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.
Hello Robo,
Yes, it's very clear now, and I am with you as regards the elegance of the EMA. I myself like the EMA better than the SMA. (And it's soo easy to put it into a spreadsheet, and with variable 'periods' too!)
I agree that Myst has turned the SMA weakness into a strength: it is precisely the lag that generates the trading opportunities. That is also why tuning is very important, as is the correct choice of channelling stocks. All this has made me lose interest, but it is a very nice system. My only doubts concern the possibility to get the backtested returns in practice. Up till now, we have mostly/only seen optimized backtests.
Regards,
Karel
Robo,
Yes, it will be interesting to wait for the ripples of Myst's epiphany to spread.
I think I understand what you are trying to do, but I just don't seem to see what you see. You see a 16 day EMA that has the same lag as the 12 day SMA, I see a 16 day EMA that reacts faster than the SMA, and has less extreme extremes. Perhaps if you would look at the 6 month version of your chart, you'ld see that the EMA actually runs before the SMA, and not with it or after it. I agree that the effect you are looking for is visible around the September dip, but it disappears quickly.
Regards,
Karel
Robo,
No need to apologize for being unclear, for I never thought that that was what you meant! And your example chart isn't very enlightening, due to the rather compressed scale, don't you think? If you enter '6 months' instead of 'Fill the chart', you miss the Big Dipper, and the effects I was talking about become visible:
- The 16 day EMA reacts faster than the 12 day SMA
- The 16 day EMA is already much more 'average' in its overall movement than the 12 day SMA. (I looked a bit further and found that this is already true for the 12 day EMA for QLGC. This may prove to be the biggest problem for the EMA.)
Oh well, why not give the URL?
http://stockcharts.com/def/servlet/SC.web?c=QLGC,uu[w,a]daolyyay[dc][pb12!c16][vc60]<i&pref=G
I have no opinion on whether this is good or bad for X_DEV or one of its clones; it's just that an EMA is a rather different beast than a SMA.
Regards,
Karel
Actually, you could do several things.
(BTW, I have a Dutch version of Excel 2000, so the names of options might be a bit off.)
1. Easiest is to right click on the upper axis (when the mouse hovers over the axis, you'll see the popup hint 'Category axis': then you do the right click). Select the option Format axis and open the tab Patterns. Select by Tick(?) labels(!) the option High, and click OK.
2. Almost as easy: Open the same dialog for the left axis (Value axis), open the tab Scale and check 'Category axis crosses? at:' and enter -110 (the lowest value in your graph). Click OK.
Regards,
Karel
Increase the EMA time span used, to the point that it's now decreased sensitivity most closely mimics the movements and deviations of the 12 day SMA. (TADA!!!)
No, no TADA here!! Once you have an EMA that moves more or less with the a 12 day SMA, you get somewhere in the 50 day range, and the overall movement is reduced a lot, compared to the SMA. And if you just search for a minimal overal difference, the EMA would react too fast.
Try it on
http://stockcharts.com/def/servlet/SC.web?c=QLGC,uu[h,a]daolyyay[pb12!c40][vc60]<i&pref=G
(Perhaps you'll need to reconstruct the link with cutting and pasting.)
The red 40 day EMA already reacts faster than the blue 12 day SMA.
Regards,
Karel
Hi karw,
I think Jibes deserves full credit for coming up with AIM ReBal. It is certainly not against the spirit of AIM as described by Lichello, as Lichello's words (about replacing stocks whenever you like, as long as the Stock Value remains equal) indicate.
And it seems to work, too!
Regards,
Karel
Hi Jibes,
You have every right to be encouraged by what you see. I am impressed. BTW, how do you pick these portfolios? Dart throwing? Trying until you find a nice one?
Regards,
Karel
Hi Myst,
your program looks impressive, the fund version also.
Now we come the But part: I see sells and buys coming exactly in time. Is it possible to time transactions in a fund in this way? I am not asking this because I want to use UOPIX, as a non-resident alien I don't seem to able to apply. Might be a blessing, too!
The problem is the crystal ball effect (again). I trust the stock version uses the predicted buy and sell levels. But I understand there is an added uncertainty with fund transactions, as you have to enter them when you don't now the price. Or so I understand.
Regards,
Karel
Hello Jibes,
I thought I had a theoretical problem with AIM ReBal. But as long as you dig up example portfolios that do well, my theoretical problem was having a problem. So I looked twice, and found that AIM ReBal is very nice, theoretically!!
