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 Barry,
Nice to hear from you again. I was searching about on Google last week and came across your Money Spinner site. Started to wonder where you'd gone.
Looking forward to your new ideas and posts.
Regards,
Mark
http://www.automaticinvestor.com
Hello Jerry,
Yes, you are correct that there can be cases when AIM has you selling below your average cost per share. My guess is, given the market conditions over the past year, that many are in just that predicament.
However I don't see this as a problem.
AIM works by buying low and selling high. Let's look at a situation where we've just started a new portfolio (I'll call this Scenario 1). We want AIM to sell some of our shares when the price rises in order to reduce some of our risk in the market.
The risk we're talking about, in Scenario 1, is the risk of the share price falling below the level where AIM recommended we sell.
That's one of the reasons we use AIM, to reduce risk. If we don't do this, there's a real possibility of seeing our shares climb to great heights, while we refuse to sell, and then seeing them descend to our original purchase price and blow right through it down to the pits of despair.
CASH IS GOOD
On the other hand, Cash we've raised from the sale of a portion of our shares can then be used to purchase additional shares when the price drops. Each time we do this we make some profit on a LIFO basis. As we repeat this cycle multiple times, we consistently increase our portfolio value slowly, but surely.
RISK IS BAD
Now consider what you are doing if you set an arbitrary value below which you will not sell any shares. When AIM recommends a sale, you check this arbitrary value and if AIM's recommended sale share price is less than this value you ignore the advice and don't sell.
What you're doing is, effectively, increasing your risk by some amount because you're refusing to take some cash off the table when AIM is suggesting you do so.
In Scenario 1, most would NOT agree with this course of action.
So in the case of a new portfolio where our sales are above our average purchase price, we'd probably just follow AIM's advice.
SHORT-TERM FOCUS IS BAD FOR LONG-TERM RESULTS
Now let's consider the case where our share price has fallen in such a way that AIM's next sell point is BELOW our average purchase price. I'll call this Scenario 2.
For all intents and purposes, the average purchase price is simply an arbitrary value (I'll be glad to expand on this if anybody disagrees). At that particular instant in time you've lost money in the stock market (hey, it happens). But your focus should not be an instant in time. You should have a LONG-TERM perspective.
RISK IS VERY BAD
If you focus on this arbitrary value and refuse to sell until it's met, you're in effect doing exactly, in Scenario 2, what we said you shouldn't do (back in Scenario 1) -- i.e. you're increasing your risk by some amount. Why should we behave differently in Scenario 1 than in Scenario 2?
SLOWLY BUT SURELY WINS THE RACE
On the other hand, if you follow AIM's recommendations, you're slowly but surely building a small profit (on a LIFO basis) with each buy/sell cycle (regardless of the fact you've sold below the average purchase price).
As the small profits grow, they eventually add up to larger profits and will eventually pull your sell points above the average price per share. If you don't do this, your risk will be increasing.
LOOK AT RISK AND REWARD -- NOT JUST REWARD
AIM is meant to do two things: Provide increased returns AND minimize risk. By focusing exclusively on the "increased returns" part (i.e. refusing to sell at a price below the average purchase price), you're ignoring the "minimize risk" part.
Therefore you're saying, "I'm willing to increase my risk in the market in order to have a chance at making a higher return." There's nothing wrong with that thinking, but it's not AIM.
The fact is, you've already lost money (i.e. the current share price is below your average purchase price). Realizing a loss in the short-term in order to reduce your risk and thus stand a better chance of coming out ahead when the game is over is a good strategy to follow.
RISK IS VERY, VERY BAD
Conversely, unneccessarily increasing your risk in an attempt to get back to your starting point in one fell swoop is a low-percentage, high-risk strategy.
FOOTBALL EXAMPLE
Here's something to think about. You're a football coach and your team has fallen behind by 7 -- in the Superbowl game, no less, with 10 seconds to play. On the next play your team scores -- touchdown!
However there's only 2 seconds left to play. What do you do? Do you kick a convert to tie the game or do you go for the two points to win? Keep in mind you might never get to the big game again.
I think most coaches in this situation would kick the convert because the risk of the 2 point play is too great even when the reward (i.e. becoming Superbowl champions) is considered. Put another way, the risk of losing is too great.
In effect they're saying we'll take a small victory now (i.e. making the convert) and set our team up for a better chance to win (i.e. using high-percentage plays) rather than risking it all on a low-percentage play.
CONCLUSION
The same kind of thinking should apply to LIFO gains and selling below average cost per share. Take the small victory (i.e. LIFO gain) and set your portfolio up for a better chance to win in the long-term.
BTW, Automatic Investor displays the average cost per share right on the screen for you, so you don't have to perform additional calculations.
