|
PureBytes Links
Trading Reference Links
|
This has nothing to do with trying to catch highs/lows. This situation
applies to *any* system that uses limit orders. TS will report a limit order
filled when the market does not trade through it even though in real life
you would probably not have been filled in this case. Out of 500 trades
using limit orders for profit exits, in about 10 cases, the high of the day
just happens to be my exit limit price. Most likely, I would not have been
filled. In 6 of those 10 cases, I am filled the next day as the next day's
high exceeds my limit. In the other 4 cases, my stop is hit before the limit
is exceeded. Turning those 4 TS simulated winners into real life losers is
about a $2000 difference between what TS reports and what you would have
experienced in real life. That's about $10 per winner. (Only 200 winners
where the market opened under my limit and traded up to / through it during
the day). So the question is, am I gonna get less than $10 slip, *on
average*, with MIT orders?
FYI, I think one should always go through each trade and verify what TS
reports. Some things to check for:
1) Limit orders where the market hit but did not trade through your price.
TS will report this as a fill. In real life, probably not.
2) Profit targets and stops hit on the same bar. TS presumes a specific OHLC
sequence, but of course this is not how it always happens.
3) Situations where the market came within 1 or 2 ticks of your stop price.
In real life, your stop would most likely have been elected, especially in
markets where your stop is triggered by best b/a rather than an actual print
at your stop price.
SH
----- Original Message -----
From: "Michael Berger" <mberger@xxxxxxxx>
To: ".OmegaList" <omega-list@xxxxxxxxxx>
Sent: Monday, August 27, 2001 11:10 AM
Subject: Re: MIT orders
> I can't proffer any info on MIT orders, but I'd suggest any
> system that dependent upon catching highs/lows is most
> likely the result of data fitting, rather than catching
> profitable trading concepts.
>
>
> ----- Original Message -----
> From: "Scott Hoffman" <trader20@xxxxxxxxxxxxxx>
> To: "omega-list" <omega-list@xxxxxxxxxx>
> Sent: Monday, August 27, 2001 2:46 PM
> Subject: MIT orders
>
>
> > Fellow traders:
> >
> > I've developed a EOD system with profit exits using limit
> orders. I went
> > through trade by trade and determined on which days my
> limit price was
> > exactly the high of the day (for longs) or the low of the
> day (for shorts).
> > For worst case testing, I presumed I would not have been
> filled on those
> > day. On some of those cases, this caused a winner, as
> reported by TS, to be
> > turned into a loser.
> >
> > I am trying to decide whether to use MIT orders instead.
> >
> > My question is this: To those of you who have used MIT
> orders, what kind of
> > *average* slippage should I expect, in terms of ticks? I
> am not looking for
> > worst case, just average over many trades. Some have told
> me that they get
> > their price most of the time. For example, if you are
> filled at your price
> > 50% of the time and average a tick slip the other half,
> you would figure
> > your *average* slip on MITs to be a half tick per fill. In
> particular case,
> > if this were true, it would be better to use MITs because
> it doesn't take
> > many cases of a winner turned into a loser to wipe out any
> advantage of no
> > slippage.
> >
> > The markets I am interested in are: NY energies, Bonds and
> Notes, and Hogs.
> >
> > TIA,
> > SH
> >
> >
>
|