What is the problem? When you AIM a basket of stocks, AIM likes a volatile basket, preferably built from positively correlated stocks. Rebalancing, on the other hand, thrives on negatively correlated or at least uncorrelated stocks. Big problem! Or not? Actually, we never get perfectly correlated stocks, either positively or negatively. So part of the action in a portfolio is positively correlated, the other part uncorrelated or negatively correlated. When all stocks go down or up, The AIM part of AIM ReBal gets a chance. OK, so far so good. On the other hand, when the action in your basket cancels out, AIM advises to lay back. AIMers that would have AIMed the stocks individually might see plenty action however. AIM ReBal is saved by the ReBal part, because it sells winners to bolster up the losers, exactly what AIMing the individual stocks would have accomplished. Of course, you have seen this all the time. It's just that I have now worked myself towards seeing too!
I think this is a very interesting way of AIMing a basket of stocks, and one of the more inspiring AIM variants I have seen. I see you have started your own thread on AIM ReBal, but certainly, the discussion wouldn't be out of place here!
Thanks for sharing your idea!
Karel
karw: Then Lichello suggests to ReBal your portfolio having approximately 50% of your portfolio in quality stocks.
But that isn't the same as rebalancing! Lichello is here in a Q&A section, and the question was: "How can we keep the dollar amounts equally invested when we have more than one stock in the portfolio?" Now that is the idea of rebalancing, and Lichello answers: "It really isn't important." He continues by saying that he may have suggested it, but that the questioner probably refers to his idea of keeping 50% of your portfolio in blue chip/quality stocks. That 50% is not written in stone; it could be 60% or 70% if you like [L doesn't mention 40%...]. But Lichello likes real quality stocks, because they won't go down the drain easily. Again, if anything, he encourages us to ignore rebalancing!
The only 'rebalancing' I see here is when the other part of your portfolio outgrows the blue chip part: Lichello might have applauded shifting part of it to the blue chip part. But never the opposite way! So it is not true rebalancing, after all.
Regards,
Karel
No, I didn't search for negatively correlated stocks. I am currently not even AIMing, having only a pittance to invest with, but I do some AIM inspired/scale trading like stuff with ETFs.
Regards,
Karel
In a way, Robert Lichello was a fund manager. He generally considered buys and sells based on his total portfolio value, and a basket of stocks is bound to be less volatile than individual stocks. (He tried to make his baskets not too static.) Now us hasty youngsters, having discovered that AIM works very well with volatile stocks, seek the most volatile stocks we can find, and AIM them not in a basket, but each stock on its own. We should not be surprised to find ourselves out of cash once in a while!
One way around this is Tome Veale's 'borrowing from Peter to give to Paul' principle. (Or was that the other way around?) When a AIM position runs out of cash (and still looks OK) let the cash go 'negative' = borrow from the cash reserve of another stock. This should work especially well with negatively correlated stocks!
Regards,
Karel
Hi Robert,
It's hard to predict the effect of the update period. In your example, the shorter period is more effective, but turn the numbers around, and it is less effective. With a price of $15 after two weeks, AIM will sell, and with $20 after a month, AIM will sell again. But now the monthly updated AIM will have the advantage. What to do?
There are two schools of thought on update periods: fixed update periods, and 'grab the first buy/sell point'.
With the fixed updates, shorter update periods might capture more action, but they also might start the action too soon. When you use weekly or biweekly updates, it would seem a good idea to lengthen the period to at least a month in a string of transactions in the same direction. This might be very important for buys, as cash is a limited resource in AIM. AIM will never sell you out of a position completely, but it is possible to exhaust your cash.
Of course buy-sell sequences can't come fast enough, and that's were the GTC folks come in. They place Good Till Cancelled orders at the buy and sell point, and lay back. Repeat when one of the orders triggers. For the same reason as above, some people prefer to enter a new buy order after a buy (and a sell after a sell) after some time has expired, say a month. (The order in the opposite direction is of course entered immediately!) Again, spacing out is especially important for buys, because of the limited cash. And a very small minority does enter a buy order directly after a buy order (or a sell after a sell), but, say, 30% lower than the calculated buy point. Every day they raise the target by 1%, and after a month they keep it at the normal buy point. Fun when it works, but it is a lot of work, and it is not guaranteed to give superior results either.
And the bottom line remains: it is hard to predict the effect of all this tinkering. Standard AIM is probably good enough for a lot of cases. When you notice that you are missing opportunities, shorten the period, ans when you notice that you are reacting too fast on the action, slow down.
DISCLOSURE: I am a GTC guy.
I hope this helps, but I'll settle for a spinning head
Have fun (and profits!) AIMing!