Regards,
Mark
http://www.automaticinvestor.com
Hello Q&P,
I'm happy to hear you like the enhancements. The market has looked quite good these past 3 weeks and I've been quite happy with my portfolio's performance.
I'm working on a few more goodies here at the Automatic Investor factory, so keep an eye out -- there's more to come.
Regards,
Mark
http://www.automaticinvestor.com
Automatic Investor 2.0 Service Pack 2 is now available.
I was able to release the new version 3 days ahead of schedule.
SP2 contains a number of new charts and also includes the SP1 enhancements such as the MACRO filter and MattMod.
If you'd like to download the software, you can do so here --> http://www.automaticinvestor.com/trial.html
If you already have AI 2.0 installed, you can download Service Pack 2 from here --> http://www.automaticinvestor.com/upgradecenter.html
If you have any questions, my email address is mhing@automaticinvestor.com
Regards,
Mark
http://www.automaticinvestor.com
Automatic Investor 2.0 Service Pack 2 is now available.
I was able to release the new version 3 days ahead of schedule so the links to the pre-release version no longer exist.
SP2 contains a number of new charts and also includes the SP1 enhancements such as the MACRO filter and MattMod.
If you'd like to download the software, you can do so here --> http://www.automaticinvestor.com/trial.html
If you already have AI 2.0 installed, you can download Service Pack 2 from here --> http://www.automaticinvestor.com/upgradecenter.html
If you have any questions, my email address is mhing@automaticinvestor.com
Regards,
Mark
http://www.automaticinvestor.com
Hello J,
If you want to link to, say, www.automaticinvestor.com, preface the link with http://
So it becomes http://www.automaticinvestor.com
That should do it.
Regards,
Mark
http://www.automaticinvestor.com
And here's the update link for existing AI 2.0 users.
http://www.automaticinvestor.com/romeo/ai20sp2setup.exe
Download it and install it into the SAME folder where you originally installed AI 2.0.
This update will upgrade both AI 2.0 sp0 as well as AI 2.0 sp1 to AI 2.0 sp2.
Note that if you are upgrading from AI 2.0 sp0, you have to import the DEFAULT Model, see http://www.automaticinvestor.com/upgradecenter.html for further details.
If you have any questions, please email me at mhing@automaticinvestor.com (please don't post to this board as it's not an AI support board).
Also note that this is a pre-release version that I've made available only to readers of this board. It's not available anywhere else (including the main AI website).
Regards,
Mark
http://www.automaticinvestor.com
Hello Everyone,
I've posted a pre-release version of Automatic Investor 2.0 sp 2 on the AI website (http://www.automaticinvestor.com/romeo/ai20setup.exe )
It's about 9 MB.
If you don't have AI 2.0 installed, feel free to download it and give it a try (it includes the latest in charts as well as MACRO AIM, MattMod and voluFilter).
If you already have AI 2.0 installed then don't download this version. An update will be available shortly.
Note that this is a pre-release version. The official release will be available on November 8th.
As always, let me know if you have any questions (please email me at mhing@automaticinvestor rather than posting to this board).
Regards,
Mark
http://www.automaticinvestor.com
Hello Arawat,
Yes, there will be some more descriptive error messages in a future update. However not all error messages can be easily described -- some can be caused by a number of different factors.
Please email me the details of what you were doing when you received the error message. My email address is mhing@automaticinvestor.com
BTW, for support questions and enhancement requests please either email me directly or use the AI 2.0 user's group at http://automaticinvestor.community.everyone.net/community/scripts/topics.pl?NodeID=332353
I'd like to keep this board focused on new AI techniques and Ideas.
Thanks,
Mark
http://www.automaticinvestor.com
Hello Irwin,
Here's a result from the Historical Analyzer...
And here's what the optimizer looks like...
The results are displayed in the "Optimization Results" field (yellow box at the bottom) and you can save the optimized parameters by simply clicking the "Save Model" button (bottom right of the screen in the "Model Details" section).
This saves the Model in the system and allows you to use it in your own portfolios as well as the Historical Analyzer.
The Automatic Investor User's guide has additional details.
Let me know if you have any further questions.
Regards,
Mark
http://www.automaticinvestor.com
Hello Steve,
Do you mean when you exit the AI website?
If so, then yes, it is cookie controlled and should only be display once every 30 days.
If you mean something else, please email me with the details (mhing@automaticinvestor.com).
Regards,
Mark
http://www.automaticinvestor.com
Hello Conrad,
No, the stock portion is blue. Perhaps you are viewing a cached image?
Regards,
Mark
http://www.automaticinvestor.com
Hello Arawat,
Thanks for the suggestion. I'll add it to the considerations list. Let me know if you have any other suggestions.
