4:20 pm:--X-- Welcome - Linn Software: Powerful Market Analysis Software for the RealTime Investor for Windows XP and Macintosh OSX. See http://www.Linnsoft.com for complete information, tour, demo, videos, etc... Formal Chat session's are Tuesday's at 4:30 est.
4:30 pm:<LS_Bill> Hello everyone
4:30 pm:<LS_Bill> Chad is on vacation this week
4:30 pm:<Julian> Good Evening, Nice holiday?
4:31 pm:<LS_Bill> Yes, I guess it was a working day for you.
4:31 pm:<Julian> So is Eddy a Chad "impostor"
4:31 pm:<LS_Bill> could be
4:31 pm:<Julian> Yeah, kind of, no markets but allowed me to catch up on other stuff
4:33 pm:<LS_Bill> I have seen very few holiday related suport emails today, a good sign I hope
4:33 pm:<Eddy> hello guys
4:33 pm:<LS_Bill> normally there are more than a few holiday related charting questions
4:33 pm:<Julian> yeah, holidays alway mess with charts
4:34 pm:<Julian> Q "Wheres my data" A: "Back a bit"
4:34 pm:<LS_Bill> Hi Eddy
4:34 pm:<Eddy> btw, in the 841 TPO version which were available on the site , the default chat room nickname is Chad :-)
4:35 pm:<LS_Bill> ah, that's been fixed now Eddy.
4:35 pm:<LS_Bill> those old versions were supposed to expire due our license agreement with CBOT
4:36 pm:<Eddy> Bill : i did further test for Ninja : this is becoming desesparate.. i am using a windows xp laptop pruchased in the US and downloading all the latest stuff and couldn't you make that Ninja dll request optional
4:36 pm:<Eddy> Bill : not yet..
4:36 pm:<LS_Bill> I think we are going to be required to obsolete those releases, it a contractual obligation to CBOT to pay the CBOT fees
4:36 pm:<Eddy> that the one i tested the IB feed - didn't help to solve the issue
4:37 pm:<Eddy> Bill : this was the only IB version with predefined futures settings - such settings should be standard i think...
4:38 pm:<LS_Bill> yes, Chad built custom installers for IB, eSignal, IQFEED
4:39 pm:<LS_Bill> with standard futures pre-defined
4:39 pm:<Eddy> that was helful for such testing purposes when you are not familar with instrument codes of other datafeed
4:39 pm:<LS_Bill> we're going to work on a way to have one installer but import the feed specific symbols
4:39 pm:<LS_Bill> now that we have a quotepage export definition, this should be easy
4:40 pm:<Eddy> good : that will be a time saver
4:40 pm:<LS_Bill> just detect if the install is a new one and do an import of the feed specific symbols
4:40 pm:<LS_Bill> we just have to roll over the symbols periodicially.
4:42 pm:<Eddy> Bill : back to the NT stuff, any way to make that 2nd dll request optional.. at least i will be able to further test the NT IRT link
4:42 pm:<LS_Bill> 2nd DLL ?
4:43 pm:<Eddy> yes, you sent to NT 3 request thru the dll
4:43 pm:<Eddy> the first one is working one
4:43 pm:<Eddy> getting order status
4:43 pm:<Eddy> but the second one is causing the crash
4:44 pm:<Eddy> getting market position
4:45 pm:<Eddy> this arerequest are sent automatically to Ninja after sending an order
4:45 pm:<Eddy> introduced first in 8.4.0
4:45 pm:<LS_Bill> yes
4:45 pm:<LS_Bill> I'm looking at the code right now
4:46 pm:<LS_Bill> what was the nature of the trading order that led to this?
4:46 pm:<Eddy> in the log, one get first : "Checking NinjaTrader Status"
4:46 pm:<Eddy> all trading order
4:46 pm:<Eddy> even simple BUY MKT
4:47 pm:<Eddy> so i do get after first request BUY1ESU60238 then Order BUY1ESU60238 - Filled
4:47 pm:<Eddy> so far so good
4:47 pm:<Eddy> but then
4:47 pm:<Eddy> when you sent the second request
4:47 pm:<Eddy> getting market position
4:48 pm:<Eddy> i do get 8.4 Rev 1 fatal error of type C0000005 at location 00000000
4:48 pm:<Eddy> always
4:48 pm:<Eddy> that particular second request is causing some issues - don't know why
4:49 pm:<LS_Bill> is the trading order setup with "Default" as the symbol?
