Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
AIM is not a religion, it is selfdefence. Against ourselves, too. I would say that a lot of the people on this board (including me) have come to AIM after finding out the hard way that their emotions, intuitions and rational assessments gave bad guidance when it came to managing their stocks. Selling out high and buying back lower is great when it works, but for me, it doesn't work often enough. Buying on the way down and selling on the way up is extremely easy to manage with some guidance from AIM.
That is why your message is singularly unconvincing to me. I am not saying you are wrong! But it didn't work for me, and just pointing out that it would work if I would make the right decisions is not enough. I need a system to remain focussed.
I am also looking out for things that will make the system better, but I am wary for enhancements that seem to me to be contrary to the spirit of AIM (expressed by the great Tom Veale as 'buy from the scared, sell to the greedy'). To me, TA seems like a method to objectify fear and greed. That is nice in its own way. But whether TA added to AIM will complement it or obstruct it, is an open question. It would need some formalization (a system) and testing.
I am easy to convince. Just show me the numbers.
Regards,
Karel
Hello whitelake, thank you for the clarification. I'll try to put it in my own words:
- TA sell: you exit the equity part of the AIM position
- AIM sell: you sell (taking vealies into account)
- AIM buy, not confirmed by TA: buy order amount is reserved
- AIM buy, confirmed by TA: you buy
- TA buy, not after a TA sell: you spend the whole order reserve
- TA buy, after a TA sell: you re-enter the equity part of the AIM position
Do I have this right? Then we might split your method into two parts: the first part concerns the TA delayed buys, the second the pure TA action you mix in.
Some remarks for a low flying goose:
- It might be difficult to keep the official account and the shadow account synchronized.
- AIM buying: you could miss the best opportunity for buying while you wait for TA confirmation. This might be no problem when the order reserve contains two or more buys, but with just one buy in the reserve, buy levels may get unattractive. That the confirmation comes rather late is illustrated in your graph, where all buys come rather late.
- TA buying and selling: this is just that, with all the risks concerned. You use four indicators, and I would suppose such an indicator to be rather slow, with all the risk of whipsaws involved. Indeed, your graph alrady shows a whipsaw: a 16% loss from a sell at 13.6 and a buy back at 15.8. Then you sell again at 15.8 and you are able to get back in a bit lower. You may be glad that TAIM did better than AIM, but this is the kind of action I don't like to see myself.
I suppose you are now testing this on more stocks and over longer periods (several years). I will be very interested in your findings.
Regards,
Karel
Hello whitelake,
What works, does work, but your graph makes me a bit uncomfortable. As AIMers, we are used to be contrarians: we sell in upmarkets and buy in downmarkets. Now if TA could help us to wait with selling and buying, that would be nice. But it would also introduce a momentum approach to the model: exactly the opposite from our contrarian approach.
You call your method technical AIM, but where is AIM? I see just the technical indicators doing their work. Could you explain a bit more what AIM does in your setup? What decisions are influenced, and how, by AIM?
Regards,
Karel
The Kelly Equation and AIM are largely incompatible, I think. The reason I think so, is a standard piece of investment advice, that IMO also applies to AIM: cut your losses and let your winners run. Standard investment advice has two things to say about gambling and investing:
- investing is part gambling, because you risk money for a greater return
- investing is not gambling, because in gambling the players are fleeced.
Just look it up on the net. According to the first take, whatever you do, how well you pick your stocks, you will always pick some bad ones. Don't hesitate to get rid of them, and try again. If you never ever pick any winners (or not enough to offset your losses), you are doing something wrong. Now letting winners run makes it very difficult to come up with a value of E, I would say. So no Kelly equation for B&H and B&H like schemes like AIM and Reverse Scale Investing.
The KE becomes applicable when you use both stop loss and stop gain schemes (and/or time limited holdings, say a month or a year or whatever). Formulate a stock picking strategy, and backtest (or forward test) your holding scheme. As soon as you find a strategy with a net gain, you look at the losers and the gainers, calculate p and E, and drop them in the KE for optimum investment. This style of investing is very uncommon, AFAIK, and the combination with the Kelly equation even more. But I might be wrong here, because both seem to occur more often in futures trading and commodities trading, where I am rather ignorant.