Regards,
Mark
http://www.automaticinvestor.com
Hello Bernie, the next update is Service Pack 2 and is scheduled for release on November 8, 2002.
It will include the charts I've been previewing here. I'll post a note once it's been released.
Regards,
Mark
http://www.automaticinvestor.com
Hello Tom,
I thought your suggestion was a good one, so I added the stacked chart with Cash Reserve and the individual Equity amounts.
Here's a preview...
Regards,
Mark
http://www.automaticinvestor.com
Hello Conrad,
"Stocks that are too hot to handle should be RED, Reliable Stocks should be BLUE"
Thanks for the suggestion. I've changed the colours to GREEN for the Cash Reserve and Blue for the Stocks (because all stocks we choose for AIM are reliable
Regards,
Mark
http://www.automaticinvestor.com
Are you kidding! Canada is a net exporter of Arctic Air
When I'm done with Automatic Investor I'm thinking of opening a little Arctic Air shop on the corner.
Regards,
Mark
http://www.automaticinvestor.com
Hello Tom,
Yes, the total Cash Reserve/Equity Value is one of the charts (see below).
I hadn't planned on splitting the equity portion into its individual components, but that's a good suggestion.
Anyone else out there think an overall total portfolio value split into Cash Reserve and each individual equity would be useful?
Regards,
Mark
http://www.automaticinvestor.com
Hello Steve,
The reason I think the 0% initial cash value is best is because stocks have an inherent tendancy to rise.
When you invest your funds in a stock, there are 3 things that can happen:
1. The price goes down.
2. The price goes up.
3. The price stays the same.
Unfortunately we don't know which one of these three options our stock will choose (if we did, then we wouldn't need AIM).
Analyzing this a bit more, we can see that if the price goes down enough to pick up extra shares, then starting with a value greater than 0% initial cash will be better.
However if the price goes up, the opposite is true (as you'll eventually end up purchasing more expensive shares with your cash).
If the price stays the same, then neither method has the advantage.
So it all comes down to whether your portfolio of stocks will go down or whether they'll go up. (Keep in mind that even if the market goes up or down, your stocks might do something else altogether.)
The only way I know to determine whether a group of stocks will go down or up, in general, is to backtest over a number of different time periods -- and use numerous stocks.
By doing this, I've come to the conclusion that most of the time the majority of stocks in a portfolio will tend to go up rather than down. Note that this says nothing about an individual stock (which can either go up or down). However it says a great deal about a large group of stocks.
Since we don't know whether the stocks in our portfolio will go up or down, I feel it's best to go with the probabilities and start with a 0% cash reserve. For one stock this might or might not be the correct action, however as we diversify our stocks (and the number we hold rises) the probability is with our portfolio tending to rise.
Basically we're starting out on equal footing with the Buy and Hold strategy and using AIM to manage our portfolio after we raise some cash by selling.
Once AIM sells, we have a profit and we're using our profit to, hopefully, purchase additional shares and therefore average down our share price over time.
However there is one tool that somewhat changes what I said above. That tool is the Idiot Wave. I believe that if all 4 IW components are in the high-risk zone, then you should not use 0% as your initial cash value. Rather you should use what the IW recommends. (Note that ALL components need to be in that high risk zone.)
Otherwise I'd recommend sticking with the 0% initial cash value.
Of course you should convince yourself of this before just taking my word for it. And the Automatic Investor multi-security Historical Analyzer allows you to do just that. You can read more about it by clicking the "Help" button in AI 2.0 and then searching for "Historical Analyzer"
Also note that the AI DEFAULT Model includes the MACRO filter and the VoluFilter. Therefore the DEFAULT model is more than simply starting with 0% initial cash. Again, the best way to convince yourself about the merits of the DEFAULT Model is to run some backtests on your chosen stocks with the various Models over a number of different time periods.
I hope that helps a bit. If you have any other questions please let me know or post a note here (where there are many experienced AIMers always ready to help).
Regards,
Mark
http://www.automaticinvestor.com
Hello Conrad,
Yes, I'm very impressed with Don Carlson's MACRO filter. Like the MattMod, it's a simple concept that has you wondering why you didn't think of it.
Interesting VORTEX results. I think optimization (or Model Building as it's called in AI) is a good thing as long as it's done correctly.
However for general comparisons, it's always best to use non-optimized results (i.e. pick one set of parameters and use them for ALL comparisons).
The thinking is that once people see how well your system does with unoptimized parameters, they'll look at it more closely. Then when they become familiar with it, they can wring out better returns and lower risk by optimizing.
Regards,
Mark
http://www.automaticinvestor.com
Hello Tom,
Yes, it's been a long time coming. Believe it or not I had this task on my AI enhancement list for over a year and a half.
But I kept getting sidetracked with things like the optimizer, automated historical multi-stock analyzer, Models, algorithm enhancements (such as MACRO AIM) and such.