4:49 pm:<Eddy> yes
4:49 pm:<Eddy> i hva tried order with and without startegy name referecne
4:50 pm:<Eddy> the order is well sent to NT - no problem on that part - even NT logs shosws order is well processed
4:51 pm:<Eddy> on IRT side, no problem as well when order is sent
4:51 pm:<Eddy> NinjaTrader command: PLACE;Sim101;ESU6;BUY;1;MARKET;0;0DAY;;BUY1ESU60238;;;
4:51 pm:<Eddy> as per IRT logs
4:51 pm:<Eddy> crash always caused by the second automatic request to Ninja, the get market position....
4:52 pm:<LS_Bill> that that show up in Ninja log?
4:52 pm:<LS_Bill> does that...
4:52 pm:<Eddy> NinjaTrader command: PLACE;Sim101;ESU6;BUY;1;MARKET;0;0DAY;;BUY1ESU60238;;; is the IRT log
4:53 pm:<Eddy> Ninja log will say that an ATI order was received and was processed
4:54 pm:<Eddy> don't have the exact log just hre
4:54 pm:<Eddy> but i know this looks ok
4:55 pm:<Eddy> as the new position (long 1 in this BUY xase) is showing up clearly on the DOM
4:55 pm:<Eddy> is there anything special witht hat 2nd request ?
4:57 pm:<LS_Bill> the second request is asking NT for info on the symbol associated with the trading order
4:57 pm:<LS_Bill> to obtain avg fill price, position size, etc.
4:58 pm:<Eddy> so this looks like to be some issue around the symbol code or so
4:59 pm:<LS_Bill> is there anything ususual about the trading order? are you using T# variables in the trading order?
4:59 pm:<LS_Bill> what is the account strategy?
4:59 pm:<Eddy> nope, just basic BUY market as per above order
4:59 pm:<Julian> Aside from the problems it sounds like you have at this moment, would you recomend ninja trader for automating (my as yet not working) TSYS strategy?
4:59 pm:<Eddy> no startegy reference
5 pm:<Eddy> just BUY MKT 1 / no stop order or limit order associated
5 pm:<LS_Bill> I have not been able to reproduce this failure here. Daniel (xortrader) has been testing IRT / Ninjatrader as well and was not having these problems
5:01 pm:<LS_Bill> Daniel has been absent for the last few chats, I don't know what he is up to recently.
5:01 pm:<Eddy> thats why i tried with the IB version, to have the same setup
5:01 pm:<Eddy> than him
5:01 pm:<LS_Bill> yes
5:01 pm:<Eddy> as this could have been esignal related
5:01 pm:<LS_Bill> did you start with a fresh system Eddy
5:01 pm:<Eddy> but still getting the same issue
5:01 pm:<LS_Bill> fresh db or did you reuse an existing database?
5:02 pm:<Eddy> fresh db - reformmated laptop - us settings - xp sp2
5:02 pm:<Eddy> tried existing db as well
5:02 pm:<LS_Bill> and the trading orders you setup were created new
5:03 pm:<Eddy> yes --- i think i never tried as many options as in that case :-) and you know i have some experience with IRT
5:03 pm:<Eddy> troubleshooting
5:04 pm:<Eddy> so brend new db - only one new trading order on a chart and here is the bug..
5:04 pm:<LS_Bill> Eddy, I will put some msg log diagnostics into a test release for you to run with, these may shed more light on the matter.
5:05 pm:<Eddy> thanks Bill, because in that case i am really stuck - and tehre is so much i would like to test with the NT IRT link
5:05 pm:<Julian> Are there any diagnostics that can be enabled in the normal versions, to try and pin down the problems I have?
