Page 1 of 1

MC 8.7 question about T&S data

Posted: Jul 16 2013
by radekj
MC 8.7 is totaly inaccurate and it is missing data !

Look at volume too, values are wrong too !

snaps:

http://screencast.com/t/ncQmIyhH5y

settings:
http://screencast.com/t/aqdHxy3o

radekj

Re: MC 7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by Dru
MC 7 is totaly inaccurate and it is missing data !

Look at volume too, values are wrong too !

snaps:

http://screencast.com/t/ncQmIyhH5y

settings:
http://screencast.com/t/aqdHxy3o

radekj

PS:
don't tell me that MC 7 release can be used to manage real money, you dont take you customers seriously with such bugy release version !
1) InsideBid & InsideAsk is Status Line fields.
2) Status Line data is updated asynchronously regarding to realtime tick data updates.
Therefore your _TestLastBar don't show nothing.

Re: MC 7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by radekj
pls dont waste my time with unqualified comments and read product documentation !

Here for you (orginale MC documentation), if you can not find it:

InsideAsk
Returns a numerical value indicating the current best Ask for the symbol that the study is applied to.

Usage
InsideAsk

Notes
Quote Fields cannot be referenced historically.

Example
InsideAsk will return the current best Ask

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by radekj
I looked at MC 8.5 beta2,

i can not test the T&S values because it does not have it !
but i see that insideask and insidebid in mc 8.5beta2 is right !
So again it seems that 8.7 release have bug in data-processing !

radekj

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by Nick
Just to state the obvious if you only have one data series in your chart, only ticks in that series will cause your code to run. So for example you could have dozens of bid ask changes that you wont see in your code as there are no ticks generated (to trigger a run of your indicator) until you get an actual trade. Put another way a change to InsideBid and InsideAsk wont make ones code run. I cant see all your code from the screenshot but I suspect that may be your problem.

Until recently even adding bid and ask ticks to your chart as data2 and data3 wouldn't help as 'ticks' in data streams data2 and data3 would still not cause an indicator to run. I believe this was scheduled to be changed and have a suspicion it might have in the latest release (haven't tested so cant say for sure). Perhaps some one form TSSupport will confirm?

In any case I don't think InsideBid and InsideAsk do what you think they are functions that return a 'snapshot' rather than a way to access complete series of bid ask data.

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by radekj
I dont know what you are trying to say.

First: you see all my code ! That are two lines not more not less.

Second: datafeed deliver a row of

...,Trade-price, ask-price, bid-price,... !

So all i want is to see the correct values for trade,ask,bid !

That can no be so hard to do this functionality right, as i wrote in MC8.5beta it works !

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by Nick
Your code only will run when there is a TRADE. There could be millions of bid ask changes and they will not be detected by your code because MC will not run it for bid ask changes.

Incidentally I am not 'trying to say' anything. Read what I wrote again and try and understand what I actually said.

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by radekj
Exact !

i will see the right ask/bid for a trade.

(RIGHT ASK/BID = the ask/bid price which was at the time as trade was generated,
this information is coming in with trade-price in same data-row,
no need for second or third data on chart)

But MC does not show the right ask/bid for a trade that why T&S are wrong !

Now one question to you, you would let software trade you money if you know it can not handle price-data right ?

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by SP
Not sure if it helps:
- At the QM use "Generate a new Ask/Bid tick if Trade changes"
- Your code says "if Close >= insideask ... else if Close<=insidebid", what are you doing with all trades between especially in the JY ?

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by radekj
The example is for test purposes !

If T&S shows trades on bids but MC does not see it in code then something is wrong,
no matter what happens with trades between ask and bid !

I am tired of all this, i showed that MC has problem with data,
if someone care about it then it is ok if not then it is ok too.
Happy profits !

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by TJ
The example is for test purposes !

If T&S shows trades on bids but MC does not see it in code then something is wrong,
no matter what happens with trades between ask and bid !

I am tired of all this, i showed that MC has problem with data,
if someone care about it then it is ok if not then it is ok too.
Happy profits !
MultiCharts is an "Event Driven" program. (so as 99% of the charting programs on the market).

"Event Driven" means that the program sits at idle all the time, it will wake up to do its calculations ONLY when an event triggers it.

In MultiCharts' case, the "event" being the arrival of a new tick. (ie. a trade).
Or, in the later versions, also by the keyword RecalcLastBarAfter.

ie. if no new tick arrives (no trade), the values of MACD, Stochastics, Moving Average, etc., will stay the same.

ie. any changes in the bid and ask value will be reported, only if it coincides with a trade.

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by radekj
TJ read the previous posts pls.

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by Nick
<sigh> TJ 'gets it' you really need to think about what people are telling you.

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by TJ
TJ read the previous posts pls.
I know what you are asking... but it is not happening.


ps. InsideAsk and InsideBid are Quote Fields. ie. they are meant for displays only.

ps. Quote Fields usually come from a different data stream. Please confirm with your dataprovider.

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 16 2013
by TJ
Maybe you do encounter a bug; please connect to live chat with the tech support.