Here's a look at the stacked chart that's also included in the upcoming AI 2.0 update...
Regards,
Mark
http://www.automaticinvestor.com
Hello Too Fuzzy,
As I mentioned to Bernie, I was using the term "Standard AIM" to refer to the algorithm rather than the actual settings (since I was comparing the MattMod algorithm and the MACRO AIM algorithm with the Standard AIM algorithm). That's the reason I listed my parameter settings as 10/10/0/5/5 in the initial note.
However if we ignore what I was really talking about for a minute, you bring up a good point. When speaking about parameter settings and update periods, there are many Standard AIM models. Lichello has come out with 3 models and suggested update periods of quarterly and monthly.
Therefore when you use the term "Standard AIM," you should always include the settings so everyone is clear what parameters are being used.
The other thing I mentioned in my original note is that these were ad hoc tests, not systematic tests. In reality I was just testing the new AI 2.0 charting functions and decided to use ORCL, among others, because I know it's a good AIM stock.
And you are correct, that in this particular case, monthly updates and a Model starting with 33% cash gives better returns.
However I didn't just pick an initial starting cash value of 0% and daily updates out of the air. Over the past 2 years I've done thousands of systematic tests for many different stocks over many different periods and the results show that these settings work better than any other in the general case over the long run -- even when the markets immediately turn against you.
The trap into which many people fall is concentrating on one stock that did well with their settings and update period and then parading it as the holy grail and the basis for using their particular settings (BTW, I'm speaking in the general sense now, I'm not saying that's what you do). However we need to forget what we THINK should work best and make decisions based ONLY on empircal evidence.
You can use AI 2.0's multi-analysis function to see for yourself. Create a text file with all the, say, S&P 500 ticker symbols (one per line) and then invoke the multi-analyzer using "standard AIM" settings and monthly updates as well as 10/10/0/5/5 settings and daily updates over various time periods.
You'll see that the "standard AIM" setting severely underperforms 10/10/0/5/5 with daily updates most of the time.
The beauty of Don Carlson's MACRO filter is that you don't have to worry about update intervals. You can update as often as you like and the MACRO filter usually filters the inefficient updates for you.
To sum up, this wordy note, I'm not against other "standard AIM" settings as long as they're being used for logical reasons. However until I see hard data suggesting that 10/10/0/5/5 with daily updates underperforms another set of parameters and update period, I'll continue to use these settings in comparisons as well as my real investments.
Regards,
Mark
http://www.automaticinvestor.com
Hello Bernie,
Thanks, I got the spreadsheet. I ran another test with the data starting at the 1st of each month and came out with 140%, so there's still a difference.
I'll check the spreadsheet when I have a chance and see if I can determine what's causing the variation. The only difference I can think of right now is the data source (Yahoo! vs. MSN).
Regards,
Mark
http://www.automaticinvestor.com
Hello Bernie,
You didn't do anything wrong. I used the term standard AIM referring to the algorithm. I also noted my settings as 10/10/0/5/5 (which is BR=10%, SR=10%, Initial Cash=0%, Min%Buy=5% and Min%Sell=5%).
The other difference is that I did daily updates rather than monthly updates.
I just tried running your settings and came up with 127% ($22,690.64), not the 172% you saw. I'm curious to know where the difference lies (the only thing I can think of is the "5% or $500" you use, while I use a straight 5%).
Where did you get your data? Mine came from Yahoo! Finance.
Here's the complete report (including all the trades).
