|
PureBytes Links
Trading Reference Links
|
eSignal has a Replay feature. You open a chart, set it back to the bar you
wish to start on, and click start. You can set the speed, from one bar/sec
to many sec/bar. You see the chart unfold in simulated real time. All your
indicators unfold also. Only problem is that their Paper Trade Module does
not work in replay mode. But you can do days of practice trading in a few
minutes, without seeing the hard right edge of the chart.
Their Paper trade module allows you to do simulated trading. It has an
order entry screen and you click on buy/sell, symbol and size and it tracks
the profit and loss for you. Great for practice trading in real time.
Tom
----- Original Message -----
From: "superfragalist" <no_reply@xxxxxxxxxxxxxxx>
To: <equismetastock@xxxxxxxxxxxxxxx>
Sent: Tuesday, February 08, 2005 5:49 PM
Subject: [EquisMetaStock Group] Re: backtesting of discretionary method
>
>
> I use discretionary trading systems all the time, and they're hard to
> back test.
>
> I write surrogate systems that come close to what I'm looking for,
> which gives me a baseline performance. Then I run explorations during
> different types of market conditions. I use my discretionary charts to
> decide which of the trades to take from the possible trades produced
> by the exploration. If the trades I pick meet my conditions on the
> open, I put them into another trading program I use which tracks their
> progress. I note the market conditions, the number of trades I find on
> a daily and weekly basis under the specific market conditions, the avg
> holding period for those trades, the win/lose ratio, etc.
>
> I've found that in live trading the results are similar to the ones
> I've found by using the hand evaluation method I described. I expect
> to do somewhat less than I achieved in simulated trading because live
> trading is imperfect. So far it's the best method I've found. Other
> systems tester programs are somewhat more flexible than MS, but the
> coding chore is often huge. I've found I can do the tests by hand
> faster than I can code them.
>
> This disadvantage is I can't add or delete variables to the simulation
> results easily. However, I've found that I can test them on the
> surrogate system and their impact will be similar to what I would have
> seen in simulated trading.
>
> MS needs a simulation model to sharpen the ability to trade with a
> particular system. Looking at historical charts is not a good way to
> learn to trade a system.
>
> A simulator is an upgrade MS could do without fixing one bug in the
> program. That's their favorite kind of upgrade.
>
>
>
>
>
>
>
> --- In equismetastock@xxxxxxxxxxxxxxx, "taforme" <taforme@xxxx> wrote:
>>
>>
>> I would like to "backtest" an EOD trading method intended primarily
>> for index futures, e.g. mini-DOW. The entries and exits are made
>> based on the condition of 3 indicators and 2 time frames with not
>> only the value of the indicators but the shape of the indicators as
>> seen on a displayed chart of each. I am experienced at writing
>> Metastock explorations and believe that writing an exploration
>> (system test script) that really captures all the factors used is
>> close to impossible. Hence, using the Metastock system test is not
>> an option.
>>
>> I have done substantial look-forward testing by scrolling forward
>> charts with the indicators one day at a time (EOD trading
>> envisioned) and making entry and exit decisions before scrolling to
>> see the next bar. Based on such manual testing, the discretionary
>> system looks very promising. I gave a live demonstration to a
>> limited audience of traders just to get objective reactions and got
>> extremely positive feedback. FYI - I have nothing to sell and seek
>> no money/funds from anyone.
>>
>> Finally the question.
>>
>> Does anyone have a suggestion/way to automate the transfer of
>> manually selected discretionary entry/exit dates (and prices) as
>> viewed from displayed charts in MetaStock EOD into a
>> database/spreadsheet so that statistics on a large number of trades
>> can be studied?
>>
>> For a non-automated approach, I could create an Excel spreadsheet
>> with daily data and then use entry and exit columns for each day
>> with manually entered flags based on what trade action was
>> determined from looking at the MetaStock chart for each
>> corresponding day. An Excel formula would then calculate
>> profit/loss of each trade based on the flags. Once that is
>> available, other statistics could be run on it. Of course, all this
>> woule be done by isolating me from seeing any "future" data.
>
>
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>
>
------------------------ Yahoo! Groups Sponsor --------------------~-->
In low income neighborhoods, 84% do not own computers.
At Network for Good, help bridge the Digital Divide!
http://us.click.yahoo.com/EpW3eD/3MnJAA/cosFAA/BefplB/TM
--------------------------------------------------------------------~->
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/equismetastock/
<*> 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/
|