R2G vs private port

Manage
41

“I like to think some of the reedy to go models are also develop with idea and not just test what work (For example my Sherman’s why to go shield wall where I try to make alpha with low risk)”

MMy point is that you chose what worked best in the past (TLT). This may not say anything for the future.

Steve
[/quote]

Sure I did choose what works best in the simulation!

Just to have an idea, and offering it to other people is in my opinion to “throw caution to the wind”
You need to test it and polish it to see if your idea really working, same you think it will
And this what I am doing

For me, safety comes from the margin of safety (from buying cheap stocks)
And from thinking a little different

If i will think and do like everyone else, you will not make more then everyone else

High liquid stocks have many eyes on them, that why the margin of safety is lower
And the performance is lower also (in simple words, it is hard to find liquid and cheap stocks)
That why I like the small cap stocks( I not trade OTC)

And let’s not forget the reword for that risk you talking about
Some people (like me) will take the risk of losing a position, because in a model that can make 100% AR
Even losing 20% is better than a model that make only 30% AR

It is much better…and yes, it is not for everyone

Amiran

42

Let’s recall the title of this thread: “R2G vs Private Port”

This approach is perfectly fine for a private port. You know who you are. You know your own financial circumstances. And you are in full control re: how diligent you will be in terms of stock selection (simply follow the model, or look at each company in the result set individually to confirm that they really do look OK) and trading (knowing that you won’t try to trade more than the market can comfortably accommodate).

But there’s a huge difference between a private model and a public model. With the latter, you don’t really know who is acting upon the model recommendations and how they are doing it. When R2G was initially being developed, I proposed limiting it to stocks with market caps of $500 million and above, based on the principle that private models and public models need to be approached differently. That idea was not accepted.

I’m active in the low-liquidity area for myself and the newsletter I publish. But I don’t just run a model, fool myself into thinking the sim results are real, and disseminate the lists of passing stocks. I do full due diligence – right now my “Reserve” research folder comprises 44 separate company files each of which contains comprehensive due-diligence information – and my archive consisting of companies less likely than in the past to appear in the model is much bigger. I also maintain a set of exclusion lists consisting of (221 – at present) tickers of firms that made it into the model but failed my due diligence to the point where I want to disqualify them from appearing in models in the future. Bottom line: I won’t put a penny of my own money into this area of the market nor will I write anything about a stock that doesn’t pass my due diligence – and the fact that a stock appears in a list produced by a model that backtested well does not count as part of the due diligence. In fact, my 221 deadly dogs couldn’t have even got on the no-no list without having come to my attention as a result of having appeared in a model that tested great.

Is anybody offering a low-liquidity R2G model doing anything like that? (And by the way, even this can’t give you 100% protection from bad outcomes; all it can do is reduce them to a more tolerable level.) Actually, nobody offering any R2G model is even allowed to ban known dogs because in order to get “proper” sims, lists cannot be used. In terms of due diligence and risk control, a non R2G model can go important places where an public R2G model can’t follow. And as to trading of these low-liquidity R2G models and all the angst and discussion and studies seen in the forums since R2G got going, my only response is: If you use your jaw to try to punch Mr. Market in the fist (which is pretty much what’s being done when low-liquidity trading is removed from the private realm and placed into the public I-can’t-control-who-is-doing-what arena), don’t complain when your jaw hurts.

There’s a lot of great stuff that can be done with private models, which is pretty much what everybody on p123 was doing before R2G debuted in early 2013. But the public arena is a very different place, something that still seems to be poorly appreciated.

43

Steve,

I don’t think that P123 has a way to differentiate between the fraud companies that cease trading and companies that just have 0 volume for a while, or even those that are acquired until some time later when the data becomes available. So how is a Sim adjusted on the fly to account for fraud companies? I agree it would be best to do it if possible but How? Also, how often does a Sim buy a fraud company, and how much difference does it really make in a 14 year Sim?