AUTOMATIC INVESTOR HISTORICAL ANALYSIS FOR ORCL
======= PERFORMANCE =======
Current Portfolio Value: $22,690.64 (2182 shares owned)
Profit or (Loss): $12,690.64
Simple Return: 127%
Annualized Return: 18%
Buy/Hold Portfolio Value: $17,241.76 (1712 shares owned)
Buy/Hold Profit or (Loss): $7,241.76
Buy/Hold Simple Return: 72%
Buy/Hold Annualized Return: 12%
Return on Capital at Risk: 243.83%
Average % Capital at Risk: 52.05%
Analysis run on 29-Oct-02
Monthly Price Data Chosen
Filters Used:
Include Buy Advice
Include Sell Advice
Exclude Hold Advice
Average Commission: $10.00
Start of Period: 29-Oct-97
End of Period: 29-Oct-02
High for Period: $46.3100 on 1-Sep-00
Low for Period: $3.0100 on 12-Jan-98
Initial Number of Shares: 1,147
Original Share Price: $5.8400
Original Share Price Date: 29-Oct-97
Original Investment: $10,000.00
Last Share Price: $10.0700
Last Share Price Date: 28-Oct-02
Average Cost Per Share: $8.63
Current Cash Reserve: $717.90
Current Shares Owned: 2,182
Current Security Value: $21,972.74
Number of Purchases: 13
Number of Sales: 11
======= CONFIGURATION =======
Buy Resistance: 10%
Sell Resistance: 10%
Initial Cash Reserve: 33%
Margin option is Off
Minimum Purchase: $0.00
Minimum % per Purchase: 5%
Minimum Sale: $0.00
Minimum % per Sale: 5%
Control Increment: 50%
Maximum Cash: 100%
Vealie: 50%
Volumizer option is Off
MACRO Filter is Off
Volume Filter is Off
MattMOD option is Off
Portfolio Control: 20,828
======= STATISTICS =======
Total Number of Records Processed: 58
Total Number of Records Printed: 26
Total Number of Records Skipped: 0
======= RECOMMENDATIONS =======
29-Oct-97 START shares @ $5.8400 own 1,147 valued @ $6,698.48 Cash: $3,301.52 Portfolio Value: $10,000.00
2-Jan-98 BUY 487 shares @ $3.8300 own 1,634 valued @ $6,258.22 Cash: $1,426.31 Portfolio Value: $7,684.53
11-Jun-98 BUY 82 shares @ $4.0600 own 1,716 valued @ $6,966.96 Cash: $1,083.39 Portfolio Value: $8,050.35
16-Nov-98 SELL 103 shares @ $5.4100 own 1,613 valued @ $8,726.33 Cash: $1,630.62 Portfolio Value: $10,356.95
17-Dec-98 SELL 258 shares @ $6.5300 own 1,355 valued @ $8,848.15 Cash: $3,305.36 Portfolio Value: $12,153.51
21-Jan-99 SELL 258 shares @ $8.1100 own 1,097 valued @ $8,896.67 Cash: $5,387.74 Portfolio Value: $14,284.41
23-Feb-99 SELL 188 shares @ $9.7500 own 909 valued @ $8,862.75 Cash: $7,210.74 Portfolio Value: $16,073.49
25-Mar-99 BUY 169 shares @ $6.6700 own 1,078 valued @ $7,190.26 Cash: $6,073.51 Portfolio Value: $13,263.77
27-May-99 BUY 234 shares @ $5.8900 own 1,312 valued @ $7,727.68 Cash: $4,685.25 Portfolio Value: $12,412.93
29-Jun-99 SELL 206 shares @ $9.2800 own 1,106 valued @ $10,263.68 Cash: $6,586.93 Portfolio Value: $16,850.61
1-Oct-99 SELL 195 shares @ $11.3100 own 911 valued @ $10,303.41 Cash: $8,782.38 Portfolio Value: $19,085.79
2-Nov-99 SELL 137 shares @ $13.2500 own 774 valued @ $10,255.50 Cash: $10,587.63 Portfolio Value: $20,843.13
3-Dec-99 SELL 236 shares @ $19.6700 own 538 valued @ $10,582.46 Cash: $15,219.75 Portfolio Value: $25,802.21
5-Jan-00 SELL 129 shares @ $25.5000 own 409 valued @ $10,429.50 Cash: $18,499.25 Portfolio Value: $28,928.75
7-Feb-00 SELL 66 shares @ $29.9700 own 343 valued @ $10,279.71 Cash: $20,467.27 Portfolio Value: $30,746.98
9-Mar-00 SELL 93 shares @ $42.0000 own 250 valued @ $10,500.00 Cash: $24,363.27 Portfolio Value: $34,863.27
16-Nov-00 BUY 56 shares @ $27.3700 own 306 valued @ $8,375.22 Cash: $22,820.55 Portfolio Value: $31,195.77
23-Feb-01 BUY 110 shares @ $22.0000 own 416 valued @ $9,152.00 Cash: $20,390.55 Portfolio Value: $29,542.55
27-Mar-01 BUY 205 shares @ $16.6500 own 621 valued @ $10,339.65 Cash: $16,967.30 Portfolio Value: $27,306.95
27-Apr-01 BUY 59 shares @ $17.1500 own 680 valued @ $11,662.00 Cash: $15,945.45 Portfolio Value: $27,607.45
30-May-01 BUY 164 shares @ $14.5100 own 844 valued @ $12,246.44 Cash: $13,555.81 Portfolio Value: $25,802.25
31-Aug-01 BUY 253 shares @ $12.2100 own 1,097 valued @ $13,394.37 Cash: $10,456.68 Portfolio Value: $23,851.05
19-Apr-02 BUY 132 shares @ $11.9300 own 1,229 valued @ $14,661.97 Cash: $8,871.92 Portfolio Value: $23,533.89
21-May-02 BUY 542 shares @ $8.8500 own 1,771 valued @ $15,673.35 Cash: $4,065.22 Portfolio Value: $19,738.57
21-Jun-02 BUY 411 shares @ $8.1200 own 2,182 valued @ $17,717.84 Cash: $717.90 Portfolio Value: $18,435.74
28-Oct-02 LAST shares @ $10.0700 own 2,182 valued @ $21,972.74 Cash: $717.90 Portfolio Value: $22,690.64
Regards,
Mark
http://www.automaticinvestor.com
I've just added charting functionality to Automatic Investor 2.0 (so for everyone who's been asking, it will be available very soon).
