[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [EquisMetaStock Group] TrendMedium



PureBytes Links

Trading Reference Links

Super,

Thanks for your evaluation on Trendmedium while giving us, the newbies, an insight into system testing.  I'm new to the metastock group and missed your posting on your trading method.  Is there a link where I can access and learn about your trading method?

John



----- Original Message ----
From: superfragalist <no_reply@xxxxxxxxxxxxxxx>
To: equismetastock@xxxxxxxxxxxxxxx
Sent: Friday, July 20, 2007 12:35:55 AM
Subject: [EquisMetaStock Group] TrendMedium

>From what I've seen, TrendMedium is short term algorithm based
optimization program. It's composed mostly of a set of moving averages. 

Essentially a user runs the optimization routine frequently and
generates new buy and sell signals based on the current optimized values. 

I tested TrendMedium like I have almost all Equis products, not
because I wanted to trade with it but because I wanted to understand
how it worked. I tested it along time ago. It looks like Eugene has
added a lot of features to it since I tested it. 

It's a curve fitting program. Curve fitting has a bad reputation but
can be very effective under certain circumstances such as when history
repeats itself. The closer the current price action is to the segment
the curve fit was performed on, the more predictability the optimized
equation has. 

Trendmedium is looking for stocks that have a price pattern that tends
to repeat itself somewhat so the optimized equation that makes up it's
expert works reasonably well. 

That's my take on it. I don't know if Eugene would agree. I suspect he
might want to put some lace on what I've said, but that's mostly it.

My impression was it trades a lot. It picks up a lot of really short
term gains with quick flip trades, and it seemed to work okay for the
most part. 

It didn't suite me. The method I use for trading was described in
Roy's newsletter. It's simple, makes a lot of money and doesn't take a
lot of time. (I spend a lot of time on trading because I like to look
at a lot of things, but picking the stocks to trade using the method I
detailed in the newsletter is not very time consuming.) While it's not
risk free, it's consistent and once someone has experience with it,
it's easy. So if I am going to change how I trade then the system
would have to be really, really good. TrendMedium wasn't my cup of
tea, but it might be yours depending on how you like to trade. 

Anyway, if you buy it from Equis you can try it for 30 days and return
it if you don't like it. Equis is really good about taking stuff back. 

As far as Eugene's post being spam, that's a bogus bunch of crap.
Eugene has a legitimate MS add-on that at least works. Just because he
doesn't post a bunch of meaningless code that's never going to make
anyone any money doesn't mean he shouldn't get to announce his product
or talk about it. I don't get the part about him "not going away". Why
should he go away?

Unfortunately I don't know what Eugene meant by his test results not
being back tests. They're walk forward tests which I and a lot of
other people use in system testing. A walk forward test is a rotation
estimation test. Simply put the historical data is divided into
subsets. An opt routine is run on one sample, and then the maxed
constraints are tested on the other data subsets to see how well the
curve fitted equation from the opted data predicts the behavior of
non-opted data subsets. I generally divide my data into four or five
parts. I test a 5 or 10 year period and then apply the equations to
the untested data in six month rolling forward data sets. In other
words, 6 months, 12 months, 18 months. I like to make sure I've
included as many different market situations as possible. 

If the equation does a good job on the walk forward data, then the
curve fit of the equation is rational and has not been distorted by
either the uniqueness of the subset or the noise in the data.
Therefore the equation is likely to give good results at least in
general for some unknown period of time. 

If someone wants to get mathematically precise, K-fold cross
validation can be used on the data, which gives the most accurate
estimation of an equations likely performance in the future. K-fold is
almost the same as what I described but more specific about regarding
the data subsets, the number of tests and order of the tests. No big
deal. 

So I guess Eugene describes his tests as walk forward tests but not
back tests. May be I missed something in the description but it sure
looks like back testing to me. May be it's a translation issue. 

Nothing replaces actual results. Too bad there aren't any of those. At
least none I know about. 

As far as any plug-in or add-on for MS goes, I've never used one or
heard of one that works very well. I guess that's why they don't come
with test results (I've seen a few weak attempts where there are test
results over very specific time frames such as in a total uptrend.
Remember everything works good in an uptrend. Even my dog made good
stock picks over the last couple of months.)