Now when you combine AIM with selling out at the first sell point, or Synchrovest with selling out at a fixed point, you might get a bit closer to the KE. But I think it really only applies to the stock and contract churners. You need huge numbers to make probabilities work.
Regards,
Karel
Hello Conrad, sometimes when I was plagued by circular references, it was possible to circumvent the problem by using the results one row lower, not in the same row. Sometimes this had to be an intermediate row, so the whole calculation would take two rows. Of course, in that case you could merge the rows into one, but two rows fit the screen width better.
And of course you just might be implementing a chaotic system!
Regards,
Karel
Conrad, perhaps it would be good to explain how you would translate this to investing, as talking about games where we get a 50% change on a 250% payout seems a bit unrealistic. How would you go about assessing p and E for an investment?
Regards,
Karel
When I look at the formula p - (1-p)/E, and we take p to be .5 (coin flipping), there is indeed a chance to win in this game, because of the occurrence of E. I thought you meant coin flipping, a dollar for a dollar. But apparently (and following the link) there is actually somebody who wants to bet more than one dollar against my dollar! Of course you can win at that game! The surprising part is that this seems to appear to you as something surprising (or something that should surprise us...)
The formula is intended to give an optimum amount to risk, given a certain probability and a certain return. Very well. Now I don't suppose there are a lot of cases were p is close to 1 and E is high. In any case, it is easy to see that from the formula follows: never risk a greater part from your money than the probability of winning, because (1-p)/E is always positive, under reasonable assumptions. So what we see here is that you don't invest everything, but stay partly in cash. That is a very well known form of risk management in investing; too much volatility may kill you. I don't see where this flies in the face of standard investment practice and/or wisdom.
The interesting thing is the optimization of the amount to risk. Now that is very, very nice.
But didn't you mention earlier, regarding risk avoidance, that there should be companies with p = 1 (or very close to one) and high E? Looking for such stocks seems to be a rather different strategy than the one outlined here. Or did I get you wrong?
Regards,
Karel
Hi Tom (and others), yes, I realize that I made it difficult for AIM, picking the SP500 over that period. I don't mind lower results with AIM. Of course, everybody likes the stocks were AIM licks B&H. These are a trivial case: return is higher, and because AIM has been buying when the market was down and selling when the market was up, risk/volatility was lower. AIM wins hands down.
But there is another claim made for AIM: even when you don't beat B&H with it, you get compensated for it by the lowered risk. But usually the claim is just that the risk is lower, which I think is just too fuzzy. The lower risk has to compensate for the lower returns. Otherwise, there might be better ways to lower your risk, with a higher return. That is were the Sharpe ratio is useful.
My conclusion for the SP500, which I consider a bad AIM target, is that AIM didn't make it. But for very low vealie settings, AIM came close. Considering the nature of the SP500, and the market over the period considered, I think that this is extremely encouraging. So let us have a look at the Nasdaq. I checked the Nasdaq100, from from Okt 85 through Mar 02. I set safe at 10% both ways, minimum transaction 5% of Equity value, monthly checkups, and initial split 66/33, unless the vealie level was lower. I kept the 5% Risk Free Return and the 3% return on cash. This gave, for the different vealie levels (100% = no vealies):
(Vealie%, Return, Risk, Sharpe Ratio:)
100% - 9.7% - 10.2% - 0.47
80% - 9.7% - 10.2% - 0.46
75% - 9.5% - 10.8% - 0.42
67% - 9.6% - 12.5% - 0.37
50% - 10.5% - 16.4% - 0.33
33% - 12.1% - 20.2% - 0.35
25% - 13.3% - 22.0% - 0.38
20% - 14.1% - 23.3% - 0.39
15% - 14.3% - 25.1% - 0.37
B&H - 16.7% - 30.6% - 0.38
This time, I have entered some vealie levels between 100 and 50, and one past 20, to get a better picture of what was happening there. There are two optimums for the SR: 100% (actually between 82% and 100% there is no difference), and around 20%. And the 100% optimum has a way better Sharpe than B&H. Now we are getting somewhere!