Take a look at the charts below to see what they look like.
One thing I noticed, while testing the AI 2.0 charts, is how the visual aids give you quite the perspective on the various AIM modifications.
For example, while testing the charts on Matt Crider's MattMod I noticed an interesting phenomenon.
The MattMod appears to be slightly more efficient than standard AIM (however keep in mind that these are just AD HOC tests). I ran a dozen or so stocks through the historical analyzer and charted them. I'm posting my Oracle results here. I also added Don Carlson's MACRO AIM to the mix.
I used ORCL from Oct. 28, 1997 to Oct. 28, 2002.
Standard AIM's return was 58%, the Buy & Hold return was 74% and the MattMod return was 143% and the MACRO AIM return was 299% over the period.
The settings were 10/10/0/5/5.
Take a look at the charts below. You'll notice that both methods performed similarly as the price rose dramatically. Then the MattMod stopped all activity until the price fell significantly. Whereas Standard AIM continued its activity and bought a couple of dips and sold a couple of peaks.
Unfortunately Standard AIM started buying too soon on the way down while the MattMod waited until a more opportune time.
Here's the standard AIM chart...
and here's the MattMod chart...
and finally here's the MACRO AIM chart...
You'll notice that MACRO AIM made fewer trades and behaved very different from standard AIM or the MattMod. However in this case it was more efficient.
Regards,
Mark
http://www.automaticinvestor.com
Hello Arawat,
This could be caused because your stock has no volume data associated with it and the VoluFilter is turned on.
See the message at http://www.investorshub.com/boards/read_msg.asp?message_id=555037 for more details.
If you don't think this applies to your case, just send a note with the details (i.e. stock symbol, what you're doing when this error occurs, etc.) to mhing@automaticinvestor.com and I'll look into it.
Regards,
Mark
http://www.automaticinvestor.com
Hello Karel/Myst,
"meaning the buy and sell SAFE, Minimum buy and Minimum sell shares and PC will have separate input boxes for quick changes. A push of a button will allow the user to find optimal AIM settings for their play."
Automatic Investor 2.0 already has this feature built into its optimizer. So a push of the button gives optimal settings for a stock based on the historical data you fed it (although I'm not sure how you'd go about optimizing the PC as this is dependent on the buy amounts. AI 2.0 allows you to optimize the Control Increment instead -- the Control Increment is the percentage you increase the PC when a purchase is made).
However I've found that optimizing (and in AI's case saving the optimum results as Models) is not as easy as pushing a button. Rather Model building requires a systematic approach to ensure the optimized parameters work well over most time periods and most market conditions.
It is quite easy to overfit a set of parameters to a specific set of historical data and see great results. However those parameters may not work well going forward. That's the big trap many fall into.
Myst, when you're ready to implement your optimizer, have Yong look at genetic algorithms. These algorithms allow for faster searches over the possible parameters' ranges. And they're relatively easy to implement.
For more on optimizing and building models, take a look at the Automatic Investor User's Guide (http://www.automaticinvestor.com/AIHelp20sp1.pdf or click the HELP button if you already have AI 2.0 installed) starting on page 65.
Regards,
Mark
http://www.automaticinvestor.com
Hello Q&P,
I just sent you a note regarding the error message you're seeing.
Mutual funds don't have volume data associated with them, so when the VoluFilter is turned on, it throws out an error.
I've fixed this bug and it will be included in the next update.
In the meantime you can ignore the error message as it doesn't affect AI's recommendations.
Regards,
Mark
http://www.automaticinvestor.com
Hello NeuralNet,
The studies assume the closing price because that data is easy to come by. And it doesn't affect the comparison.
The intent of all the studies was to show how AI results compared with the Buy and Hold strategy's results. Therefore as long as the same prices were used for both Automatic Investor and the Buy and Hold strategy, the comparisons are valid.
I could have used an average of the high, low and close prices, but the close was adequate in this case.
In the real world, you would most likely update AI at some time (let's say at noon), and then place your order based on any AI recommendations. Alternatively you could wait until the markets close, update AI and then place your order for the next day.
These details certainly matter when you're actually managing your portfolio on a day-to-day basis, however they're not as important for comparison purposes.
