jpcavin's Account Talk

[jpcavin]

I'm not good at all that back testing stuff. Was hoping that someone else could tested going forward from 2004 to present. Per the book http://ptgmedia.pearsoncmg.com/image...79024_ch03.pdf , "This is a short-term hit an drun timing model that was invested 45.9% of the time from 1972 to May 2004, yet it outperformed buy-and-hold strategies during 20 of the 32 years included in the study. Gains from winning trades were more than five times the size of losses taken during losing trades. You can look at the performance or the Triple Momentum Timing Model as compared to Buy and Hold, Nasdaq Composite on Pages 60-61

Summary of Performance Results for Buy and Hold vs Triple Momentum

Gain per annum 9% vs 19.8%
Open drawdown -77.4% vs -17.5%
Rate of gain while invested, annualized 9% vs 43.1%

Round trip trades 288 (8.9 per year)
Percentage of trades profitable 54.4%
Percentage of time invested 45.9%
Average gain per profitable trade 4.8%
Average loss per losing trade -0.9%
Gain/loss per trade ratio 5.3
Total gain/loss ratio 6.2

This timing model has shown consistend performance relative to buy-and-hold over the past three decades. During the 1990s, there were five years in which buy-and-hold strategies outperformed Triple Momentum, and five years in which the model outperformed buy-and-hold. Keep in mind that this system is only invested 45.9% of the time.

The book can really explain all this better than I can. I urge anyone interested to read pages 59-63 for clarification.

 

[JTH]

I'm not good at all that back testing stuff. Was hoping that someone else could tested going forward from 2004 to present. Per the book http://ptgmedia.pearsoncmg.com/image...79024_ch03.pdf , "This is a short-term hit an drun timing model that was invested 45.9% of the time from 1972 to May 2004, yet it outperformed buy-and-hold strategies during 20 of the 32 years included in the study. Gains from winning trades were more than five times the size of losses taken during losing trades. You can look at the performance or the Triple Momentum Timing Model as compared to Buy and Hold, Nasdaq Composite on Pages 60-61

Summary of Performance Results for Buy and Hold vs Triple Momentum

Gain per annum 9% vs 19.8%
Open drawdown -77.4% vs -17.5%
Rate of gain while invested, annualized 9% vs 43.1%

Round trip trades 288 (8.9 per year)
Percentage of trades profitable 54.4%
Percentage of time invested 45.9%
Average gain per profitable trade 4.8%
Average loss per losing trade -0.9%
Gain/loss per trade ratio 5.3
Total gain/loss ratio 6.2

This timing model has shown consistend performance relative to buy-and-hold over the past three decades. During the 1990s, there were five years in which buy-and-hold strategies outperformed Triple Momentum, and five years in which the model outperformed buy-and-hold. Keep in mind that this system is only invested 45.9% of the time.

The book can really explain all this better than I can. I urge anyone interested to read pages 59-63 for clarification.

I threw the system over the Transports, S&P 500, and Wilshire 4500 last night, using a buy/sell trigger when all three indexes hit <>4 in unison. I only tested the last 2 years.

A few observations:

It's a trending system, it works great for long up/down trends it's a long-term system.

Within an oscillating market this system appears to perform rather poorly, getting whipsawed, losing small amounts on each false signal.

I don't know how long you've been reading this forum, I will tell you I've done a fairly extensive amount of backtesting with various systems. I'm not telling you that makes me quallified but I do feel it gives me a broad & basic understanding of various TA systems.

My observations are this. We are not in the same market we were 30 years ago, if you've noticed we have larger intraday ranges than we did in the past, this is due to faster money moving in & out of the markets at the speed of electrons. This type of system is more easily manipulated under these conditions, it's going to naturally produce more false signals. Smoothing out the 3 ROCs may dampin the number of false signals, but it may do so to the point where the enty/exits are too late to the game, thus underperforming the B&Hers.

I realize those backtested results are in a book, but this doesn't mean they will work well under TSP's 12-noon cutoff EOB price restrictions. My own personal rule of thumb, if I can't backtest it myself than I don't trust the results, JMHO.

 

[jpcavin]

I agree with you JTH. And the book clearly states that times change and systems need to be adjusted and forward tested and smoothed, etc. One of the forums I referred to in my post actually discusses using QQQ instead and smoothing for noise, etc. I'm sure it can be backtested like everything else. I can't back test it because I don't know how.

But like I have stated many times...I don't know much of anything about the markets. I understand terminology but I can't put it all together. And when I see it on a graph, I don't see anything at all. I understand numbers....Looking at the spreadsheet, I can clearly see negative direction and posite direction, even if it's erratic. I can see when the Bond yield is down. I can see if the NYSE or NASDAQ is leading. I can't see that on a graph. I know you all can, but I can't. I can see the fake sell's and the fake buys as well. Gotta work with what I can understand right now and this is as basic as it gets.