5:06 pm:<Eddy> Julian, i think you need to use some ES #F(V#1) type of custom instrument to solve your TSYS issues... if the same as the one you explained last time
5:07 pm:<Julian> In the email I sent last week?
5:07 pm:<Eddy> and store historically V# value (managed by your TSYS) in such instrument
5:07 pm:<Eddy> in the chat of few weeks ago
5:08 pm:<Eddy> but that's really no easy stuff to implement, even if this is working nicely when all set up
5:08 pm:<Julian> I got past that bit Eddy, its now down to TSYS trades disappearing and re-appearing, I sent an email to Bill about it last week with lots of attachments and examples
5:09 pm:<Eddy> TSYS live trade or offline ?
5:09 pm:<Julian> Live
5:09 pm:<Eddy> not the same beast...
5:09 pm:<Julian> Offline the charts are fixed
5:09 pm:<Julian> during market hours the trade actions work about 20% of the time
5:10 pm:<Julian> if the chart is refreshed it changes (even if the data has not) and shows trades it entered say 10 bars ago as current
5:10 pm:<Eddy> then probably some recalc settings, or chart time range settings
5:11 pm:<Julian> even if you refresh twice the TSYS usually appears and disappears with each data refresh
5:11 pm:<Julian> Got the chart time frames covered too - very careful
5:11 pm:<Julian> about that
5:11 pm:<Eddy> never use for example last 50 bars, always use charts with a fixed start up time
5:11 pm:<Eddy> ie starting 08.30 2 days ago
5:12 pm:<Julian> The first tsys rule is IF POS=1 or POS=390 then SET V#x,0 etc etc to clear all the V# vars down
5:12 pm:<Eddy> whoooo
5:12 pm:<Eddy> dangerous life you have
5:12 pm:<Julian> Back tests work absolutely perfectly
5:13 pm:<Eddy> as is said, this is another beast in live conditions
5:13 pm:<Julian> And the last TSYS is IF POS_STATE=POS_LONG THEN CLOSE etc
5:13 pm:<Julian> So nothing is carried overnight
5:14 pm:<Eddy> V# are good for backtest
5:14 pm:<Eddy> as only one calc per bar
5:14 pm:<Julian> even just one stock open (i.e 1 stock marked intraday) one chart - still does not catch every trade, refreshes mess it up
5:14 pm:<Eddy> as soon as you refresh, you recalc all bars
5:15 pm:<Eddy> tahts what i say, in order to do live trading, you can't use the same rules than in backtest
5:15 pm:<Julian> I realise it recalcs, thats the only thing that will get the trades to appear on a chart
5:15 pm:<Eddy> tahts because of recalc there are not there at 1st place
5:15 pm:<Julian> But is a TSYS not for live trading then?
5:15 pm:<Eddy> yes of course
5:16 pm:<Eddy> but sompe RTL code have nt the same maening in backtest or live
5:16 pm:<Eddy> like TIME
5:16 pm:<Julian> Time is not used
5:16 pm:<Eddy> or like resetting v#
5:16 pm:<Julian> The bar number of a trade thats all
5:17 pm:<Julian> I cannot see a way of implementing the TSYS I need without using V# variables
5:17 pm:<Eddy> you need to collect live your V#
5:17 pm:<Eddy> managed by your TSYS
5:18 pm:<Julian> Sorry, what do you mean?
5:18 pm:<Eddy> and make sure there arent modified by a recalc / refresh
5:18 pm:<Eddy> you don't want to modify one of your V# after it was calculated at the end of a given bar
5:19 pm:<Julian> why would the answer not be the same, as it is in a back test?
5:19 pm:<Eddy> a recalc will do that
5:19 pm:<Eddy> beacuse in a backtest you never modify your data after the bar
5:19 pm:<Eddy> it is calculated bar per bar
5:19 pm:*** IRT_USER has joined us
5:19 pm:<Julian> To take a trade there are 3 steps in 3 time frames to go through, before a trade is taken
5:20 pm:<Eddy> but in a refresh, at 10.30, you modfiy your V" calcualted at 08.30
5:20 pm:<Julian> but the first thing it does is reset, then calc each closed bar up to the current time?