Most of the time, you won't have an order filled at the recommended AI price anyway (because by the time you place your order, the price will have most likely changed). However as long as the order amount is "close" to the AI recommended amount, everything works well.
I hope that helped a bit, please let me know if you have any other questions.
Regards,
Mark
http://www.automaticinvestor.com
Hello Q&P,
I just replied to your email.
Let me know how you make out.
Regards,
Mark
http://www.automaticinvestor.com
Hello Q&P,
I just sent you an email, but I'll post some comments here too...
It sounds like the update wasn't installed correctly or the shortcut you're using to invoke the software is pointing to the wrong place.
Go to the folder where you installed AI 2.0 sp1 and look for a file called AutomaticInvestor2.exe
When you find it, double-click on it and then check the "About" function on the "Help" menu. It should display Service Pack 1.
Also, in the same installation folder, do you see a file called "default.xml?"
If things still aren't working I can send you the new files via email and we can try installing them manually.
Regards,
Mark
http://www.automaticinvestor.com
Hello NN,
Yes, that's right. The trial version you download now has MACRO AIM, VoluFilter and MattMod included.
Regards,
Mark
http://www.automaticinvestor.com
Hello Conrad,
"This leads me to think that the definition of randomness is ultimately a practical issue rather than a theoretical one. Any set of numbers that we produce are constructions, no matter how hard one tries to deny this. Numbers are abstractions and Number Sets do not appear magically out of nowhere. As long as we can not make any sense out a of set of numbers then we may call it random, even if there happens to be some structure found in it."
I like your statement. Especially the last sentence. It reminds me of the Arthur C. Clarke quote about how sufficiently advanced technology would be indistinguishable from magic.
My view is that we call events (not just numbers) random because we lack the ability to calculate the causes of the effect. Hence we have sayings such as, "random weather patterns."
If we could calculate how a butterfly's wing beat starts a chain reaction of events that combine with still more events and eventually create a storm many miles away, then our weather patterns wouldn't be random.
To generate random numbers, we can base it on something that nobody can calculate, such as the number of neutrinos from the sun hitting a certain spot on the Earth.
This would serve as an excellent source of random numbers... until someone discovered how to calculate the number by starting at the sun, following each neutrino's journey to Earth and determining how many actually hit that spot at a particular time. Then the numbers wouldn't be random anymore.
Even a number you choose in your head would not be random if technology advances enough to map your brain and determine what influences your choice of number.
Given that we can't choose absolutely random numbers, it would appear that the best course of action is to choose something that is seemingly random in the domain in which we're interested (e.g. stocks).
And it would also make sense to choose the simplest method of generation we can (but as Einstein said, "no simpler"). After all, why bother with counting neutrinos when you can use your PC's random number generator.
So I'd agree with you that a computer's random number generator is good enough for your stock tests. Keep in mind that a computer actually uses a pseudo-random number generator and it will produce the same sequence of "random" numbers given the same seed. So you might want to seed it with something that changes (such as the current time).
Anyway, I enjoyed your thoughts on random numbers.
Regards,
Mark
http://www.automaticinvestor.com
Hello Karel,
Glad to hear everything went well. Let me know if you have any questions or comments once you have a chance to use the new functionality.
Regards,
Mark
http://www.automaticinvestor.com
Automatic Investor 2.0 Service Pack 1 is Now Available!
Service Pack 1 contains the latest and greatest AIM enhancements.
It includes Don Carlson's MACRO AIM, the VoluFilter (that filters inefficent AIM trades based on Volume) and Matt Crider's MattMod.
The Filters have combined with the AIM algorithm to SIGNIFICANTLY improve results.
There are a number of detailed historical studies available for your perusal. You can start here --> http://www.automaticinvestor.com/study.html and here --> http://www.automaticinvestor.com/PerformanceStocks.html
You can view a couple of online video tutorials here --> http://www.automaticinvestor.com/Tutorials.html
And if you're not an Automatic Investor user, you can download your own Trial copy here --> http://www.automaticinvestor.com/trial.html
If you are an Automatic Investor user (well, Congratulations! You're in for a treat , you can download the FREE Service Pack upgrade from here --> http://www.automaticinvestor.com/upgradecenter.html
If you have any questions, please don't hesitate to ask.
Regards,
Mark
http://www.automaticinvestor.com
Automatic Investor 2.0 Service Pack 1 is Now Available!
Service Pack 1 contains the latest and greatest AIM enhancements.
It includes Don Carlson's MACRO AIM, the VoluFilter (that filters inefficent AIM trades based on Volume) and Matt Crider's MattMod.
The Filters have combined with the AIM algorithm to SIGNIFICANTLY improve results.