Denny :sunglasses:

44

Steve and Denny,

The media notwithstanding, out-and-out fraud may be the least of the problems (remember the journalistic adage about dog bites man; we hear about it because it is so unusual). The more frequent scenario is getting into companies that operate legally but are just plain lousy. Numerical filtering, analysis, etc. can carry one just so far.

With larger companies, there tends to be a lot going on (even a supposedly single-industry company can, actually, have quite a lot of very different businesses) and there are the protective benefits of just-plain size. With smaller companies, there tend to be far fewer activities, meaning bad businesses (temporarily bad or structurally bad) are right out there – not shielded by who-knows-how-many-other things. Combine that with less, often much less, fixed cost coverage and you have an entirely different sort of risk profile and the ability to lose just as much in shares of law abiding companies as can occur with the less frequent but more newsworthy fraudulent firms. And those who confine themselves to five-stock portfolios are even more vulnerable.

45

Someone should create an R2G for stocks to be shorted (and publish the criteria). Many more stocks go down than go up…I might use this for my own exclusion list.

46

I’ve been with P123 for about a year and I have seen companies have problems with their reports twice (that I noticed). Of course, everyone remembers FU. But also Barnes and Noble (BKS) had to redo or amend one of its reports and if I recall it was required to do so by the SEC. I love Barnes and Noble so I held on to it. Liking Starbuck’s coffee is not a good reason to hold onto a stock it turns out.

I got bit twice in a year and even a hint of inaccuracies or fraud in a report is a reason for me to sell at market, now, as far as I am concerned. Of course, this does not mean I know how to find problems ahead of time and I suspect a lot of people already sell quickly when this occurs–I’m trying to learn.

47

Denny - The point was made that the low liquidity stocks are a huge opportunity. But how can you assess an opportunity when you can’t assess the risks? If even a couple of stocks fail in 14 years in a five stock model then there will be a huge dent in the system performance and the perceived risk. I understand that P123 can detect most mergers so why not assign a zero price for the remainder of stocks that disappear? Doing so would be a good incentive for avoiding such stocks.

And who should do the research behind the companies? People subscribe to models to tap into others’ expertise, not to perform the work (research). So it should be on the model provider’s shoulders to screen the companies and read the news daily looking for issues. When an issue is found then the model provider should send an EMAIL to subscribers warning of potential danger. And it gets more complicated when autotrading is introduced.

Amirans - what was the idea behind adding TLT? If you had said “TLT is negatively correlated in times of market stress, and at other times potentially offers better performance than a short (stock market) ETF”, then that would be an idea. But as I understood it, you chose TLT over other ETFs because it gave better backtest results, without considering the “fundamentals” which suggest that interest rates will rise in the future, hence TLT will fall.

Steve

48

Marc & Jim,

It is fairly easy to sell a stock we are holding after a hint of illegalities or fraud if we are paying attention to the stocks we hold. My point above is that I don’t see any way that we can do what Steve wanted to do. That is, have a Sim either not buy the stock or set the price to 0 after the fact. Without knowing if a stock that has 0 volume is a “good” stock or a “fraud” stock, how does a Sim make the correct judgment. Unless we can get “The Information” real time, how can it be used in a Sim?

My guess is that even if we loose 100% of the investment in a fraud stock once every few years, if we are well diversified it wont make a big difference over time. That includes the 20% loss in a 5 stock port. I HOPE no one has all their eggs in only 1, 5 stock port. If they are holding 20 stocks in 4 diversified 5 stock Ports then the loss is 5%. And I hope they are holding 50 stocks and/or other investments overall.

I just don’t see any way yet that it can be realistically handled in a Sim.

Denny :sunglasses:

49

Denny, agree. I’m not going to give up the potential for 50% or more return in some of my ports (e.g., the one that was holding FU: I did not hold until trading stopped but could have gotten out sooner). Was really commenting on the narrow issue of selling at the hint of fraud. If I hold 20 stocks, that’s just a 5% loss in a day, at worse. I’ve seen worse days without a stock going bankrupt.