5:20 pm:<Eddy> taht will again modify your V# at 10.31
5:21 pm:<Julian> But if do not touch a chart, it does not catch the trades either, thats the only reason I refresh, because it generally shows nothing unless I do?
5:22 pm:<Eddy> then you have issue with your recalc settings
5:22 pm:<Julian> Recalc settings on the chart?
5:22 pm:<Eddy> i cant tell more much more w/o seeing your particular code
5:22 pm:<Julian> I sent the email to Bill, should I forward it to you too?
5:22 pm:<LS_Bill> when you refresh the chart (download fresh data or just adjust prefs) it's like running the trading system from the oldest data to present
5:22 pm:<Eddy> yes on the charts
5:23 pm:<Eddy> you know, i am just another user...
5:23 pm:<Eddy> even if not everyone manage 500+ V# in its trading systeme :-)
5:23 pm:<Julian> Sorry, I thought you were part of Linnsoft
5:23 pm:<Eddy> no problem
5:23 pm:<LS_Bill> Eddy probably done more with backtesting and TSYS than anyone
5:24 pm:*** IRT_USER is now known as efranco
5:24 pm:<Eddy> that was a nice creation of yours Bill
5:25 pm:<Eddy> still remember the TSYS precursor, the SSIM ?
5:25 pm:<LS_Bill> oh yes
5:25 pm:<Eddy> that was a nice token...
5:25 pm:<Eddy> except the setup windows was getting crowded
5:25 pm:<Julian> If it really was recalc problem, why would a recalc outside market hours work?
5:26 pm:<Eddy> But Julian, you can always sent me your email, i ll have a look
5:26 pm:<Eddy> a recalc outside mkt hours ?
5:27 pm:<Julian> YES - THIS IS LOCKED IN CAPS NOW, EVEN WITH CAPS LOCK OFF!
5:27 pm:<Eddy> Julian, you say you have 3 different timeframe : you are using MPD ?
5:27 pm:<efranco> I have a question about the POS_SIZE and ENTRY tokens, are they linked to a symbol? As in, could I use POS_SIZE for symbol XYZ in the order logic for symbol ABC? (e.g. if POS_SIZE for XYZ > 100 then sell 100 ABC)
5:27 pm:<Julian> Yes MPD
5:28 pm:<Eddy> you are sure you are not stuck in the looking forward trap ?
5:28 pm:<Eddy> that might happens with MPD
5:28 pm:<Julian> looking forward trap?
5:28 pm:<LS_Bill> You can qualify a token efranco, e.g. POS_SIZE(IBM) would give you the postion size at present for IBM
5:29 pm:<Julian> I.e knowing what the 3 min bar does before it gets there?
5:29 pm:<Eddy> yep, trigerring signal using data that have not yet occured
5:29 pm:<efranco> super
5:29 pm:<Eddy> ie looking ahead of the horizon
5:29 pm:<Julian> Would a backtest let you do that?
5:29 pm:<Eddy> time horizon
5:29 pm:<LS_Bill> same with V# variables, V#1 is user variable #1 for the instrument in question while V#!(IBM) refers to the value of user variable #1 for IBM
5:29 pm:<Eddy> a godd programmer : no
5:30 pm:<Eddy> every program let you do that
5:30 pm:<Eddy> the point is
5:30 pm:<Eddy> this can easily happens when miwing timeframe
5:30 pm:<Eddy> signals
5:31 pm:<efranco> ok, thanks bill
5:31 pm:<efranco> btw are you bill linn?
5:31 pm:<Eddy> the only Bill Linn :-)
5:31 pm:<Julian> I get what you mean, but even signals 2 hours ago change when recalcing, not the current ones
5:32 pm:<LS_Bill> yes, linnsoft bill
5:32 pm:<efranco> wow i guess you guys really do support your product well
5:32 pm:<efranco> i never get to chat with the tradestation guys :-(
5:33 pm:<Eddy> lol
5:33 pm:<LS_Bill> as best we can, julian and eddy are on the leading edge, if you've been following this dialog ;-)
5:33 pm:<Julian> Its great fun
5:34 pm:<Eddy> Julian : how lo,g are you using TSYS if you don't mind me asking that ?