It also appears that the more risk we allow (low vealie levels) the higher the return is. But when we start at the highest level, the extra return does not compensate for the extra risk at all. This reaches a low around 50%, after which the extra returns start to gain on the extra risk. Past 20%, the extra action no longer compensates for the rising risk.
Of course, this is all highly anecdotal: we have looked at the SP500 and the N100 only. But I think the results are very interesting. Any opinions are welcome.
Regards,
Karel
Conrad, while you are working on the circular reference problem, it might be good to know that there is a way to make Excel accept circular references. Just go to Tools\Options\Calculation(?) (Extra\Opties\Berekenen) and check Iteration (Iteratie). You probably won't need to change the iteration values.
To debug a circular reference, you select the problem cell and then you choose Tools\Check(?) (Extra\Controleren) and then the option to point to source and/or target cells.
Success!
Karel
Hi Tom,
Good you asked, it made me check my spreadsheet again! I didn't test standard AIM at all, but AIM with Vealies stopping sells at cash over 25%! Back to the drawing board. I also changed my minimum transaction to 5% of PV, and allowed .25% interest/month.
The bad news is that AIM without vealies (or vealies at CA = 100% PV) is much worse off: Return 6.8%, Risk 6.4%, Sharpe Ratio .29. That is Very Bad, compared to the SR .44 of B&H. Things look up when the vealies kick in, however:
Vealies at 100%: 6.8%/6.4%, SR .29
Vealies at 50%: 7.9%/7.7%, SR .37
Vealies at 33%: 9.1%/10.0%, SR .41
Vealies at 25%: 9.7%/11.2%, SR .42
Vealies at 20%: 10.3%/12.2%, SR .43
Buy and Hold: 12.0%/15.8%, SR .44
Now the last two AIM options differ less than 5% from the B&H Sharpe, so let us call those a draw. But who would set the Vealie level at 25%, let alone 20%? Perhaps this analysis is an encouragement to do so with very stolid stocks/funds!
More bad news: my estimates for bond rate and interest make this exercise a bit dubious. If anything, I seem to have underestimated the bond rate, and higher Risk Free Return favors the higher return option; B&H in this case. (I checked RFR 6%, interest 4%/year.)
The conclusion is: AIM is at best struggling to make the reduced risk compensate for the reduced return; in this case of course. I am convinced there are better AIM candidates than the SP500!
Regards,
Karel
Hi Jonathan,
Taking the first reversal order seems like a good idea to me. After all, for a reversal order the stock price has to traverse almost the whole hold zone. Waiting for more price action is just plain greedy!
Regards,
Karel
AIM and the Sharpe Ratio
With AIM, we hope to realize two things for our investments: to reduce risk and to raise return. At least in our hopes, AIM is the Holy Grail of investing. How can we check this? Perhaps the Sharpe Ratio will give us some idea. The SR relates (excess) return to risk/volatility, and comes up with a number. A high number means relatively much return for relatively little risk. We could use this for a comparison between AIM and B&H, or for a comparison between AIMs.
To calculate the SR, we need the annualized return for an investment, its volatility, based on periodic checkups (I use monthly), and the risk free return. We leave the risk free return out for a moment.
Let us have a look at the Lichello cycle. This is a trivial case, as the B&H return is zero. Relating zero return to whatever else results in zero, so AIM wins if it can make a profit, and Lichello showed it can. But it is still interesting to look at the risk: in a spreadsheet with 40+ cycles, I calculated the Geometrical Standard Deviation to be 136% for B&H, and 89% for AIM. AIM's millions are gained with less risk. BTW, that 89% is still a lot of volatility.
Now for almost the opposite extreme: the SP500 index. The SP500 is a bit dubious as an AIM investment, as it has very little volatility. But couldn't the risk reducing factor still work wonders? B&H for the SP500 from Jan 79 through March 01 gives a return of 12.0% (I think this ignores dividends) and a volatility of 15.8%. The Return/Volatility ratio is then 0.756. A fairly BTB AIM with Safe settings of 5% both ways gives an annualized return of 9.2% and a volatility of 11.6%, wih a Return/Risk ratio of 0.792. So AIM has a better Return/Risk ratio, but that is not the same as the Sharpe Ratio. The SR uses Excess Return, defined as (annualized) Return - Risk free return. Usually, the 90 day Bond Rate is used for the RFR. I don't know that rate over this period, but 5% doesn't seem to be a bad guess. For B&H we then get (12-5)/15.8 = .44, for AIM we get (9.2-5)/11.6 = .36.