50

If someone thinks that the size of a company or the nationality is a guarantee to honesty
They have learn nothing form the 2008 crises

We are lucky, frauds are rare, but sure diversification is always welcome especially is a system that working statistically

Marc,
[color=darkblue]“With larger companies, there tends to be a lot going on (even a supposedly single-industry company can, actually, have quite a lot of very different businesses) and there are the protective benefits of just-plain size”[/color]
There is also one big disadvantage to size
For example, imagine what one new and good medicine will do to a small size Drug Manufacturers company stock price?
Compere with the same new medicine will do to Pfizer?
(Hint in the small company will more than double its profit and the large company will add 1% to its profit)

[color=darkblue]“When R2G was initially being developed, I proposed limiting it to stocks with market caps of $500 million and above, based on the principle that private models and public models need to be approached differently. That idea was not accepted.”[/color]

When I offer a small cap model I look at it almost as a privet portfolio
I limit the number of subscribers, to only a few and also keep min liquidity

I was very surprise that with an opinion like you have, you also offer a small cap R2G portfolio
With a min liquidity of only 80000$ and that buy stocks below 500$ million and offer it to 98 Subscribers, yes 98!!

Is someone put a gun to your head and make you do something so irresponsible, not only in my opinion but also in your opinion?

Steve
Yes, you got it wrong
Treasury bond are not just about interest rates,they also a safe place to be when the markets uncertainty rise
And the “shield wall” model is not only a bonds model, the stocks part do most of the performance

And sure I will trust a simulation of 14 years more than the “fundamentals “you are talking about, because if you know the fed will move up the interest rate tomorrow

Why you not go short with your entire portfolio today?

Amiran

51

“If someone thinks that the size of a company or the nationality is a guarantee to honesty. They have learn nothing form the 2008 crises”

Amirans - Enron is a good example of a large American company that committed fraud. Another example was Bre-X back in the 1990s. We can’t avoid all fraud. But there is a saying “Don’t let the things you have no control over stop you from doing the things you do (have control over)”.

Everyone - It is plain wrong to suggest that anyone can get out of these stocks at the first whiff of illegal activity. I just did a search of P123 and here are a list of quotes from different people. These are all intelligent people and many of this comes from the people on this thread.

“You can add FU to the list. I just got killed with it.”
https://www.portfolio123.com/mvnforum/viewthread_thread,6686#32542

“My experience was LPH which turned out to be a total fraud, and caused me to lose my entire investment in the stock.”
https://www.portfolio123.com/mvnforum/viewthread_thread,6686#32542

“I got stuck by LPH also, cost me a little over 10k.”
https://www.portfolio123.com/mvnforum/viewthread_thread,6686#32542

“I’m in FU too. Only sold part of my holdings when a sell was triggered in one of my ports Friday. Run up was spectacular. Remains to be seen where it goes from here.”
https://www.portfolio123.com/mvnforum/viewthread_thread,6686#32542

“I learned my lesson after taking a hit on LPH.”
https://www.portfolio123.com/mvnforum/viewthread_thread,6011#28579

“You’re not alone, I was on LPH too when it fell. Fortunately I was on my computer that day and was able to sell at -50% before it was halted…”
https://www.portfolio123.com/mvnforum/viewthread_thread,6011#28579

"I also owned LPH and lost 43% due to a 30% trailing stop loss I had on the stock. "
https://www.portfolio123.com/mvnforum/viewthread_thread,6011#28579

“For a long time, I rejected anything with headquarters in China due to the RTO fraud issues. Recently allowed them back in, purchased NEP, and that’s been halted since 3/1. Fool me once shame on you, fool me twice shame on me.”
https://www.portfolio123.com/mvnforum/viewthread_thread,5273#25055

