SystemTrader's Account Talk

Status
Not open for further replies.
imported post

Discuss SystemTrader's account and general approach of timing/fund selection based on tested, mechanical models. Thanks for the comments and encouragement, Rod and Chaplain!
 
imported post

Right now you are 100% I --an interesting place to be. Wouldn't it be safer to spread the risk in the other stock funds?
 
imported post

Safetyguy wrote:
Right now you are 100% I --an interesting place to be. Wouldn't it be safer to spread the risk in the other stock funds?
Well, days like today are an awesome time to be 100% I!:^

Please keep in mind, this is part of John's system he is testing.
 
imported post

"Kind regards,
John, AKA "SytemTrader," "SystemGeek" or whatever else (within reason) you want to call me. "

How 'bout my new best friend.

Welcome ST. I love the idea that your system always has you 100% in something. I guess that's the gambler in me. I have 2 questions for you. First, look at today as an example, where the I fund appears to be overachieving and the S fund appears to be underachieving (badly), would your system tell you to get into the one that's hot, or the one that's due? Second, there has been a ton of discussion on this board about the I fund and particularly to what extent the rising/falling dollareffects the price of the I fund. Does your system consider the effect of the dollar and to what extent?

Thanks in advance for your answers. I'll be watching (and perhaps mirroring) closely.

Dave
 
imported post

I'm answering FundSurfer's questions/comments here. I'll try to answerother questions (such as why the system only invests in one fund) later. What I've written here and a few other places on the board is enough for one evening! I'm trying to address the concern that timing models may not work in the future and/or that they may be anticipated
by other investors.


First, thanks for your comments. Yes, it is true that backtested results could be faked orbe the result of data mining (this essentially means looking for something that “worked” in the past until you find something that's probably a statistical fluke). And you're right--no system or investing guru can guarantee future results.But because of this, I prefer to use something that has performedwell in many different market climates—in up, down and sideways markets including the '87 crash, the Gulf War, the roaring mid-late 90s, the subsequent bear market, etc.

While timing systems aren't perfect, you can take measures to increase the odds that they'll bereliable in the future. First,they should ideally be built/tested in one time frame then tested in an “out-of-sample” period andin real time. Most timing models I use were built by others and were subjected to this sort of rigorous testing. Some have been around since the 90s (and even a few since the 80s) and work just as welltoday as they did 10 years ago. Practically all of them avoided most/all of the 2000-2003 bear market though they were designed well before it—not in hindsight. To my knowledge, these models aren't known or used by many people. (We're not talking 200-day moving averages or VIX reversals.) And if you think about it, how many of the 3.5 million TSP participants,22+ million 401k participants, or ?? million IRA owners actually do this sort of thing? Very, very few. And if they use their local stock broker or financial planner's services, I can almost guaranteethey don't.

Timing systems should also "make sense" and have some theoretical basis, whether it be trend-following (crowd behavior), economic, market relations/comparisons (like how the NASDAQ usually leads blue chips heading into a rally...it happened yet again this fall!, etc.) I have designed several of the timing systems I use, but all are based on some sort of established relationship with a logical basis. I have just combined them in unique ways and added my own parameters. If I had just found some strange model that happened to work (like basing it on the score of the Super Bowl each year), that would be a different story.




Another way to bolster timing model reliability is to test for “parameter sensitivity.” If a model is too reliant on strict parameters, that's a danger sign. For example, if a 20-day moving average produces beautiful backtested results, but the 19-day and 21-day variations do terribly, you probably have a fluke. By optimizing (i.e., testing every possible combination and viewing the results), it's easy to see if a wide range of parameters outperformthe market. I've done this.




Finally, I use a composite model for both stocks and bonds. This assigns numerical “weights” to different timing models, and produces a buy signal when a certain weighted score is reached.




The beauty of composites is that they require multiple conditions (price, volume, oil prices, activities of professional investors, intermarket relationships, interest rates, seasonal trends, etc.) to enter or exit the market, and these conditions may be different each time a signal is given. As I mentioned, my composite contains 9 timing systems, a few of which I designed, and I've run optimizations that tested over 3 million combinations. If anyone thinks they can crack the code or anticipatemy next signal, then have at it! ;)




One of my favorite market researchers, Nelson Freeburg, who builds “composites” for big-time institutional money managers, said the following about them: “they combine the individual signals from a variety of components into a structured whole. The consensus of indicators includes a diverse mix of data--market breadth, sentiment, monetary and trend-sensitive inputs. All of our custom models have outperformed the S&P 500 in real-time in terms of risk and reward.” Since Nelson has a lot of experience and prominence in this field, I don't think most of us can afford his services. But hopefullyfolks can afford mine, especially now while you can follow them for free. :D