The conclusion should be that in the case of the S&P500 over the period Jan 79-Mar 01, the reduced return of the AIM version considered is not compensated by the reduced risk. You could have gotten a better return and a with AIM comparable risk by splitting your investment over the index and bonds.
So, maybe the SP500 is a bad target to AIM. Suggestions or analyses for other AIM targets are welcome. And your opinions about these calculations, of course.
Regards,
Karel
Sarmad,
Thanks, and thanks
Perhaps it is as well that there two BBs for AIM.
Regards,
Karel
That reminds me, does anybody know were Jonathan Lyons hangs out?
Regards,
Karel
Well, Synchrovest in LC's version sure looks interesting. But I tried putting in the SP500 numbers since January 1979, and the results are pathetic compared to Dollar Cost Averaging. Now the SP500 is not very AIMable, but still.
Synchrovest seems to really rock when you can use a sell stop to lock in a fat profit, in the knowledge that a lot of buys are going to follow below your sell point. The combination of the Lichello cycle and Synchrovest is a marriage made in heaven.
Regards,
Karel
Hello Conrad,
no need to be surprised. The initial investment for your test is near the top of the price range. No surprise to see someone who invests gradually into a position get a very nice result. Now try the same test for an inverted curve!
Regards,
Karel
For Don Carlson
The email address you gave in a Private Message doesn't seem to work. Please try again,
Regards,
Karel
Thank you Mark, Modern Portfolio Theory, now I remember reading about it on SI. That would have been shortly after the AIM convention. I am very interested were it brings you (and us).
My own idea has single stocks as its focus. What happens to the Sharpe ratio (or something like it) of an investment when you AIM it? I have some ideas, but need to get a bit more information, and to set up backtests.
But first, a long weekend Paris with my wife!
See you all,
Karel
Thank you Tom. This discussion on risk and return has given me the idea to check how AIM influences the Sharpe Ratio for an investment. I'll try to make it work, but that requires some study. Ideally, we would expect to see the Sharpe go up, because of AIM. Or that is how I see it now, untutored as I am myself. Or has something like this been done before?
Regards,
Karel
Everything you want to know about Risk Management and more:
http://www.contingencyanalysis.com/
I really should study that site! I found it easily enough, and then discovered that I already had it bookmarked.
Regards,
Karel
Profits result from risk:
http://www.google.nl/search?q=%22profits+result+from+risk%22&hl=en&lr=
It may be a generalization, but one not generally known.
Karel
Why I would welcome the rebuttal? Because it would increase the sum total of human knowledge, which in my book is a Good Thing.
Perhaps Risk management seeks to eliminate and/or avoid risks, but then there is more to risk and investing than Risk management. In investing we generally seek to maximize return and to minimize risk. A useful metric is the Sharpe Ratio. A high SR means a return that is relatively high compared to the risk involved. However, if an investment with a high SR has a return that is too low, and you are willing to take more risk, you could use leverage. Thereby you raise both risk and return, and the SR remains the same. Or, when the risk is too high, and a return a bit lower is still appealing, you do the opposite: invest only a part with the high risk, and the rest risk free (as far as possible), lowering both risk and return = same Sharpe Ratio.
For your information: the Sharpe ratio cannot be calculated for a totally and completely risk free investment, because it would approach infinity. Instead, economists hold that a completely risk free investment does not exist. That is difficult to prove, but easy to disprove: just show us one. Preferably with high return! (And please show it first to me by private mail, because as soon as it is known, people will start bidding it up to low return.)
Regards,
Karel
Why is the investment risk free? Next day, another scientist finds a cure that will result in a cheaper medicine, and your bubble is bust. Or an epidemic breaks out that makes all cancer research seem academic. Or ...
And such incredibly fabricated stories (100% cures, venture capitalists that make no mistakes, you name it, the investment got it) really don't prove nothing.
Except of course, that when you know everything past, present and future, it is possible to invest risk free. Now were did I leave tomorrows paper? I want to check the quotes for some stocks
Karel
Conrad, please!