Regards,
Karel
Perhaps he could check with the company what is going on and how to avoid it, if possible. Opening a bank account probably is a good idea, as karw already remarked. We Dutch are even less check minded than credit card minded. Our plastic money comes close to a debit card, and is linked directly to your bank account.
Regards,
Karel
Myst, you say:
X_DEV is designed to capitalize on a stocks weakness and strength by selling at average extreme highs and buying at average extreme lows within it's trend. Therefore, it can not have a "losing trade". We might be able to have an argument over this if a restriction to a certain time frame was a critical component to your case, but I don't think it is.
I don't think this is correct, and that X_DEV can have losing trades, and moreover should be able to have them. When a stock drops out of a range and starts trading in a lower range, that may be the only way to raise cash for new buys. I have heard this mentioned as a feature over AIM, that insists on selling at a sell price over the last buy price.
Now when this feauture hits you repeatedly, you get hurt badly. This is not to blame X_DEV: you might have picked the wrong stock, or didn't get the bands right. And every once in a while you are bound to get the bands wrong.
Regards,
Karel
feedback appreciated...
Great looking window! (You knew that, didn't you?)
Minor quip: the goofy Start icon.
And the & in Standard & Poors gets interpreted as
underscore next char (like an access key).
Regards,
Karel
Hi,
Try right-clicking on those links. You might see an option to save the target of the link; that's what you want.
Regards,
Karel
Hello,
to further sleep deprivation
X_DEV is seeing a SMA with a constant value in your example. That is a bit thick. Also, you have a daily update, not a monthly as in the backtest. What about using the (co)sine for the daily prices? A 'Lichello cycle' takes 6 months, so you could use the following formula for the prices:
=3*(COS((ROW()-1)*2*PI()/130))+7
(ROW()-1) is for a times series starting at Row 1, adjust for series starting lower.
Have fun,
Karel
No, I didn't try the demo. As I said, they target experienced investors, and they convinced me rather quickly that they didn't target me. My real problem was the impossibility (then) to buy odd lots, or so I learned, and no interest on the cash reserve (for my low levels of everything). Now LowTrades even has a money market sweep! Boy, do I feel lucky! I can't wait till the interest rates are starting the hike upward!
Regards,
Karel
Well, IB do their best to discourage inexperienced traders, and they succeeded in my case
All those things like: 'you need to consider the trade rules of the exchanges your order may be routed to' don't sound like much fun. I understood, when I visited the Motley Fool, that you may have trouble buying odd lots, precisely because of this. But that was some time ago, and I don't know the current situation. And it may not bother you!
Regards,
Karel@holding2NVRforaprofit.com
Rien:
My trading costs are about $40 per transaction. It is one of the disadvantages of using a dutch broker for US stocks.
Why don't you get an American broker? At LowTrades (OK, OK, very, very bare bones) I pay only $4... But then, my positions are ridiculously small, so I really need it. (And even then they are too small to AIM.)
Regards,
Karel
Rien, I sincerely hope there are no (former) Enron employees on this board, for they will not enjoy your playful destruction of the concept of market capitalization. You do have a point: prices were inflated. There was a bubble. We can shake our heads at it now. But lots of people lost money. And all those losses add up to something in the 2T ballpark. Not more, but certainly very little less, as turnover in bubble stocks usually is very fast.
But I have a new office, especially for you. Whenever the price of a stock rises, people will have to ask your permission to raise the market cap. Then you can have a look at the PE and you say Go ahead! or No way! This way, a lot of trouble could have been avoided.
Karel
No, of course you didn't, and neither did I! Now, how long can we keep this ball up in the air...
Karel
Great, Lou, I hope you will have great time, for business and otherwise! You are just in time for one of the icons of our country: the tulips!
http://www.keukenhof.nl (a showcase)
http://www.holland.com (official site of the Netherlands Board of Tourism)
Have fun!
Karel
Hi Rien, I am not considering myself very sophisticated. I asked a question on risk management on another board and got referred to the stuff a (responsible) hedge fund manager needs to handle risk. My Sharpe Ratio pales in comparison!
Of course, when something works for you, by all means follow it through. Stop loss is one good way to manage risk.
Regards,
Karel
Cool! You're building up the suspense! Is it allowed to say that I always immediately click through to the newsletter?
Thanks for all you work, it is a real inspiration,
Karel
That depends on your definition of risk. In my view the risk between LTBH and AIM is equal.