I've shared what I have found not only for review by more knowledgeable members but also for those that know nothing and want something simple until they get their learning curve going. Anyone can take it or leave it. Would have liked seeing someone with more know-how tweak it.

I do appreciate your opinion because that is exactly what I wanted to know. What type of markets does this system work best in? Is this system worth looking into with some adjustments or not? etc, etc. I don't know how nor what adjustments to make to fit this market...

Anyway, I'm only on chapter 4. With a few more chapters and a few more books, I may be able to graduate from a basic understanding to semi-intermediate understanding of technicals.

 

[jpcavin]

I AM IN THE WRONG FUND!!! I heard a couple of times that currently small caps were weaker than large caps. I always wondered how to choose between C and S Fund. I didn't understand until Uptrend explained it. A simple RSI chart that compares the S&P and the W4500 would have helped me make the right choice. Lesson learned. Damn it, I need to go get me some ice cream!

 

[JTH]

I'm sorry I can't be more help, with what I have, certian things are easy to backtest, but I don't have the capability to backtest a cumulative sum from 3 seperate ROCs, everything would have to be handjammed and that takes a great deal of time. You may find (as I have) that when developting a system, your imagination isn't the limiting factor, the tools you have to work with are. If you like to work with numbers then maybe creating excel formulas will work for you, it's all a matter of what you know and what you're willing to learn, Some are willing to do the work while others would rather ride the coat-tails of others. I'm sure the milkman knows exactly what I'm talking about

 

[JTH]

I AM IN THE WRONG FUND!!! I heard a couple of times that currently small caps were weaker than large caps. I always wondered how to choose between C and S Fund. I didn't understand until Uptrend explained it. A simple RSI chart that compares the S&P and the W4500 would have helped me make the right choice. Lesson learned. Damn it, I need to go get me some ice cream!

What you might be looking for is a price performance chart where you pick a timeframe and see which chart performs better (percentage wide) over the same period of time.

Here are two examples over a 6-month period, the first one shows them side by side, the second is a direct comparison shown as 1 line.

View attachment 15184
http://stockcharts.com/h-sc/ui?s=$SPX&p=D&yr=0&mn=6&dy=0&id=p73872750144


View attachment 15185
http://stockcharts.com/h-sc/ui?s=$SPX:$EMW&p=D&yr=0&mn=6&dy=0&id=p15045636864

 

[jkenjohnson]

I'm sorry I can't be more help, with what I have, certian things are easy to backtest, but I don't have the capability to backtest a cumulative sum from 3 seperate ROCs, everything would have to be handjammed and that takes a great deal of time. You may find (as I have) that when developting a system, your imagination isn't the limiting factor, the tools you have to work with are. If you like to work with numbers then maybe creating excel formulas will work for you, it's all a matter of what you know and what you're willing to learn, Some are willing to do the work while others would rather ride the coat-tails of others. I'm sure the milkman knows exactly what I'm talking about

Do you have a model that you developed yourself that you use to make trades, or do you use a developed model with trends? I don't think I have read what kind of model you utilize, if at all.

 

[nnuut]

I just compare the 2 charts at the bottom of the page, and do it daily. EASY PEASY!:laugh:

 

[Happy_Trails]

I AM IN THE WRONG FUND!!! I heard a couple of times that currently small caps were weaker than large caps. I always wondered how to choose between C and S Fund. I didn't understand until Uptrend explained it. A simple RSI chart that compares the S&P and the W4500 would have helped me make the right choice. Lesson learned. Damn it, I need to go get me some ice cream!

I wish that I could pick the wrong funds as well as you! Seriously though, the trending models like you followed last month did incredibly well during August. Another one is laid out in the book "Buy Don't Hold" by Les Masonson. His timing system sold right around the end of July and just got back in a few days ago.

 

[JTH]

Do you have a model that you developed yourself that you use to make trades, or do you use a developed model with trends? I don't think I have read what kind of model you utilize, if at all.

I do have timing models I use, but I don't base my trades on those factors alone, some of it is also based on timing, resources, and trends. I am ever evolving and still in refinment, I get smarter with each mistake

 

[Happy_Trails]

What is $compq? is that the NASDAQ? I look at .IXIC or ^IXIC (google or yahoo).
Based on the NASDAQ Composite close today at 2579.46 the 5d ROC=4.53, 15 Roc=8.33 and 25d ROC=-6.70. Composite=6.16

The 15 day ROC appears to have dropped a lot by today's close, around 2.14. This drop appears to have flipped the system to a sell. Is that what your readings show, JP?

