are several draws in a row where the house loses. You can give the house starting capital and let it - uh - roll. The house can bankrupt under Roulette conditions even with positive expectancy. This is the same as saying that re-bal can lose you tonnes of money relative to buy-hold. It can and will happen. Casinos will go bust under certain scenarios. Re-bal will lose (a lot of) money under certain scenarios.
3. Where did you get this from?
Good for you ! Were you trading for a living with your own capital (and no other income) at that time?Seemed to work alright in my lifetime. And that included the periods like LTCM, Russia/Asia Crisis, Tech Wreck, Iraq and GFC.
Well it's a good thing that you're not reliant upon the accuracy of your guesses for income then, isn't it?!!I would guess, because I cannot ever know for sure when we're talking Probability Theory, that you have probably no real concept of convergence and how fast it occurs and how it can be further stabilized.
Well what was the point of that!4. Yes they do. No argument here. Please see Post #73, Item 4. Where I ran re-bal for portfolios that were composed entirely of stocks that went down.
It demonstrated that rebal has positive expectancy under these conditions. Re-bal doesn't care if stocks are going up or down. All that matters is how they move relative to one another. The underlying idea behind this re-bal stuff comes directly from Probability Theory and from that, the positive expectancy just falls out.
Much of this debate arises because people can't work through the maths or, failing that, won't take the time to run a basic Monte Carlo simulation to get a feel for it.
Good to hear! Enjoy your retirement!I'm done with this debate. Please just do the work to figure it out. I'm not here to be your pillar of abuse or punching bag. I could care less.
Sorry TPI, I signed off too soon.
1. I think you are looking to low turnover strategies.
If that is so, then Premium Based Strategies are suitable candidates. They are long term in focus an involve some notion of risk bearing that can't otherwise be absorbed by others in the market. Things like deep value which is screened for credit worthiness, buying small caps. You may have heard of Low Vol. That's achievable but more intense and needs valuation and earnings quality overrides. You can add trend stuff into the mix, which works very powerfully in Australia for some reason, but no where near as powerfully elsewhere. All that is simple.
...
Otherwise, you get rich by not losing money. Watch for tax. Knowing your tax rate, you should either tilt to or away from dividend payers. Watch comm. There is massive excess turnover. Just doing nothing is often better than watching screens. Harvest the right parcels and know the consequences of harvesting a large CGT parcel requires a very powerful idea to justify it. This really adds up over time.
You sound like an adviser or planner. You would be aware of the horrible record of mistiming markets in retail (let along insto). So a huge part of this is simply sticking to the game plan. But, it's hard because you don't know if your game plan is busted. To me, you invest in a way that allows for really bad scenarios but not the extreme ones. When that's set, you find a way of protecting tail risks so you can hack it in the event of really awful outcomes and stick to the game plan and stay cool as the world melts around you.
That's the simple truth of it.
2. Actually, you can build a portfolio with an income focus and add re-bal on top.
Hey RY,
I am really enjoying your contributions to this forum, you have already given me enough ideas to research and play around with for a long time. Thank you.
If I could please keep you signed in to this thread for a little bit longer..
What are your thoughts on black swan events? A system may have a positive expectancy, but a significant enough chance of a total meltdown. And the longer you run the system, the greater chance of running into one eventually.
Australian market is an interesting example at the moment. A lot of value strategies, at the moment would be almost totally invested in mining services companies. Chance of a catastrophic event for a portfolio is thus significantly higher, as it only needs to happen in one industry.
What are your thoughts on controlling systems to lower this risk, what kind of manual overrides to allow, and how to not let those overrides be abused?
While I wont change my tune about the dangers of rebal on stock portfolios, I have done a little preliminary research on an area where rebal should ALWAYS work.
This being in the field of ETFs. Taking any of our long term well managed ETFs, say AFI, MLT, WAM, STW etc
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?