5:34 pm:<LS_Bill> I have an appointment to go to, have to leave early today...
5:34 pm:<LS_Bill> any last questions?
5:34 pm:<Julian> You mean, in programming terms?
5:34 pm:<Eddy> yep
5:34 pm:<efranco> oh question
5:34 pm:<efranco> the problem i emailed you about
5:35 pm:<efranco> can i use TIME
5:35 pm:<Julian> Only about 1 month, but rtl a few months before that. But I am a contract programmer by profession
5:35 pm:<efranco> to keep the strategy open only from 5AM to 5 PM
5:35 pm:<efranco> for the london exchange
5:35 pm:<Eddy> ok, thks, in that chat room, you have user with anything between 5 hours and 5 yrs + of IRT experience
5:35 pm:<LS_Bill> I have that email efranco, I will have to respond to you later, it's rather detailed as you know
5:35 pm:<efranco> and then from 5 PM to 9 PM use TIME and the POS_SIZE (LONDONSYMBOL) >= some number
5:35 pm:<efranco> oh ok
5:35 pm:<efranco> thank you!
5:36 pm:<LS_Bill> have a nice evening everyone.
5:36 pm:<Julian> Bye Bill
5:36 pm:<Eddy> bye bye Bill
5:36 pm:<efranco> take care Bill
5:37 pm:<Eddy> Julian : switching from TSYS backtest to live mode is the most complex stuff in IRT - only thing i can tell is that i always was able to have same result live than backtest
5:37 pm:<Eddy> as long as the backtest result were not cheated
5:37 pm:*** LS_Bill has left
5:37 pm:<Julian> The live ones that work, match the backtest, it just does not catch them all real time
5:37 pm:<efranco> hey guys, maybe you have some ideas for me? when you're done discussing tsys just yell
5:38 pm:<efranco> so i can recap the item
5:39 pm:<Julian> Well the backtests do match live trades but just not often enough
5:39 pm:<Eddy> i see
5:39 pm:<Eddy> i never had time to write tutorial about TSYS
5:40 pm:<Eddy> but i wrote one about backtest few years ago
5:40 pm:<Eddy> http://www.linnsoft.com/tutorials/F_trader_backtest.htm
5:40 pm:<Eddy> there was no TSYS yet at that time
5:40 pm:<Eddy> niether POS token
5:41 pm:<Eddy> or MPD :-)
5:41 pm:<Julian> Oddly the thing is way more reliable if.....
5:41 pm:<Julian> I place a tsys on a chart,ignore it, set some v# variables, use the POS_STATE on a quote page
5:42 pm:<Julian> and set trade actions from the pos_state token being set
5:42 pm:<Julian> the charts do not show the trade boxes but the quote page tends to update more reliably, not perfectly, but better
5:42 pm:<Eddy> you will still have issue
5:42 pm:<Eddy> let me take an example
5:43 pm:<Eddy> remnber the re is only one V# value at agiven time
5:44 pm:<Eddy> V# are not historical value
5:44 pm:<Julian> Yep - no historical
5:44 pm:<Eddy> so if you have some rules linked to a V#, becareful with the trigeering
5:44 pm:<Eddy> of trades
5:45 pm:<Eddy> but i just need to see how you use these V# in your trading rules
5:45 pm:*** Mpt has joined us
5:45 pm:<Eddy> as the issue is just overthere
5:45 pm:<Eddy> most probably
5:46 pm:<Eddy> you know INTC(V#1) type of instrument ?
5:46 pm:<Eddy> ie how set them up ?
5:47 pm:<Julian> No - I never encountered that use of retrieving a V#
5:47 pm:<Eddy> of storing an historical V# array of data
5:51 pm:<Julian> So using INTC(V#1) creates an array?