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 17 2013
by CrazyNasdaq
I've pick up the picture in the first post (http://screencast.com/t/ncQmIyhH5y) and elaborate it putting side by side the quotefileds of T&S and the print of your code.
This because suddenly one thing came to my eyes.........the time stamp of your print file through your code.
Image
Your file has 20 lines about trades, the T&S has 38 lines.
Your file has always a time stamp different from the T&S, specially regard milliseconds time and this is because snapshot doesn't mean accuracy.
You can have accuracy only through data that is stored and not data that is temporally dispalyed as a snapshot, and InsideBid and InsideAsk are not stored but only displayed/snapshots of the system at a particular point in time.
http://en.wikipedia.org/wiki/Snapshot_( ... r_storage)

This means that insideBid and InsideAsk probably are not the way to achieve what you are searching and testing.
In the past in this forum we have faced this problem regards Cumulative Delta before its implementation by default in Multicharts (http://www.multicharts.com/discussion/v ... lta#p37675) and you can find that InsideBid and InsideAsk are not an accuracy way to test "real Bid and Real Ask" of the market.
If you make a search in the forum regard "insideBid and InsideAsk", you will find what I'm referring to.
Probably you have to make your test by an other way.
Regards

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 17 2013
by radekj
Thanks for the info !

radekj

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 17 2013
by CrazyNasdaq
Thanks for the info !

radekj
You are welcome

Re: MC 7 is totaly inaccurate and it is missing data !

Posted: Jul 17 2013
by Dru
pls dont waste my time with unqualified comments and read product documentation !
Read is not enough. It also needs to be properly understood ;-)

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 24 2013
by furytrader
I hereby recommend this post for the award of "Most Annoying Message Title Of 2013"

Speech! Speech!

Re: MC 8.7 is totaly inaccurate and it is missing data !

Posted: Jul 24 2013
by TJ
I would suggest anybody who needs help here heeds Dru's recommendations. It will save you some time.



Just my humble observation. YMMV

Re: MC 8.7 question about T&S data

Posted: Aug 21 2013
by PD Quig
I'm trading the spread between two instruments, shorting one and going long the other simultaneously (or vice versa), when the spread between the two hits a certain value. My strategy currently uses a spread calculation that is based on the last trade price of each. Normal entry and exit slippage issues are compounded because I'm taking hits on two instruments, and I cannot use limit orders because one side of the pair may not get filled.

I came across this thread while researching an approach using the InsideBid and InsideAsk functions to retrieve bid and ask prices instead of last price. Ideally, I would like entries and exits to be triggered for a spread between the two instruments that is calculated on selling at the bid of one and buying at the ask of the other. That way my calculation is essentially a worst case and orders are more likely to be filled closer to my targets.

After reading this thread, I wanted to be clear whether or not MC would be able to trigger order actions based on the Instrument_1_Bid – Instrument_2_Ask spread reaching a certain level—it sounds like a tick / trade event is required before the strategy can act. If that is the case getting bid and ask data may not do me any good anyway.

Could you please confirm? Any workarounds?

Re: MC 8.7 question about T&S data

Posted: Aug 22 2013
by piranhaxp
I'm not sure I understood the hole inside_bid/-ask theory, but for me seems to work when u use bid & ask set in your symbol settings for symbol 1 as data1 and symbol 2 as data2 with . Here one example for symbol setting on ask prices :

Image

As you can see in the following chart, bid & ask are updated (if set as data) when they change.

Image
(red : ASK / black : TRADE / blue : BID ... for the same symbol so will work on diff. sym's as well .. I guess :) )

So you are able to calculate the "real" spread with ask on data 1 and bid on data2 when you are data1 SHORT and data2 LONG. Vice versa if position for each symbol is checked. Further you could record the size of ask & bid and maybe adapt your execution price as limit for the next bid/ask level till your order would be filled on the first bid/ask price.

BTW ... what execution platform are you using for your spread trading ?

EDIT : ONE EXAMPLE HOW IT COULD LOOK LIKE .....

Image

Re: MC 8.7 question about T&S data

Posted: Aug 22 2013
by PD Quig
Piranha,

Hey, thank you very much. It had not occurred to me to chart both data1 and data2 as three separate streams (bid, ask, trade). Lots to think about and test there. The open question is whether or not an MC strategy can execute unless a new tick event occurs? With your approach the bid+ask+trade data for both sides of the pair are available, but can an new order be triggered without a new trade occurring? I guess I'll have to test that on live data.

Thanks again for the ideas.

PDQ

Edit: BTW, execution is all through the MC / IB interface. I used to use a 3rd party API but their extra layer of complexity got to be too painful.

Re: MC 8.7 question about T&S data

Posted: Aug 22 2013
by TJ
::
but can an new order be triggered without a new trade occurring? I guess I'll have to test that on live data.
Look up RecalcLastBarAfter.

Re: MC 8.7 question about T&S data

Posted: Aug 22 2013
by PD Quig
Thanks, TJ. I'm aware of that function from a prior thread where you pointed me in that direction. What I'm having trouble wrapping my head around is whether using the RecalcLastBarAfter timer to poll the data1_bid / data2_ask data stream every second is responsive enough. The spread criterion to execute orders has to be present at the moment the RecalcLastBarAfter timer activates. If the criterion is met momentarily between timer ticks--but then disappears again--the orders will not be launched, right?

Re: MC 8.7 question about T&S data

Posted: Aug 22 2013
by TJ
Thanks, TJ. I'm aware of that function from a prior thread where you pointed me in that direction. What I'm having trouble wrapping my head around is whether using the RecalcLastBarAfter timer to poll the data1_bid / data2_ask data stream every second is responsive enough. The spread criterion to execute orders has to be present at the moment the RecalcLastBarAfter timer activates. If the criterion is met momentarily between timer ticks--but then disappears again--the orders will not be launched, right?
That is correct. The computer is dumb, it does not have emotions; it only process explicit instructions. During the "Recal", your trade criterion must be met for an order to trigger. If the criterion is met, an order will be sent at the next "event". ie. next tick, or at the next Recal, whichever comes first.