 

[jpcavin]

Yes. The market is really erratic so this system would not help anyone right now. At least not for the TSP IFT limitations. This system is based on closing prices so technically you would buy or sell the next day based on the previous day's closing composite. Using the last few days as an example...

Buy on the 30th based on composite on 29th of 8.04% (w4500=622.63)
Sell on the 31st based on composite of -0.20% on the 30th (w4500=623.49)
Buy on the 1st based on composite of 6.16% on the 31st (w4500=611.71)
Sell on the 2nd based on composite of -0.60% on the 1st (w=????) Hopefully it closes higher than 611.71???

Also, for the long term, if you wait until the composite is back to under 4% to get out, you would have given up a lot of the gains. I like to use this system as an entry point and then pick my exit. And the entry point carries some risk as well. But to me, it's better than having no system.

 

[FAB1]

Yes. The market is really erratic so this system would not help anyone right now. At least not for the TSP IFT limitations. This system is based on closing prices so technically you would buy or sell the next day based on the previous day's closing composite. Using the last few days as an example...

Buy on the 30th based on composite on 29th of 8.04% (w4500=622.63)
Sell on the 31st based on composite of -0.20% on the 30th (w4500=623.49)
Buy on the 1st based on composite of 6.16% on the 31st (w4500=611.71)
Sell on the 2nd based on composite of -0.60% on the 1st (w=????) Hopefully it closes higher than 611.71???

Also, for the long term, if you wait until the composite is back to under 4% to get out, you would have given up a lot of the gains. I like to use this system as an entry point and then pick my exit. And the entry point carries some risk as well. But to me, it's better than having no system.

What did you do with the real JP? Wheres the Pod?

 

[jpcavin]

What did you do with the real JP? Wheres the Pod?

What are you talking about? I was never JP anyway. Judy, Judy, Judy!

 

[jpcavin]

Ok, for any mods reading this, can I change my account name from jpcavin to JudyBreeze or am I stuck?

 

[burrocrat]

I'll start calling you JudyBreeze, hell I'll call you anything you want, if you send me that coffee cup.

You can always go get another one.

 

[jpcavin]

Which coffee cup do you want. I'll try to make the top 10 again. :laugh: They say history repeats itself. We'll just have to wait and see.

 

[jpcavin]

I'm sorry I can't be more help, with what I have, certian things are easy to backtest, but I don't have the capability to backtest a cumulative sum from 3 seperate ROCs, everything would have to be handjammed and that takes a great deal of time. You may find (as I have) that when developting a system, your imagination isn't the limiting factor, the tools you have to work with are. If you like to work with numbers then maybe creating excel formulas will work for you, it's all a matter of what you know and what you're willing to learn, Some are willing to do the work while others would rather ride the coat-tails of others. I'm sure the milkman knows exactly what I'm talking about

Regarding the back testing....
What do I have to do or what can I do. I have all the numbers going back to 1971 or 1972 (not writing from my computer right now). I am refering to the NASDAQ Composite. I have all the closing prices back to the 70's. I have to results of all 3 separate ROCs and I have the cumulative figures for all three. Everything is on an excel spreadsheet. What other information do I need to get? I'm willing to do backtest or at a minimum to all the legwork to get the data needed. Just need someone to tell me what to do one step at a time. How about it? What's the first step

 

[burrocrat]

Which coffee cup do you want. I'll try to make the top 10 again. :laugh: They say history repeats itself. We'll just have to wait and see.

a traveller, like this one, or else whatever is available with the most stainless steel and the least rubber grips or other useless stuff, that looks nice on your desk and in a car.

View attachment 15198

Thanks JudyBreeze,

donkey.

p.s. I looked for the TSPTalk store button but couldn't find it, had to search a username's posts and found the link in the monthly winners thread finally.

 

[JTH]

Judy, I wrote you a long reply, then it got lost when my internet crashed. I've seen many systems come and go, some of them I devised then got bored with. When it comes down to brass tacks whatever lets you sleep well at night is what works best. I gave you an honest assessment of the Triple ROC based on how I think it would fair within the TSP world. If you were so inclined you could perhaps figure out a way to combine the TSP returns calculator with the triple ROC using the combination of the 2 to backtest your results. I'm by no means an Excel guru, in fact it's often a very painful experience when I'm trying to write formulas, as I've mentioned before, my imagination exceeds my ability to use the tools I have to work with. This is going to sound corny but it's true, all indicators do is interpret prices, the best indicator you can use is the one between your ears...