Well, I'd like to address a few other things, but I'll wait until later. The bottom line: I'm not some self-proclaimed "market genius" who claims to have found the flawless holy grail of market timing. The Internet is already full of such characters! I've simply compiled a diverse, "best of breed" set of timing models, add a few models/tweaks of my own, and customized it for the TSP. This has taken considerable time and expense but I'm finally satisfied with the results. My system may not be for everybody, and that's fine. I hope everyone here does great with whatever approach they use.




I will leave one question for some pondering, though, and it's what I ask myself about all trading/investing activities I do: willyour approachcontinue to work wellno matter how the market behaves? I startedinvesting byusing a combination of news events, chart reading, indicators, advice from others, hunches, etc. I quickly and painfully found out that I'm not a very good intuitive trader. I can't filter thenear infinite amountof market info we're bombarded with and consistently make good decisions with it. That's why I prefer systematic approaches. If done correctly, they filter the most important parts of this data and remove all of your emotions and second guessing. No, they're not perfect, but if done properly, they are,in my humble opinion, the best way to approach the future. My system is doing just as well (if not better) right now as it did in the early 90s, so Ihave no reason to doubt it will continue to perform well in the future.




Thanks again for your input.



John
 
imported post

Oh, and re: other paid services, I'm aware of them. When I first thought about doing this, I thought I may be the first, but I soon discovered a few others. I don't subscribe to any others, so I can only judge them by what's on their websites. And in this highly litigious age, I'm a little gunshy about talking about them. I will say that one service uses some of the same concepts I do, but it uses a completely different framework to select the funds. (Again, this is based only on what I've read at their site.) I'm not very clear on the methodology used by the other services.




There are a couple of things that appear to bedifferent about my system. First, it's a top-down approach, meaning it selects the proper asset class (stocks, bonds, cash) from a macro-level first, and only then is it concerned with selecting the particular fund. Next, and related to this, is the fact that it analyzes stocks and bonds using completely different criteria. While there are some similarities between stocks and bonds (i.e., they both generally perform better when interest rates are low), they are really two different animals in the end. That's why I don't use a "relative strength" system to compare the C/S/I funds to the F or G fund.

~John
 
imported post

The I FUND is acting funny again...

The EAFE was UP 0.70%

YET

The I FUND was down 5 cents.:P

What's up with that, John???

Must have been those smalllosses within the Asian markets.
 
imported post

Since there have been a few questions about being in the I Fund with the uncertain Dollar situation, et. al., here are a few comments. First, I think Mike is right. Over time, the I Fund should track the EAFE closely, but the daily valuations can be quite different, as we saw yesterday. Rod's right, too...it'sa bummer to see a 0.7% gain on the EAFE yesterday and then find out the I Fund lost a nickel! :shock:

Re: concerns with the I Fund and the Dollar vs. other currencies, I don't consider this--at least not directly. I'm just not knowleable enough. If I could really forecast the Dollar vs. Yen, Dollar vs. Euro, etc., I'd be daytrading the currency markets and making a killing. :P Seriously,some ofthe best minds and trading systems in the world are employed on the currency markets.And even if I really couldtrade currencies well (very few people can), that would only help me understand a part of what makes the I Fund tick. There's more to it than just the Dollar vs. EAFE currencies.

Instead,themodel I use to select the C, S, or I fund issimply based on theirpricemovements.It's a little complex, and uses a lotof "smoothing"to detect the long-term trend, but it's solely based on price.It's a variation of a system that others have used to successfully switch among sector funds, value/growth funds, and even U.S. and international stocks. ButI only invest in C, S or I when theoverall stock model determinesstocks are theplace to be. If conditions are too risky, I stay out of stocks completely.

When you think about it, the price of a fundhas all the relevent information: it's what all investors, money managers, trading systems, etc., think the fundis worth based on future expectations, economic trends, currency projections, etc. That's why I stay in a fund as long as the overall price trend is strong, regardless of current events or daily market "noise." Looking back, even withthe dailyhubbub over what the Dollar will do next, I'm happy to have been in the I Fund since mid-Oct. It's appreciated about 11%. The S Fund hasdonea littlebetter, and my model may switch to the S Fund very soon. But this is all about trying to find something that works well (not perfectly) over time. I've found ifyou find something that works and then overtweak it, it almost always backfires on you.