I respond to a message that, according to IHub, was written in response to my message 1432. It doesn't show from the content, I am afraid. I just explained what risk means in investment terms, and you go on talking about bombs ticking on planes. This is totally beside the point, as in your case you only look at the people on the unfortunate side. Assessing risk is not done by supposing that everything will go bust, but by assessing how much on average will go bust. On that you build your expectations, and on that you assess your risk.
But you did react on Jonathan Lyons in the first place. Now, to extend your message to Jonathan (1392): do you really suggest that your bomb on plane argument is an effective rejoinder to his Classical economics holds that profit is the reward for taking risk? I am afraid it just shows your ignorance, and you magnify the insult by just butting into a discussion, and going off on a tangent in full heat.
Small wonder that Jonathan was mightily pissed off, and I haven't seen you apologize yet. It would be a nice thing to do. Of course, a formal rebuttal of the risk vs. reward thesis would also be welcome. It might win you a Nobel prize (if you can leave the bombs out). But after your apologies to Jonathan please.
Karel
Conrad,
all that heat spent on a semantic misunderstanding. 'Risk' in the context of investing doesn't mean the same as risk at the gambling table or in the snake pit. Its meaning is closer to volatility. High risk investments are fluctuating so much that short term profits are hard to get. Over the long term the fluctuations cancel out, and then the high risk investments prove to be the high return investments. That's all.
Karel
About copy and paste from Notepad not working: what about putting the pasted stuff between whose nifty pre-tags?
Regards,
Karel
Portfolio Control or Portfolio Set Point? From your description, it is clear that the meaning of the term Set Point, as used in industrial control systems, covers the function of the Portfolio Control. But is that reason enough to change the name? I am always wary of that: it causes no end of confusion.
To counter your proposal, I would just say that Control as one of its meanings has: Standard of comparison. So we might say that the Set Point is the Control of an industrial control system.
And another point, on the Vortex method: you talk about Vortex buying more before the reference point, and less after it, as something that could be a feature or a flaw. A more important aspect might be, that a Vortex setup that wants to be comparable with AIM will have a scaling factor <1. This means that the PC drops with each buy, and rises with each sell, because not all the losses are compensated with the new buy (or not all gains sold off), and PC always equals Share Value after the last transaction. How do we like the movements of the Vortex PC?
Regards,
Karel
Conrad, in message 1214 you wrote: The important thing I wanted to convey was that my method is not substantially different than the Licehllo Method in its mathematical form.
If you now claim that that was not the focus of the discussion, I am only too ready to agree, because
a) it is not correct
b) it detracts from the value of the Vortex method
In fact the differences you note between the methods are caused precisely by the mathematical differences, which therefore have to be substantial, otherwise we would see no differences. A simple graph illustrates the differences, and explains my statement that 'your' S is a scaling factor and Lichello's s(afe) is a translation factor. I take note of you intention that you want the advice (order?) to be the same at the border of the hold zone/dead zone.
V L / L V
V L / L V
* / *
L V / V L
L V / V L
--------------L-----V-----L--------------
L / L
L / L
L
/
I just *love* blinking ads! :)
Karel
A bit boring, that error in my demonstration. The multiplications towards the end should have been divisions. Let's try again and be more direct:
[1] S * (pc-y) (Conrad)
[2] pc - y * (1+s) (Lichello).
We consider when both formula's are balanced, i.e. are zero. Then we get:
[3] pc - y = 0 (S<>0) (from [1]), and
[4] pc - (y * (1+s) = 0 (from [2]).
Combining [3] and [4] gives:
[5] y = y * (1+s) and throwing y out gives
[6] 1 = 1 + s
Which can only be true when
[7] s = 0
Conclusion: Conrad's method is mathematically equivalent to Lichello's method without Safe. Conrad's S is a scaling factor, Lichello's (1+s) is a translation factor.
Regards,
Karel
I have followed this discussion with some interest, as I find the Vortex method intriguing. Barry has shown a mathematical difference between the methods that is real. Conrad claims the methods are mathematically equivalent, but they are only equivalent in intent.