Well, I think the fact that I went into very great detail in my description of how I measured risk, proves I am aware of that. And I am very suprised to hear you say that the risk between LTBH and AIM is equal, as anyone with Financial Risk Calculation 101 will disagree (and be able to point out why). You seem to view risk only as: is there a chance that you lose all your money? With both LTBH and AIM, the anwswer is 'yes', and therefore you say that the risk is equal. Of, course, more things can happen to your money than losing everything, and this should be included in a risk measurement. To give a very basic example: suppose the stock doesn't fall to 0 but to 10%, which of the two strategies has you exposed to more risk?
(And yes, LD-AIM also reduces risk.)
BTW, I am not an adherent of the 'with AIM you have only part of your money at risk' school. I quite agree with you there. Indeed, it was one of my points in a private email to Tom. His reply made me understand that his ROCAR was not so much meant as a risk measurement, but as an effectivity measurement. Then ROCAR makes sense. I think my reply to his explanation has found its way to the "In Depth" board (yes, it has: http://www.investorshub.com/boards/read_msg.asp?message_id=291224 )
Have fun, and don't oversimplify
Karel
Case A: Sell $250, gain: $35.99, ROIM = 0.72%, ROMAR = 0.36%
Hum ho. R(eturn) O(n) I(nvested) M(oney) = .72%?
Now Case A was were you invested 50% of your money. Shouldn't the above then read: Realized gains instead of Return? I hope that when the price rose of the stocks that you then decided to sell, the inventory that you held also gained about the same amount
Generally speaking, the more sell points your investment passes, the farther Standard AIM will get ahead, because it has more stocks with a bigger and bigger profit.
Regards,
Karel
Hello Rien, I am not quite sure why you consider stocks with not enough volatility to beat the return of B&H not AIMable. Of course, when your only criterion is return, your statement is true. But as I have shown in my analysis of the Nasdaq 100 from Oct 1985 through March 2002, a smaller return can be compensated by lower risk:
AIM 10-10-5, no vealies: Return 9.7%, Risk 10.2%, Sharpe 0.47
B&H: Return 16.7%, Risk 30.6%, Sharpe 0.38
In this case, AIM would have beaten B&H on a risk/return basis.
I am a bit concerned about a 'trading' AIM, with all that loose cash. You may be tempted to start another LD-AIM, and another. All with the most volatile stocks you can find, for LD-AIM needs volatility even more than the BTB AIM. And when they all tank, you are wiped out in a hurry.
Just some thoughts,
Karel
Rien, you are right that investments that beat B&H don't lose stocks. They gain stocks, and that is why they beat B&H. My concern are the AIM investments that don't beat B&H on return, but compensate this with the reduced risk: these investments lose inventory (in # of stocks) in the long run.
Regards,
Karel
Well, I have these spreadsheets with N100 and SPY lying around, so. SPY isn't interesting but QQQ might come close to a stock you like. In Jan 1995 the N100 was at ~400. If you then had started to AIM with 100 'shares' N100, you would have sold 89 shares all the way up to ~4,400 in April 2000. In December 2000 AIM would have started buying again, at ~2500. No vealies, of course. With vealies at 33% (rather tight), you sell 60.
Another point to consider is that a lot of the AIMable stocks that don't really throw B&H lose stocks over their whole holding period. That is the price you pay for reduced risk. You couldn't Low-Down-AIM them without running out of stocks.
And why worry? When you run out of stocks, you just start a new LD-AIM position.
Regards,
Karel
Hi Sarmad, yes, I am the Sharpe guy
I assume you mention me in connection with risk assessment. Your measurement, $$$ below start, is pretty useless for that, but has a more respectable family member, called Maximum Drawdown. For the calculation of maximum drawdown, you take a historical maximum, and the lowest minimum after that. Take the difference in % from the high, and you have the Maximum Drawdown, a measurement for downside risk. Measure preferably over longer periods, and not only in a bull market!
BTW, I do some momentum investing myself. The aim (lowercase) is +15% or sell after 6-8 weeks. I am forward testing it at the moment, but I had a little success already with ATN (ACTN when I bought it). I bought before I thought about testing the method first, but it made the 15% mark, so no harm done.
Regards,
Karel
Thanks Tom, that will keep me off the street for some time! I am investing in the US myself, so checking for correlations with the AEX (our Dow) will have low priority. I just want to test out some asset allocation stuff.
Thanks again,
Karel
Hi Tom, I am interested in the IW history. On your Web site, you show the major turning points, but do you perhaps have a table with weekly or monthly IW data laying around somewhere? Or do I need to dig through your newsletter archive? I was looking for SP500 P/E's, then remembered http://www.yardeni.com , and there found nothing, as the site is gone! Perhaps I should have gone looking for your IW first!
Regards,
Karel
What I meant was, the broker warned that during the day it might seem as if you were charged for each fill, but that this would be corrected automatically.