I've read the argument that this or that plug-in or add-on is a tool
and the user has to come up with a system to use the tool effectively.
Okay, well, if that's the way you want to go, no problem. 

Basically from my experience trading is about having a strategy, a
simple chart that gives the trader some data on which to determine if
this is a good time to apply the strategy and enough sense to make
reasonable decisions on getting in and getting out. I think that
summarizes every book I've ever read. 

I just never have found a high IQ indicator (tool) that is smart
enough to make buy and sell decisions for me based on price and volume
data alone. I haven't found 3 indicators together that can do that for
me. 

Problem is if one of my buddies has an IQ of 30 and I need help making
a decision, I'm not going to ask that buddy for help. If I get two
more buddies with IQ's of 30 and ask all three of them to come up with
a joint decision, is the decision they give me based on an IQ of 90 or
an IQ of 30 collectively. 

May be for you TrendMedium has found a solution to that problem. 

Super

--- In equismetastock@ yahoogroups. com, "Wayne Matthews" <Wayne@xxx> wrote:
>
> Preston
> 
> 
> 
> Would you be able to give me any info on this plug in and how
effective it
> could be in the forex market.
> 
> Thanks
> 
> Wayne
> 
> 
> 
> _____ 
> 
> From: equismetastock@ yahoogroups. com
[mailto:equismetastock@ yahoogroups. com]
> On Behalf Of pumrysh
> Sent: Thursday, July 19, 2007 4:16 PM
> To: equismetastock@ yahoogroups. com
> Subject: Re: [EquisMetaStock Group] TrendMedium 2.75 PreRelease major
> update.
> 
> 
> 
> Since its a MetaStock Add-on I allow it.
> 
> I would hope any newcomer would be cautious about their selection of 
> add-ons. Eugene has been good about uploading the beta updates which 
> are free.
> 
> Preston
> 
> --- In equismetastock@ <mailto:equismetast ock%40yahoogroup s.com>
> yahoogroups. com, Code 2 <Code2@> wrote:
> >
> > Preston, are you the list administrator? If so, what is your policy
> > on spam?
> > 
> > This guy just won't go away and I'm concerned for the newbies here.
> > 
> > 
> > 
> > From: Eugene Labunsky <elabunsky@>
> > To: equismetastock@ <mailto:equismetast ock%40yahoogroup s.com>
> yahoogroups. com
> > Date: Thursday, July 19, 2007, 5:17:05 AM
> > Subject: [EquisMetaStock Group] TrendMedium 2.75 PreRelease major 
> update.
> > 
> > Dear Traders,
> > 
> > Update TrendMedium build, please. GUI & MSX DLL fixed (Added wave 
> profit
> > lost control, fixed GUI memory leak bug, increased speed of MSX DLL
> > (MetaStock show calculated system faster)). Download it:
> > http://www.trendmed
> <http://www.trendmed ium.com/download /trendmedium275- setup.exe>
> ium.com/download/ trendmedium275- setup.exe
> > PreRelease web page: http://www.trendmed
> <http://www.trendmed ium.com/prerelea se/> ium.com/prerelease/
> > 
> > PS. Estimate TrendMedium perfomance for QQQQ 1725 trading day 
> simulation
> > smilar with real life using of TrendMedium / paper trading or WFT - 
> this is
> > not BACKTEST!.
> > http://www.trendmed <http://www.trendmed ium.com/wft/ LongTest. htm>
> ium.com/wft/ LongTest. htm (equity chart)
> > http://www.trendmed <http://www.trendmed ium.com/wft/ SystemReport. txt>
> ium.com/wft/ SystemReport. txt (system's report)
> > 
> > -- 
> > Regards,
> > Eugene Labunsky.
> > http://www.TrendMed <http://www.TrendMed ium.com> ium.com - MetaStock
> Add-on
> >
> 
> 
> 
> 
> 
> [Non-text portions of this message have been removed]
>




[Non-text portions of this message have been removed]



 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/equismetastock/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/equismetastock/join
    (Yahoo! ID required)

<*> To change settings via email:
    mailto:equismetastock-digest@xxxxxxxxxxxxxxx 
    mailto:equismetastock-fullfeatured@xxxxxxxxxxxxxxx

<*> To unsubscribe from this group, send an email to:
    equismetastock-unsubscribe@xxxxxxxxxxxxxxx

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/