Take the two equations [1] S * (pc-y) (Conrad) and [2] pc - y * (1+s) (Lichello). It is clear that [1] can only be negative with y > pc, while [2] can still be negative when y < pc, or, put differently, only becomes negative when y > pc * (1+s). Obviously, a harmonization of these cases would require the equality pc = pc * (1+s), which requires s=0, or zero safe.
But more interesting than the differences would be the respective behaviors, IMO! Perhaps we could study that?
Regards,
karel
Frankly, this is a rather tweaked version of AIM, especially that Beta booster thingy. It might very well that not AIM itself is responsible for what you notice, but the beta booster tweak. Just run the test again without it.
I supposed you were testing plain AIM...
Regards,
Karel
Hello banjanxed,
I think your method of speeding up buys and sells only works when the stock trades in a narrow band. When the band widens, perhaps you would prefer slowing the buys and sells! Not being able to see into the future, AIM prefers neither speeding up nor slowing down. (Actually, AIM may be buying a bit too fast, but that is another point.) You may disagree, of course.
A second point: it is easy to see with perfect hindsight that you could have done better than AIM. But would you have taken your profits when AIM did, all on your own? And would you have started to buy into the stock when AIM did? If you are certain you can do better than AIM on your own, why use AIM?
A third point: from your figures I deduce that you use some kind of fixed period checkup, about monthly or so. AIM buys more stocks at big jumps! When you take a look at your figures, you'll see that the big orders follow big price jumps. Seems about right to me! (To be precise, this is more clear at the sell side. At the buy side, the relation is a bit more complicated and the second buy after a big jump can still be bigger than you would expect. Some call it a flaw, others call it a feature: 'pumping the brakes'.)
Regards,
Karel
Hoi Conrad,
Nice link. Nothing new for us, but interesting indeed that AIM gets coverage. I don't think that happens very much outside the AIM "network".
Regards,
Karel
Hello XT, when I had posted my message, I realized that it doesn't just apply to AIM without vealies, it also applies to AIM with vealies. Vealies don't prohibit all sells, they just keep Cash Value below a ratio compared to the Stock Value. My point remains: don't sell a winner unless it changes its characteristics and no more fluctuations occur.
But it is possible to backtest this if you would like to: Just AIM the downswings of mr. Lichello's example (or another cycle, or historical stock prices) and stay in cash when the position generates a sell over the first buy point. Buy back when the cycle on the way down passes the initial buy point again and start a new AIM position.
The reason I propose this, and not to string only downswings together, is that part of the swings are caused by general market conditions. So this way you stay on the safe, conservative side.
Regards,
Karel
I think Lichello rules out selling a winner by not talking about it
But he also gives as a rule of thumb "It's my job to make my portfolio fluctuate." So that could be a reason to sell a winner: when it no longer wins, but has gotten into the doldrums. Then the stock no longer fluctuates, and you are free to look for another.
And of course AIM wants us to sell winners automatically. It sells winners all the time. If you don't do vealies, a winner generates a heap of cash that doesn't fluctuate much. Why not convert part of it into other attractive positions.
But don't just cash out of your winners. Count your blessings! They come in at every sell point!
Regards,
Karel
Well, about Ted Tesser, with the current state of search engines on the web, it isn't difficult to find http://www.taxtrader.com/
Regards,
Karel
Perhaps, but I will stay with stocks and index trackers. I see some scared little mice scurrying below right now. When they get too far from their little burrows, I'll take a dive!!!
Regards,
Karel
It is a fascinating idea, but here is an article that shows what can go wrong:
http://moneycentral.msn.com/articles/invest/funds/8664.asp?special=msn
Regards,
Karel
Jonathan, what about this calculation?
Assume Cash is at 50% of Portfolio Value, and the minimum buy is 5% of Share Value. Assume further that you use 10% Safe both sides. If you have one buy and sale a year, 5% of your cash has made about 35% in that year. So you can safely add .7% to your cash returns for each round trip/year in this situation.
Now assume Cash is at 25% of Portfolio Value, and the minimum buy is still 5% of Share Value. If you now have one buy and sale a year, 15% of your cash has made about 35% in that year. So you can safely add 2% to your cash returns for each round trip/year in this case.
Regards,
Karel