“Right now ChinaCast (CAST) is halted although the issue here is a board fight. I am also done dealing with Chinese companies.”
https://www.portfolio123.com/mvnforum/viewthread_thread,5273#25055

“And it’s not just a few bad apples. At least 105 Chinese companies listed in the United States have been delisted, are under investigation or have financial problems, according to The Pittsburgh Tribune-Review.”
https://www.portfolio123.com/mvnforum/viewthread_thread,5273#25055

"FUQI changed its ticker last week to FUQI.PK Duoyuan Printing(DYP), a Chinese reverse-merger company under investigation by the Securities and Exchange Commission for fraud, said the New York Stock Exchange has decided to delist its stock as of April 4. "
https://www.portfolio123.com/mvnforum/viewthread_thread,4927#23082

“The level of fraud in China is much worse than in the US. Even Enron really had employees. On the other hand, many Chinese companies have no business and no employees but just make up numbers.”
https://www.portfolio123.com/mvnforum/viewthread_thread,4863#22856

So this is very convenient for model developers. Create the riskiest models possible and when a stock disappears, it doesn’t impact the stats. As I said earlier, the onus is on model developers to warn subscribers of dubious stocks. Do not tell subscribers “It was in the news so you should have known to sell it”.

Denny - there is absolutely no reason why this can’t be handled.

To quote Marco: “This is one of the few situations left in the sim where returns are over estimated (under estimation is always better). They reality is that a 99% loss should occur in that position. Any mechanism to avoid this loss is incorrectly recreating the past.” https://www.portfolio123.com/mvnforum/viewthread_thread,6011#28579

“And sure I will trust a simulation of 14 years more than the “fundamentals “you are talking about, because if you know the fed will move up the interest rate tomorrow…”

Amirans - I got it right.

I think you are missing the point that I was trying to make. I am not disagreeing with the “idea” as you stole it from me!

I have been saying all along to go with what works. But Marc has raised some (legitimate) objections. What burns me is when model developers indicate that their models are traceable back to fundamentals to get brownie points (appreciation) from subscribers. Use of TLT is a case in point. You are not tracing it back to some fundamental idea, at least until now. You were just saying it is an “idea”, and making reference to backtest results.

As for shorting a port, we don’t know when interest rates will begin to rise. But we do know that they can’t fall too far from here. I would prefer to hedge my stocks as this is safer and doesn’t require perfect prediction of the future.

Steve

52

Not when trading is halted it isn’t. Sometimes there are hints ahead of time and sometimes not. Sometimes hints simply create value by driving the price of a good company down, and sometimes the company wasn’t so good.

I’m not sure why this is even being discussed. If the stock has 0 volume then obviously it can’t be traded and P123 should not record a buy or sell on that date. What value to assign the stock is a little more difficult, but certainly not impossible. These are simulations after all and a perfect solution is not required, but I would like to see some improvement over recording a sell at the last traded price on a stock that essentially went to 0. For a sim this really shouldn’t be so difficult, after all it is known what price the stock traded at when trading resumed. This is one time looking ahead in a backtest would make the simulation more realistic.

53

Can’t some of these be takeovers, though?

54

Plan_Trader - apparently P123 can detect mergers and acquisitions.
Steve

55

That’s an interesting idea and it would seem to work also for situations where trading is halted for good reasons such as a pending takeover announcement, or any other buy-order imbalance.

I’m not sure we could make it a universal default solution – P123 is about point in time and i don’t think we could prejudge, in a mandatory manner, when subscribers should be willing to suspend it. But I can see it as an option in test/sim settings, and of course as a required choice for situations where p123 establishes ground rules (such as for R2G sims). Assuming this would be technically feasible (that’s Marco’s call), can anybody see any downside to this solution?