And about the other partial fills, you are right to mention that the coffee is rather weak. At home I drink my coffee with lots of hot milk, but when I was in the USA, I drank my coffee black. You do the math. My darling coffee loving wife couldn't get used to the stuff. Not to mention the 'flavoured' varieties. And cappuccino was considered a 'flavoured' variety too, horror of horrors!
Regards,
Karel
Hi Grabber! Perhaps you got partial fills. I remember seeing on some broker's site that in that case they would seem to charge you for each fill, but that it would come out in the wash.
Regards,
Karel@dontqualifyforpartialfillsyet.nl
And he is right! (Past performance ...) I like the way his 'High PE' measurement goes with the flow. And it makes me wonder where I could find historical P/E info for the big indices (stocks would be nice too). Anybody any idea?
Regards,
Karel
Sharpe vs UPI! Thank you Don. Of course, when I started my foray into Risk Analysis territory, I started to stumble over lots of other ways to measure Risk/Reward. Another one I found (and liked) was:
http://www.aegisfund.com/english/library/articles/art_risks.html
But I had already started on the Sharpe Ratio, so I stayed with that measurement.
The Ulcer Index (or Ulcer Prevention Index; beautiful names!) only measures downside risk, and the Sharpe ratio both downside and upside risk. That the UI still doesn't make much difference from the Sharpe Ratio becomes clear when we look at the calculation:
1. Square all the daily drawdowns of the vehicle or strategy.
2. Sum all the daily squared drawdowns.
3. UI = the square-root of the sum.
Now an investment can have a higher UI for a combination of two reasons:
- it has a lower return (by comparable volatility), and so more (or greater) daily drawdowns
- it is more volatile (by comparable return), and the greater drawdowns are compensated by greater upswings.
We try to avoid investments with the first characteristic, so measuring only the downside volatility for UI in practice means measuring both upside and downside volatility.
Regards,
Karel
Tom, congratulations on a nice pick. I couldn't resist going for the Sharpe Ratio of this investment of yours. It should be obvious that this beats B&H, for the stock grew from 1.31+ to 2.43 (last entry), a rise of some 85%. In this period, your investment grew from almost 20,900 to over 55,000, for a gain of 167%. AIM always reduces volatility, so add a greater gain and you blow Mr. B&H's socks off.
I didn't follow all your transactions exactly, but set up a backtest for STKL, using monthly checks for buying and selling, 10% Safe both ways, and minimum transaction 5% of equity value. I didn't use vealies, as they don't appear in your stock trading history. I got a 162% gain, or 22.0% annualized. The volatility was still a shocking 63.7%, but this was a significant improvement over the 79.5% for B&H. Mr. B&H was not displeased with his return of 13.7% annualized, until he heard of your return, of course. AIM reached a Sharpe Ratio of .27, B&H got stuck at .11.
Is this investor's heaven? Perhaps it is, but the SR is rather low. The SRs calculated earlier, for SP500 and N100, were higher. That would mean that those would have been better investments from a Risk/Reward perspective. But we need to compare SRs over the same period, so here follow the benchmarks:
SPY (Jun 97-Mar 02): Return 6.8%, Volatility 18.4%, SR .10
SPY (AIMed): Return 6.4%, Volatility 10.2%, SR .14
N100 (Jun 97-Mar 02): Return 9.3%, Volatility 45.0%, SR .10
N100 (AIMed): Return 12.8%, Volatility 17.3%, SR .45
So, it may not be investor's heaven, but it comes close. SPY and N100 (QQQ was not available for the whole period) don't come even close, with one exception: N100 AIMed. It beats N100 B&H for a very nice return, and with an amazingly low volatility, for a Sharpe Ratio that outshines STKL. From a Risk/Reward standpoint, the N100 would have been the best investment, if AIMed. But to reach the same return as STKL, you would have needed leverage, and leverage usually costs. Tom, congratulations on a very nice investment!
Regards,
Karel
PS. Perhaps there is another aspect to volatility. I am not exactly AIMing myself, I use a kind of Scale Trading approach, but I catch myself thinking, when stock prices drop: perhaps they drop so far that I can buy again! While it is not exactly fun to see my portfolio, or parts of it, diminish, it is still exciting to get some action. What I mean is: an AIMer might experience less Risk than the Objective Risk of the investment. So, instead of trying to formulate the experienced risk and calculate an ExpSR, I will let this be the last of my excursions on the Risk/Reward territory (until something new crops up, of course).
K
Hi whitelake, good luck with the test! And don't forget to report back, one in a while!
Regards,
Karel