|
PureBytes Links
Trading Reference Links
|
I think that the only way to interpret Robert's question is of how reliable
you can expect your system to work if you place market orders at the end of
a bar. the approach that I mentioned is a way to evaluate just how much
variation you might expect, and how to measure it for stocks.
Jim
-----Original Message-----
From: Roger Shepherd [mailto:mailrs@xxxxxxxxxx]
Sent: Tuesday, March 28, 2006 7:43 PM
To: omega-list@xxxxxxxxxx
Subject: Re[2]: simulation of trades w/TS Securites
Hello Omega List:
I trade a system using the ER2 russell futures on 1 minute bars. It
trades 4 or 5 times a day and issues orders to buy or sell limit at
the bid and sell at the ask at the close of the bar. The system has
made about 200 trades since mid December 2005 and only 4 or 2% did not
get a fill. Sometimes the wait for a fill can be 2 or 3 minutes but
that is unusual. Usually the fill comes in seconds. Althought the ES
doesn't bounce around as much as the ER2, my guess would be that with
a little patience you would in most cases get a fill in the ES as
well. If you are talking stocks, I don't have a clue.
Tuesday, March 28, 2006, 5:02:50 PM, you wrote:
BM> Hello,
BM> I tried limit orders in real trading, results on ES
BM> futures: On 1 minute, almost impossible to get a limit fill,
BM> usually you get filled if the trend goes against you. On 5 min,
BM> you get filled 75% of the times, but the odds are that you will
BM> perform worse than the simulated trading. If you want to get a
BM> realistec limit results, I think you should test your stratigy on
BM> 15 + minutes, otherwise, you may play around, by increasing the
BM> commission for each trade. Please note I only tested the limit
BM> system on ES, not sure if the same results goes the same on heavy
BM> traded stocks, if any one has expearianced the limit fills, please
BM> let me know.
BM> Hope this helps,
BM> -----Original Message-----
BM> From: Jim Bronke [mailto:jvbronke01@xxxxxxxxxxx]
BM> Sent: Tuesday, March 28, 2006 1:37 PM
BM> To: robert pisani; omega-list@xxxxxxxxxx
BM> Subject: RE: simulation of trades w/TS Securites
BM> I don't know of any way better than actually trading the market with it.
I
BM> have gotten better fills sometimes and worse others. The odds of a poor
fill
BM> are directly related to average daily volume for the stock and the
amount of
BM> your dollars per trade. you could probably write some code that would
look
BM> at the ATR for a 1 min bar whenever your system would make the trade.
you
BM> could then just use the worst case for that minute.
BM> that's the best I can do.
BM> Jim
BM> -----Original Message-----
BM> From: robert pisani [mailto:r.pisani@xxxxxxx]
BM> Sent: Tuesday, March 28, 2006 1:26 PM
BM> To: omega-list@xxxxxxxxxx
BM> Subject: simulation of trades w/TS Securites
BM> There are a number of execution platforms much superior to
BM> TradeStation's (TradeMaven, TradeBolt, X_Trader, etc). They all allow
BM> realistic simulation, so that you can simulate an automatic strategy and
get
BM> an idea of the kinds of fills (on limit orders,
BM> etc.) to expect in actual trading. However, in all cases if your
BM> signals are generated in TS you have to go through an API. I'm
BM> considering using TS Securities brokerage/clearing, which of course
accepts
BM> signals directly from TS, but I don't see any way to simulate there (b/c
the
BM> fills in TS backtesting are not realistic at all, TS
BM> backtesting is no help). Does anyone know how a realistic
BM> simulation of trading with TS Securities can be done?
--
Best regards,
Roger mailto:mailrs@xxxxxxxxxx
|