I've started testing some of these other ETF's IWM, and QQQ for
next bar open trades for the reversion to mean systems - instead of
the new entries occurring at the close, new entries would be taken
the next day on the open - which as I described before makes it
easier to communicate these and avoid the mad dash end of day
scramble.
The initial entries to the systems are fine, no problem. It's
the 2nd and 3rd entries that I ran into an obvious issue that I
hadn't considered but I think I came up with a solution.
Here's where you get issues. The 2nd and 3rd entries are
generally based on an X% better price than the 1st entry or 2nd
entry. The PROBLEM is, after your 1st entry - let's use of
the short systems as an example. For the standard SPY and ES end of
day systems - if you enter a short, and let's say price closes 2 or
3% higher than the 1st, then the system will enter a 2nd entry at
that closing bar. Well, the problem arises from entering these 2nd
or 3rd entries on the next day's open.
Let's say you entered a 1st entry short, and price rallied over
a few days and was now 3% higher and the system called for a 2nd
entry to that short. Well, what can happen if you take the 2nd
entry short on the next day's open is that price could
theoretically gap down and be LOWER than where you entered the 1st
entry. This doesn't happen often but in my testing I see
examples of it - see the attached image.
So....a solution to this would be: for 2nd and 3rd entries,
instead of simply entering them on the next day's open with a
market order, do so with a limit order i.e. price short IWM at
>=178, so if price opened at 177 and sold off you would not
fill.
the imaged shows examples of the IWM HI Mid Low short systems -
there's a couple examples here showing this weird behavior of where
a 2nd or 3rd entry short could occur at lower (worse instead of
better) prices than the previous entry by waiting for the next day-
then entered a 2nd entry the next bar but at lower prices. While
you make more money on that, it doesn't make sense for what these
2nd and 3rd entries are meant to be based on i.e. better prices.
The limit will prevent stuff like that
anyway nerd stuff - but that's the kind of stuff you deal with
in system building
Posted by timebandit on 19th of May 2022 at 04:22 pm
Matt, I've been thinking about the sytems you are testing for
sector ETFs, and I wonder if the KISS system could work in tandem
as it does for the market ETFs like SPY, etc. Could the buy and
sell signals of the KISS system be used to take advantage of sector
rotation?
Yeah I like the marriage of the two for sure. However he
KISS stuff is not as clear cut yet as the reversion to mean
systems. The reversion to mean systems are cut and dry like a
calculator, 6x6=36, it can never equal 38 or 35. The KISS stuff has
can have some slight subjectively to it, it's more of a methodology
I consider it 90% there with more refinement I'd like to do.
I don't need need a bunch of complication, and the reversion
to mean systems are cut and dry so easy for me to just say here's
the trade and go. Maybe one day we'll do that KISS ETF combo
as well
Posted by timebandit on 19th of May 2022 at 08:41 pm
Oh, yeah. Makes sense. It’s already tons of work to get the
systems tested and up and running. And here I am asking for more!
Will definitely appreciate anything you can offer regarding
sectors
you are saying to just have all the systems execute at the close
like the SPY instead of bothering with next day open, let them
trigger at the close - then sometime after the market email out
saying IWM, and QQQ took a trade etc. Because from a logistics
standpoint I couldn't issue the trades for all of these like I do
with the SPY. For me after hrs is no big deal - heck most of
those trades I enter in after hrs anyway. But the issue here
is that the trades would always be well after the market
close maybe 15 and 30 min or longer because I wouldn't keep the
workspaces open during the day for stability, and then would open
them all up after hrs to see what triggered.
The point here is that all the signals would be well after the
close. For you guys that's really no big deal (except for options
of course) - but if I tried to market this to a wider audience,
there's a lot of people that wouldn't like that every one of the
trades would have to be executed in after hrs - especially if it's
all the trades. I was thinking that the next day open might have
more appeal to a wider audience. And, it's less work on me because
I don't need to have an urgent need to try and confirm all those
trades after the close - I can leisurely do it - look them over,
maybe run a quick backtest to test something if I wanted to - then
send the trade notification email out that evening.
Posted by focus175 on 19th of May 2022 at 06:44 pm
I love the chance to be able to utilize the systems for playing
IWM and the Qs in the future. I'm not able to trade the first
several hours of the day, so if you're able to send out the
notification for those in the evening the day before like you
mentioned, that'd definitely be helpful.
Newsletter
Subscribe to our email list for regular free market updates
as well as a chance to get coupons!
The new systems next bar open trades - a unforeseen complication
Posted by matt on 19th of May 2022 at 04:03 pm
I've started testing some of these other ETF's IWM, and QQQ for next bar open trades for the reversion to mean systems - instead of the new entries occurring at the close, new entries would be taken the next day on the open - which as I described before makes it easier to communicate these and avoid the mad dash end of day scramble.