5:51 pm:<Eddy> let make it easy
5:51 pm:<Eddy> you have a 1 minute chart
5:51 pm:<Eddy> with one custom indciator which is SET(LAST,V#1)
5:52 pm:<Eddy> recalc is once per bar
5:52 pm:<Julian> yep
5:52 pm:<Eddy> then INTC(V#1) will be populated with one data every minute which is teh CL of the bar
5:52 pm:<Eddy> if recalc is 1 sec
5:52 pm:<Eddy> then INTC(V#1) instrument will be populaed with 60 data per minute
5:53 pm:<Eddy> one evry sec, ie the last tick of the given hh:mm:ss
5:53 pm:<Julian> INTC being the symbol not an RTL TOken?
5:53 pm:<Eddy> INTC(V#1) is than an instrument like any other
5:53 pm:<Eddy> yep
5:53 pm:<Eddy> INTEL
5:54 pm:<Eddy> you just need to define a custom instrument, name it like that, and say it is a cash instrument
5:54 pm:<Julian> Okay - historic V#1 - or at least a historical trail
5:54 pm:<Eddy> some do probably in the linnsoft manuals
5:54 pm:<Eddy> search for INTC(V"1) as i think this is the example taken by Bill in its pdf
5:55 pm:<Eddy> yep, this is collecting the historical V#
5:55 pm:<Julian> So it basically creates a custom instrument to "hold" the data
5:55 pm:<Eddy> yep
5:55 pm:<Eddy> only works live of course, not in backtest
5:56 pm:<Eddy> as you can only populate such instrument live
5:56 pm:<Julian> quite it could happen every 3 seconds
5:56 pm:<Eddy> yep
5:57 pm:<Eddy> but in 90% of the case, i am sure you could avoid V# in your trading rules
5:57 pm:<Julian> if its done in a TSYS calc'd at end of bar - will it be the value at end of bar or end of bar plus a second or two, depending on when IRT gets round to calculating it
5:58 pm:<Eddy> the timestamp will be the close of the bar if i remember
5:58 pm:<Eddy> of course make sure your CPU is not at 100%
5:58 pm:<Eddy> i run all IRT on my RAM
5:58 pm:<Julian> can each array element be accessed individually? like INTC(V#1).1 etc?
5:58 pm:<Eddy> much quicker as the limit is the access to your HD
5:59 pm:<Julian> Do you use MAC or Windows?
5:59 pm:<Eddy> INTC(V#1).1 will be CL.1 for the INTC(V#1) instrument...
5:59 pm:<Eddy> Windows
5:59 pm:<Eddy> got esignal and IRT and NT on RAM...
5:59 pm:<Eddy> much quicker
6 pm:<Julian> The whole of IRT , or just eh Data_f folder?
6 pm:<Eddy> whole
6 pm:<Julian> Do you use a ramdisk type utility
6 pm:<Eddy> yep
6:01 pm:<Eddy> the one of cenatek, but many works fine
6:01 pm:<Julian> I suppose for RT day trading speed is the key for slippage
6:01 pm:<Eddy> i dont have slippage issue because of the design of my trading system :-)
6:02 pm:<Eddy> i only using STOP LIMIT order at predefined levels
6:02 pm:<Eddy> this is explained in the FAQ i posted...
6:02 pm:<Julian> I will be sure to read, and re-read that
6:02 pm:<Eddy> eitehr i am filled, either i am not... but no slippage
6:03 pm:<Eddy> a sytem buying at the CLOSE of a given bar is just worth nothing..
6:03 pm:<Eddy> you must know your entry price in adavance..
6:03 pm:<Eddy> same for exit..
6:03 pm:<Julian> And collect the ECN rebate too!
6:04 pm:<Eddy> i am just trading futures, no ECN issue
6:05 pm:<Julian> futures seem (from the little I know) less muddy to trade
6:05 pm:<Eddy> i started with nasdaq "hot" stocks 5 years ago
6:05 pm:<Eddy> too much work
6:06 pm:<Eddy> an idiot steal a Quarterly report and publish it on the net and the stocks halt
6:06 pm:<Eddy> so much hassle for such siutations..
6:06 pm:<Eddy> CME is sometimes down, but only once a year or so
6:18 pm:<Mike> hi all
6:19 pm:*** gloveday has left