Deciding on a Backtesting and Trading Platform

Quote from fundjunkie:

No, you could use anything that can handle intra-day and eod data. that said, I generally encourage people to avoid Tradestation when they have a choice. It has too many limitations. If you browse this and other threads those limitations will become readily apparent. But hey, it might be right up your street.

By auto-execution I mean unattended triggering of position entries and closures. Part of the attraction of eod trading is you can do that manually. You couldn't do that in a "once during the day" model cos you wouldn't know when the entry might get triggered.

Not from me, I've not done any backtesting for quite a while. I'm in research mode and writing code. I use Tradersstudio and Murray ruggiero is often around here and should be able to address your query directly. he would be able to explain the "virtual backtester" too which would help with what I "think" you'd like to do. I don't want to explain it because although I'm aware of it I have not used it.

For your information, I believe a virtual backtester would help where you have a conventional eod system but oerlay a metastrategy that uses intra-day data. So there are three variations I've touched upon:

1) EOD
2) Once during the day (EOD style system with intra-day entry/exit)
2) EOD with intra-day meta-strategy overlay.

Be aware that when I talk about intra-day execution I don't mean day trading but rather the capability to signal entries and exits intra-day with a system who's holding period might average one or two weeks. However, what i added to that is that ideally one would have systems traded in harmony modelling different market regimes and timeframes (including day trading systems) in order to find an overall more robust approach.


Thx
D

Thanks fundjunkie, that was interesting. Shall ask Murray for the same.

Read about back-testing, it looks that it is a kind of "Technical Analysis". But, so many people are against TA. :confused: What does it say about TA & back-testing?

Please explain. Thanks! :)

Murray ruggiero, please give examples of back-testing. Thanks! :)
 
Quote from droskill:

Murray - can you provide details on what virtual backtesting is? I'm a Tradersstudio user and I don't know what it is but I've heard it is interesting!

I'm a user of Amibroker and Tradersstudio - I think they're both great tools but are different - AB is great prototyping and TS for doing more detailed analysis.

Multichart I've never found compelling - and Tradestation is cool but too expensive. I really like Wealth-lab but their current software strategy is a loser since I'm US-based.

Another interesting one: Seer.

What I mean by virtual backtesting is having two different backtesting channels one which records the backtested results using one set of buy and sell statements and another set of them which have the statisics available but the results are not used in the standard backtesting reports. This way you can use the virtual one to filter the set of rules used in the backtest, for example; skip a trade if the last trade was a large winning trade.

This can only be done by have two sets of backtested signals one which is unfiltered and another which is filtered by the unfiltered one.
 
I love Multicharts from what i've seen so far, there is just one significant issue I have. Any way to disable expansion of chart range so I can exclude analysis techniques from chart scaling?

An example of what i mean is included in screenshot

I can do this operation in Tradestation but not in MC.

Thanks!

<img src=http://elitetrader.com/vb/attachment.php?s=&postid=2016921>
 

Attachments

Quote from iluvMC:

I love Multicharts from what i've seen so far, there is just one significant issue I have.

Hmmn, I wonder whether you've posted here by mistake. This is not a Multicharts support thread. I recommend that you go to the Multicharts support forum on their website (it is quite active), contact their support people directly or, alternatively, seek out the Multicharts users thread on this site where you'll find MC using peers with whom to discuss your issues.

Up to now, this thread has been generally sane without too much off topic "noise" pollution. It will be good if it can stay that way.


Thx
D
 
Can you provide an example so I get what you're talking about?

Quote from Murray Ruggiero:

What I mean by virtual backtesting is having two different backtesting channels one which records the backtested results using one set of buy and sell statements and another set of them which have the statisics available but the results are not used in the standard backtesting reports. This way you can use the virtual one to filter the set of rules used in the backtest, for example; skip a trade if the last trade was a large winning trade.

This can only be done by have two sets of backtested signals one which is unfiltered and another which is filtered by the unfiltered one.
 
Quote from droskill:

Can you provide an example so I get what you're talking about?

Code:
'********************************
' Channel Breakout  Filtered by Equity of Channel Breakout
' This system trades a virtual channel breakout which takes all trades and filters the 
' performance of the actual live system. It only takes a trade if the previous trade
' for the virtual system  lost money
' TradersStudio(r) 2004 , All rights reserved

Sub ChannelWithFilters(SLen as integer)
Dim MinMove
Dim  TrProfit
Dim LowPrice as double
Dim HighPrice as double
LowPrice=Lowest(low,SLen,0)
HighPrice=Highest(high,SLen,0)
MinMove=GetActiveMinMove()

' Take a trade for the virtual system 
VirtualBuy("ChanBuy",1,HighPrice+MinMove ,STOP,DAY)
VirtualSell("ChanSell",1,LowPrice-MinMove,STOP,DAY)

' If the virtual system has at least two trades
if VirLiveTradeCount()>2 then  

' Trade 0 is the last completed trade, Trade 1 would be two closed trades ago
    TrProfit=VirTradeProfit(0)

' If last trade in virtual system lost money then trade a real trade
   if TrProfit<0 then 
Buy("ChanBuy",1,HighPrice+MinMove ,STOP,DAY)
Sell("ChanSell",1,LowPrice-MinMove,STOP,DAY)
    else

' Just exit the old trade and don't start a new one 
ExitLongallpos("XBuyBrk","ChanBuy",LowPrice-MinMove,STOP,DAY)
ExitShortallpos("XSellBrk","ChanSell",HighPrice+MinMove,STOP,DAY)   
  end if 
end if 
end Sub
 
Quote from Derrick1983:

Thanks fundjunkie, that was interesting. Shall ask Murray for the same.

Read about back-testing, it looks that it is a kind of "Technical Analysis". But, so many people are against TA. :confused: What does it say about TA & back-testing?

Please explain. Thanks! :)

Murray ruggiero, please give examples of back-testing. Thanks! :)

Backtesting is doing a historical look at the performance of a given set of rules. It has good points and bad points because you can change the rules over and over again to produce good results. This is called Curve fitting. Curve fitted systems don't perform well in the future, which is important because you can't trade the past.

In terms of technical analysis most classic rules in books don't work. For example buying when RSI crosses from below 30 to above 30. They type of rules don't produce robust results.

Most technical system that do well are trend following breakout or pattern based systems.
 
yo [XXX Deleted by Moderator],

I am EVALUATING MultiCharts. I am "Deciding on a Backtesting and Trading Platform". Get a clue.

Do you work for Tradestation or some other competitor and trying to box out questions directed at other vendors?

iluvMC
(soon to be changed to MCsux)
 
For those into switching to Multicharts, I would recommend against it based on my research.

There are some serious usability issues in Multicharts. Its a bagful of features - including seconds charts and many others, however basic usability is quite inferior to Tradestation 8.3.

Multicharts has a chance to be the leading charting software a couple of years from now but not for now (considering how disinterested they appeared to be at resolving basic usability issues).

A subset of the issues:

1. scaling issue (see previous post). This was a deal breaker for me.
2. you will lose your workspace if you try saving to a non Windows compliant name. No warning.
3. highlighting price on right axis based on plots can only be done gloabally, not on an individual plot basis.
4. Sessions can be only applied on a per symbol basis
etc...

Cheers,
MCsux
 
Back
Top