The initial entries to the systems are fine, no problem. It's the 2nd and 3rd entries that I ran into an obvious issue that I hadn't considered but I think I came up with a solution.
Here's where you get issues. The 2nd and 3rd entries are generally based on an X% better price than the 1st entry or 2nd entry. The PROBLEM is, after your 1st entry - let's use of the short systems as an example. For the standard SPY and ES end of day systems - if you enter a short, and let's say price closes 2 or 3% higher than the 1st, then the system will enter a 2nd entry at that closing bar. Well, the problem arises from entering these 2nd or 3rd entries on the next day's open.
Let's say you entered a 1st entry short, and price rallied over a few days and was now 3% higher and the system called for a 2nd entry to that short. Well, what can happen if you take the 2nd entry short on the next day's open is that price could theoretically gap down and be LOWER than where you entered the 1st entry. This doesn't happen often but in my testing I see examples of it - see the attached image.
So....a solution to this would be: for 2nd and 3rd entries, instead of simply entering them on the next day's open with a market order, do so with a limit order i.e. price short IWM at >=178, so if price opened at 177 and sold off you would not fill.
the imaged shows examples of the IWM HI Mid Low short systems - there's a couple examples here showing this weird behavior of where a 2nd or 3rd entry short could occur at lower (worse instead of better) prices than the previous entry by waiting for the next day- then entered a 2nd entry the next bar but at lower prices. While you make more money on that, it doesn't make sense for what these 2nd and 3rd entries are meant to be based on i.e. better prices. The limit will prevent stuff like that
anyway nerd stuff - but that's the kind of stuff you deal with in system building
Matt, I've been thinking about
Posted by timebandit on 19th of May 2022 at 04:22 pm
Matt, I've been thinking about the sytems you are testing for sector ETFs, and I wonder if the KISS system could work in tandem as it does for the market ETFs like SPY, etc. Could the buy and sell signals of the KISS system be used to take advantage of sector rotation?
Yeah I like the marriage
Posted by matt on 19th of May 2022 at 06:18 pm
Yeah I like the marriage of the two for sure. However he KISS stuff is not as clear cut yet as the reversion to mean systems. The reversion to mean systems are cut and dry like a calculator, 6x6=36, it can never equal 38 or 35. The KISS stuff has can have some slight subjectively to it, it's more of a methodology I consider it 90% there with more refinement I'd like to do. I don't need need a bunch of complication, and the reversion to mean systems are cut and dry so easy for me to just say here's the trade and go. Maybe one day we'll do that KISS ETF combo as well
Oh, yeah. Makes sense. It’s
Posted by timebandit on 19th of May 2022 at 08:41 pm
Oh, yeah. Makes sense. It’s already tons of work to get the systems tested and up and running. And here I am asking for more! Will definitely appreciate anything you can offer regarding sectors
Why go through this. There
Posted by jlevinthal on 19th of May 2022 at 04:12 pm
Why go through this. There is plenty of time in the aftermarket after the cash and futures close. Except for the option trades
you are saying to just
Posted by matt on 19th of May 2022 at 06:05 pm
you are saying to just have all the systems execute at the close like the SPY instead of bothering with next day open, let them trigger at the close - then sometime after the market email out saying IWM, and QQQ took a trade etc. Because from a logistics standpoint I couldn't issue the trades for all of these like I do with the SPY. For me after hrs is no big deal - heck most of those trades I enter in after hrs anyway. But the issue here is that the trades would always be well after the market close maybe 15 and 30 min or longer because I wouldn't keep the workspaces open during the day for stability, and then would open them all up after hrs to see what triggered.
The point here is that all the signals would be well after the close. For you guys that's really no big deal (except for options of course) - but if I tried to market this to a wider audience, there's a lot of people that wouldn't like that every one of the trades would have to be executed in after hrs - especially if it's all the trades. I was thinking that the next day open might have more appeal to a wider audience. And, it's less work on me because I don't need to have an urgent need to try and confirm all those trades after the close - I can leisurely do it - look them over, maybe run a quick backtest to test something if I wanted to - then send the trade notification email out that evening.
I agree Matt, next day
Posted by littlegrasshopper on 19th of May 2022 at 06:31 pm
I agree Matt, next day at the open seems like it would work better for a wider audience taking 2nd and 3rd entries with limit orders.
Great work, thank you!!
BTW, will the next VVIX signal be considered just as valid with a close above 110?
I love the chance to
Posted by focus175 on 19th of May 2022 at 06:44 pm
I love the chance to be able to utilize the systems for playing IWM and the Qs in the future. I'm not able to trade the first several hours of the day, so if you're able to send out the notification for those in the evening the day before like you mentioned, that'd definitely be helpful.