Ernie: Limit orders are definitely a challenge. If you can find a platform that will give you good execution of the types of entries you describe, you’ll probably have an edge over many traders. The trade-off is that it becomes more about hardware and execution issues than the strategy, and execution-related issues are largely outside of your control.
I’ll consider other platforms, such as Zen Fire or CQG, or TT, for the future. The challenge with that, from my point of view, is in committing to another platform that may not be around or be one of the better ones in the future.
Mike Bryant
Frankly Ernie, I doubt the platform has all that much to do with your
limit orders not executing. I've had this problem a *lot* over the
last 5 years and I've never been able to demonstrate that the platform
itself, even if I include how TradeStation's network routes orders,
has much to do with the root cause of the problem, but that's just my
opinion and I encourage you to really dig in and figure it out
yourself.
I will give you a couple of things to think about:
1) How 'over-optimized' are your strategy entry/exit points. There is
an option in TradeStation's simulator to fill a limit order when
touched, or when exceeded. If you change to fill it only when exceeded
how much does your equity curve change? I suspect it changes a lot.
Most of my strategies coming from Builder show this effect.
2) Where are you in the order queue? If your position in the queue is
number 700 but only 650 contracts get filled at that price then of
course you don't get filled. Any strategy that places it's order late,
or changes the order level continually, will push itself back in the
queue and thus delay when you fill. Look at market depth and try to
figure out for yourself where you are when the order is placed.
Also, there is an option to convert limit orders to market orders
after X seconds. This results in slippage but does get you in around
the price you were looking for.
HTH,
Mark
+1, your points are well worth highlighting.
I've used and coded with TradeLink with IB and also used NT with Vision Financial in my exploratory path and pursuit of better fills, reduced latency etc… the end result, TS may not be perfect and it does have it issues as you'll no doubt see in the forums however so do the others and IMO TS delivers comparable even slightly better results than the others mentioned here. Oh and run these platforms side by side on real servers in a central data facility as I do with my production trading platform running Mikes strategies.
Cheers,
Bruce.
I would just add:
1) I have seen significant differences between the backtested strategies and real time trading (haven't see this problem with Builder-generated strategies). This is a link to a discussion we had with TS people about the discrepancies: http://strategytraders.org/index.php?topic=6.0
2) Your strategy execution can be enhanced by whether or not you keep the orders on your computer or if you have them kept at TS. This can affect several aspects of execution as well as where you are in the order queue.
Question about speed how about a virtual server is this ok or is it better
to go for real server(s) ?
thanks ,
Chris
----- Original Message -----
From: "Bruce" <br...@tradingwireless.com>
To: <adaptrad...@googlegroups.com>
Sent: Tuesday, December 13, 2011 10:20 PM
Subject: Re: Alternative to TradeStation
Mark,
+1, your points are well worth highlighting.
I've used and coded with TradeLink with IB and also used NT with Vision
Financial in my exploratory path and pursuit of better fills, reduced
latency etc� the end result, TS may not be perfect and it does have it
Hello Bruce,
Question about speed how about a virtual server is this ok or is it better to go for real server(s) ?
thanks ,
Chris
----- Original Message ----- From: "Bruce" <br...@tradingwireless.com>
Sent: Tuesday, December 13, 2011 10:20 PM
Subject: Re: Alternative to TradeStation
Mark,
+1, your points are well worth highlighting.
I've used and coded with TradeLink with IB and also used NT with Vision Financial in my exploratory path and pursuit of better fills, reduced latency etc… the end result, TS may not be perfect and it does have it issues as you'll no doubt see in the forums however so do the others and IMO TS delivers comparable even slightly better results than the others mentioned here. Oh and run these platforms side by side on real servers in a central data facility as I do with my production trading platform running Mikes strategies.
Bruce, et. all,
I hope it was clear that I'm not saying TradeStation is a platform
without problems as that would be about as far from the truth as could
be and I don't want to risk Santa taking all my toys away. Far from
it, TS has lots of problems and TS as a company has a real tendency to
ignore problems in favor of developing new widgets which frustrates me
to no end. I do, however more or less agree that TS performance is
probably average if not slightly better than most. The devil you know
for the devil you don't know...
My real point was, I think, well understood and something the OP
can investigate on his own. If your strategy is changing the entry
level of a limit order every bar then you will be taking a new
position at the end of the queue at every price level, at least until
others line up behind you. There are also subtle issue having to do
with order thrashing between profit exits and stops that can have a
big effect with auto-traded strategies.
Whether someone needs to co-locate or not I wouldn't know much
about but would guess has a lot to do with what type of strategy they
are running. It's my opinion that a strategy that buys once and holds
to the end of the day won't see much advantage being co-located on the
floor of the commodity exchange, but what to I know? Strategies that
hold for just a couple of seconds would see huge improvements.
If it matters, I run Gentoo Linux on a 12 core processor here at my
home office and then run TradeStation in a Win 7 VM. (I run Builder in
a different VM.) I see _no_ negative effects vs running TS on a native
Windows machine. My belief is that it's actually more stable and more
maintainable to run this way because Windows doesn't manage any
hardware and I can back up and restore a Windows VM in literally
minutes, whereas when a Windows machine crashes it's a real drag to
rebuild the whole box.
Just some things for the OP to consider.
Cheers,
Mark