There are a number of detailed historical studies available for your perusal. You can start here --> http://www.automaticinvestor.com/study.html and here --> http://www.automaticinvestor.com/PerformanceStocks.html
You can view a couple of online video tutorials here --> http://www.automaticinvestor.com/Tutorials.html
And if you're not an Automatic Investor user, you can download your own Trial copy here --> http://www.automaticinvestor.com/trial.html
If you are an Automatic Investor user (well, Congratulations! You're in for a treat , you can download the FREE Service Pack upgrade from here --> http://www.automaticinvestor.com/upgradecenter.html
If you have any questions, please don't hesitate to ask.
Regards,
Mark
http://www.automaticinvestor.com
Hello Neuralnet,
I see leapyear has already answered your question. The next update is scheduled for release October 31, 2002 (although it might be sooner).
I think everyone will be quite happy with the new version as it consistently beats standard AIM, Buy and Hold and all the other AIM-like investing software out there.
If you have any other questions, please let me know.
Regards,
Mark
http://www.automaticinvestor.com
Hello Don,
"However, you didn’t mention the oppressive, cream-skimming consequences of periodic profit selling using AIM-like strategies in a taxable account."
You are correct in saying that the tax consequences have an impact on the results. However there is no standardized way to include taxes when running a generalized comparison.
On the other hand, once an Automatic Investor user in a particular country has decided to invest in a particular stock, he can run the historical analysis on that stock and see exactly how many trades took place and exactly what they were. This allows him to estimate his tax consequences for his specific case.
There are two further notes I should make.
1. With the Volume Filter and Don Carlson's MACRO AIM, a good number of unneccessary trades are filtered. This dramatically decreases the number of trades and thus the tax problems (in fact the number of trades decreased so rapidly in my tests that I think that some people won't like it because there won't be enough action. However the results speak for themselves, so if action is what someone values over results, then they should look into another strategy -- such as day trading for example).
"This is skimming that buy-and-holders do not suffer."
2. What most Buy and Hold advocates don't mention is that you will still be taxed on your Buy and Hold proceeds, the tax is only deferred, not eliminated. Buy and Hold is not a completely accurate description of that strategy. A more apt description is Buy and Hold a long time and Sell.
When you sell, you'll have to deal with "the oppressive, cream-skimming consequences" of the tax system.
Granted that it is still more desirable to defer taxes than to pay them right away, but I don't believe this disadvantage outweighs the advantages of using the AIM variant.
However your point is well taken and I will add a note regarding my tax assumption to the Web page.
Regards,
Mark
http://www.automaticinvestor.com
Hello Bernie,
Yes, I am showing the results of the stocks that made up the index.
However the point of the exercise is to choose a stock (any stock will do) and then calculate the results for both the AIM variant and the Buy and Hold strategy.
I'd then like to repeat this process a number of times in order to minimize any errors (i.e. those caused by the fact my stock's share price might have had a price pattern that just happened to do better with the AIM variant or with the Buy and Hold strategy).
Once I've completed my analysis on, say, 500 different stocks, I'd like to sum the results to see which method did better over the realm of the chosen stocks.
The problem with this method is that I've been backtesting stocks with AIM for over 2 years now. So I (either consciously or unconsciously) know which stocks do better with AIM. For example, without even looking I can tell you that over the past 5 years both MSTR and DCLK did MUCH better than GE when used with AIM.
However I don't want to taint my results by choosing all the DCLK's of the stock world and eliminating all the GE's. That would definitely result in AIM coming out ahead, but what's the point since I've filtered the inputs and stacked the deck against Buy and Hold.
To get around this I decided to use the major indicies. That way I have NO input into selecting which stocks I'm going to use for my tests. Someone else selects the input and all I have to do is run these stocks through the AIM variant and note the results.
The results simply state that for the stocks tested over the given periods, the AIM variant did better than the Buy and Hold strategy. They don't attempt to state anything further.
I hope that clarifies things a little.
Regards,
Mark
http://www.automaticinvestor.com
Hello Bernie,
Yes, you are correct about the indicies's makeup being different today than they were in the past.
However that fact doesn't apply in this case as the results are meant to simply compare the AIM variant's performance to the Buy and Hold strategy's, not to prove that particular stocks are/were good investments.
The reason for choosing the major indicies is to ensure there is no bias in selecting stocks that I know (or unconsciously know) will perform better with AIM than with Buy and Hold.
In effect, I'm allowing others to choose a standard set of stocks (e.g. the Dow) and then running AIM on whatever they choose.
If I'd performed the study 10 years ago some stocks would have been different, but it doesn't matter since I'm only using them for comparison purposes.
On the other hand, if I'd been using them to say that it is better to have invested in , say, all of the Nasdaq stocks 10 years ago rather than any other index, then survivorship bias would come into play and your point would be valid.
Regards,
Mark
http://www.automaticinvestor.com