56

It would be good if at least in sims, this could be “adjusted” so that positions in these stocks show as total losses instead of sold at last price like it currently is. It would be a case by case addition so probably not 100% reliable, but something is better than nothing. I would not expect the impact to be that big on decently diversified portfolios of 30+stocks though.

57

Maybe I’m missing something, but I don’t think this ‘no trading if no volume’ requires any serious ‘look ahead’ bias for ‘mark to market’. Sim’s are already using that current trading day’s price data to determine transaction prices. Just need to add a daily volume constraint. No stock held can be sold if day’s volume is lower than X. No position can be bought in such cases either. Ideally, with a user specified input field for X.

So, sim’s could just not allow a buy or sell if there is zero volume (or volume below user specified min. threshhold volume). The position and ‘cash in it’ is frozen if holding a position. And positions can’t be bought if no volume. Ideally, we would also have the option to not allow a buy or sell if the position we are seeking to sell is X% or more of the day’s trading volume…or daily volume is below X…as this may mean that slippage would be HUGE…but this makes low liquidity sim’s harder slightly to backtest and might require other features (i.e. fix the number of positions and dollar amount in each).

Then…When the volume is over some min. threshhold again (maybe better to allow a user field for min. threshhold…with default of zero), then the stock can be bought or sold again…and transaction price is figured as it is currently done. We don’t need to ever peak ahead more than 1 day (as per now) and ‘mark to market’…unless the stock never trades again and gets delisted…in which case it’s sold at zero dollars. So…if Volume has been zero for X days…the stock gets marked to zero.

Seems doable?

But…This raises one other point. For individual user sim’s to be more accurate, the variable slippage formula needs to let users take into account the position size we are trading relative to the daily trading volume. This is in addition to the liquidity of the stock. As our daily position size/total trading volume rises, so does slippage. Allowing users to input a custom 'slippage formula could be helpful. Why does it matter? Slippage assumptions in hi-turnover systems are very, very significant drivers of sim results. Dynamic slippage might reveal things that static fixed amounts aren’t showing.

58

Steve

You bring very good examples

In some of these companies I got hit myself(and survived).

There is always a risk when you invest in the stock market
But there is also a reward

You can say the risk /rewards in Chinese companies are not good for you and you exclude them

You can say the same about small market cap stacks

Or you can say it about the whole stock market.

The only risk you really control is, if you want to invest, or not.

Don’t blame Marc, for not following your ideas
He post his opinion, but you are the one who choose to accept it

Probably I was in Thailand in holiday during that time
So I did not read your post and also i not read Marc answer to you.

But I not think the interest rate will jump tomorrow, because the recovery is very slow

(Extracts from the Federal Reserve Press Release:
The unemployment rate, however, remains elevated
Inflation has been running below the Committee’s longer-run objective,
the recovery in the housing sector remained slow.)

Having said that, I am not ashamed to say many of the ideas that I am using are “stolen” from great investors

I try to implement these ideas where there are less great investors to compete with

I find this place in the small cap category, and I feel safer that I will make more money there

You have the right to think different and it is ok.

I wish you a good luck

Amiran

59

It doesn’t, but what value is to be assigned a stock that has been suspended when calculating the value for the simulated portfolio? The value is unknown but in practical terms it may be closer to the future value than the last. If so then why not just look ahead and make use of that value in determining the current daily valuation?

Don

60

Yes, you are missing something. This is not simply a no-volume situation. There’s a huge difference between a stock that doesn’t happen to trade on a particular day versus one for which trading has been formally suspended. For valuing a portfolio containing such a stock, the last trading price may be as good an assumption as any for the former situation. But for the latter, the probabilities are extremely high that the last-recorded trade my be the worst possible assumption with true value being most likely reflected by the so-far-unknowable next future trade. That’s why the imposition of look-ahead-bias may be the most reasonable solution for trading suspensions (even in the rarer cases where trading does pick up at a level close to the last pre-suspension trade, then it wuld be pretty much a no harm-no foul situation).