Take care, and hope everyone has a great Christmas!

John
 
imported post

I said it before and I'll say it again. I think that use of a systematic approach is the best way to "surf" the "funds". C-C-Cattcchh the wave... Ride each wave as long as you can and still avoid the undertoe.

I agree with everything you said systemtrader. I'd like nothing better than to see your system do really well.

My system is simple, I haven't had time to refine it. Not sure when I'd get the time. Which is why I've said in the past pay sites are good if the advice is good. I just can't help being a skeptic at heart. You'd think I grew up in Missouri, the Show-me state.
 
imported post

Here's a goodsystem: when you get nervous and scared, buy. When you feel comfortable and confident, sell! :^
 
imported post

Mike wrote:
Here's a goodsystem: when you get nervous and scared, buy. When you feel comfortable and confident, sell! :^
LOL!:D Yep! Sometimes you just gotta go with that GUT!!!;)

Thanx for the comments, John!:^ And BTW, your grammar is excellent!

It shows you are an educated man, which goes a long way inbringing intelligence & trustto your posts. (I hope this came across the right way)

God Bless & Happy Christmas!:)
 
imported post

FundSurfer,

I hope my system performs well, too! I realize I'mtaking a risk by posting "live signals" here, because it could perform poorly in the short term. In backtesting, there were a few lousy quarters and a fair number of mediocre quarters. It always seemed to correct itself and give a decent annual performance, however.I figureI have tostart giving real-time recommendations at some point, though, so why not be the first service to do this publicly before I ask for money?

Rod,

Thanks for your comments as well.



Hope everyone has agood week and greatnew year!

John
 
imported post

SystemTrader wrote:
I hope my system performs well, too! I realize I'mtaking a risk by posting "live signals" here, because it could perform poorly in the short term.
I figureI have tostart giving real-time recommendations at some point, though, so why not be the first service to do this publicly before I ask for money?
Are you still staying 100 % I ??
 
imported post

As I said in my "Account" folder, I'm still100% in the I Fund. I don't know if its streak will continue, but it's up about 3.89% since I first posted two weeks ago. It even jumped ahead of the S Fund as the top performer forthe year and4th quarter, so I'm glad my "fund selection model" has continued to favor it.Whenever the S Fund comes close to overtaking it in the model,it's like theI Fund starts hearing footsteps and really kicks itin for a few weeks.:D

Speaking of the stock market, I agree with Tom that there a few danger signs. The stock composite model I use is still on a buy signal, but themodels/indicators that compriseitdon't look quite as healthy as they did in November and early December. In addition, all 4 of the major surveys of investor optimism are at extremes according to SentimentTrader.com. Bulls outnumbered bears every week in 2004 in theoldest of thesesurveys (Investor's Intelligence), which is quite unusual.

In fact,this hasonly happened 7 times since Investor's Intelligence began, and 5 of those 7 years were followed by double-digit negative returns.

But I'll followthe system,of course, until I get a sell signal for stocks or a signal to switch to a different stock fund.

Happy new year,

John
 
imported post

Hi, Rod.



Nope, I'm still 100% in the I Fund. If I switch among the 3 stock funds, it will only be on a weekly basis (based on each Friday's closing price). I'll post the recommendation before Monday's opening. However, I could pull out of stocks entirely and move to the F or G Fund during the week. If this happens I'llpost this before the next morning's opening.



Here's a brief recap of my method. I control risk through the stock timing model, the first "tier" of my system. When it says to sell, I'll pull out of stocks completely. Again, this could change during the middle of the week, as the model uses both daily and weekly data. The second tier of the system only comes into play on a stock "buy" signal. Its purpose is to determine the best performing stock fund over the long term. It gives more significance or "weight" to recent price changes, however, to prevent being "locked" into a deteriorating fund for too long.



I realize that following a system like this may seem difficult or counterintuitive at times. It can almost be like buy-and-hold investing, as it's not always easy to stick witha plan and ignore daily news and market fluctuations.However, an effectivemechanical system should get you outduring a significant downturn, unlike buy-and-hold. A buy-and-hold stock fund investor will stay fully invested no matter what, even through something like the '29 crash/Great Depression, the '73-74 bear market, or the 2000-2003 market. Historically, it can takeover adecade--or several decades in the case of the Great Depression--to recoup your losses in real (inflation-adjusted) terms if you hold through a major bear market. No thanks!



Hope that answers your question,

John
 
Status
Not open for further replies.
Back
Top