Strange spikes in the e-mini from IB's TWS

Questions about MultiCharts and user contributed studies.
bowlesj3
Posts: 2180
Joined: 21 Jul 2007
Has thanked: 227 times
Been thanked: 429 times

Strange spikes in the e-mini from IB's TWS

Postby bowlesj3 » 05 Aug 2011

Hi,

I got an email from MultiCharts about IB TWS spikes (email shown below). Up until the email I figured all was normal in the "e-mini ES". Yes there were lot of long 10 second bars due to the big drop in the market (this always happens and sometimes these bars can be $400 per contract in length - my program paints then red if they are $200 or more per contract). However a very strange one which I attached came through after the email. It effected 10 second bars and all bar sizes right up to the 60 minute bars. I restarted MC and it corrected itself except for the 10 second bars. A few questions.

1/ Am I to assume that this is the problem mentioned in the email?
2/ When did this issue start. I once saw a very strange long bar one time before but only that one time.
3/ I am assuming that if this is the issue a long trade would not have been stopped out by this.
4/ Did anyone get the exact same thing or is this a random per user thing?

Thanks,
John
Dear users,

If you are using Interactive Brokers as a data source you may sometimes get bad ticks, so-called "spikes" with the price 10-15 points higher or lower than current price. This is a known issue on IB API side and has been confirmed by IB as a bug.

In order to solve the issue please enable "Generate new tick if total volume changes" option in Quote Manager --> Tools --> Data Sources --> Interactive Brokers --> Settings. This will eliminate spikes however some ticks may slightly differ from IB TWS data as they will be formed on volume change basis.

Happy Trading
MultiCharts Team
http://www.multicharts.com
Attachments
20110805_144508_ES_--_LongBarMarketsSpikes.PNG
(26.4 KiB) Downloaded 660 times

User avatar
aiti
Posts: 43
Joined: 25 May 2009
Location: Germany
Has thanked: 43 times
Been thanked: 20 times

Re: Strange spikes in the e-mini from IB's TWS

Postby aiti » 06 Aug 2011

Hello bowlesj3,

I got the same spikes in the ES-chart yesterday.
When it happened I go to tab "View" -> "reload data" -> "1 day back" and the chart was okay.

Last year I got these strange Bars in some currencies and it was alot of work to eliminate these wrong data with the editor of quote manager.But with the new version of multicharts and the reload-data-tab it is more easy.

To your point 3:
If you have a strategy active on the chart and your entry- or stop-order is triggered by this long bar the strategy will send the order.
But when this bar happened I have not checked whether the wrong data is trade,ask or bid or all together?

aiti :smile:

P.S. I didn´t get any email from TSsupport about this issue !!

bowlesj3
Posts: 2180
Joined: 21 Jul 2007
Has thanked: 227 times
Been thanked: 429 times

Re: Strange spikes in the e-mini from IB's TWS

Postby bowlesj3 » 06 Aug 2011

Hi aiti,

Thanks for the info. An interesting thing happened after I created the original post. When I restarted MC shortly after the spikes (when I had the png picture done) the spikes in the 1 minute bars through 60 minute bars went away. However after the trading day when I brought MC down and brought it back up off-line the spikes came back. I never tried a reload. I guess I should try a reload 2 days back (or whatever).

Regarding execution of a stop order, I do not execute through MC. Being what I call a system discretionary trader (a discretionary trader with very precise rules that reads the wave input manually rather than trying the difficult task of doing it 100% via EL code) I use the IB TWS app for orders (and I use a target/stop bracket). I was not in a trade but I am curious as to whether being a trade it would trip the TWS bracket stop order. I guess in MC it would trip a stop and send a market order, but the big question (probably for IB rather than TSS) is whether that market order would be executed at the spike values or the normal market values. If it is just an API bug then I am assuming it would be at the normal (none spike) prices.

John.

User avatar
JoshM
Posts: 2195
Joined: 20 May 2011
Location: The Netherlands
Has thanked: 1544 times
Been thanked: 1565 times
Contact:

Re: Strange spikes in the e-mini from IB's TWS

Postby JoshM » 06 Aug 2011

Thanks for this thread John, you're touching on quite some important points here.
Regarding execution of a stop order, I do not execute through MC. Being what I call a system discretionary trader (a discretionary trader with very precise rules that reads the wave input manually rather than trying the difficult task of doing it 100% via EL code) I use the IB TWS app for orders (and I use a target/stop bracket). I was not in a trade but I am curious as to whether being a trade it would trip the TWS bracket stop order. I guess in MC it would trip a stop and send a market order, but the big question (probably for IB rather than TSS) is whether that market order would be executed at the spike values or the normal market values. If it is just an API bug then I am assuming it would be at the normal (none spike) prices.
Good question. If the bug lays in the IB API, I'd assume this would not trigger a stop order, since I assume that the IB Order Server checks with the IB Historical Data Server if the price level is violated (and thus the order should be submitted). The API bug would only show in the connection between the Historical Data server and the client (MultiCharts in this case). In other words, I think this means that this API bugs causes MultiCharts to "see" different prices (with bad ticks) which are not there in the original data.

But, truth be told, there are an awful lot of assumptions in that line of thought ;) and I'm just a beginner, so it would be great if someone else can give their take on this.


On a related note: don't we need a 'bad tick filter' in MultiCharts to prevent these types of situations? I think that would be a great add-on.


Like Aiti I also didn't receive that email - how did you get subscribed to that email list John? Perhaps MC Support can comment on this, because in the future I'd also like to receive these mails. I think it would be a great added form of 'customer support'. :)



Regards,
Josh

User avatar
aiti
Posts: 43
Joined: 25 May 2009
Location: Germany
Has thanked: 43 times
Been thanked: 20 times

Re: Strange spikes in the e-mini from IB's TWS

Postby aiti » 06 Aug 2011

Hi aiti,

..... I was not in a trade but I am curious as to whether being a trade it would trip the TWS bracket stop order. I guess in MC it would trip a stop and send a market order, but the big question (probably for IB rather than TSS) is whether that market order would be executed at the spike values or the normal market values. If it is just an API bug then I am assuming it would be at the normal (none spike) prices.

John.
Hello bowlesj3,

this spikes were only in the MC chart.I have opened the IB charttrader at the same time and there was no spike.

The marketorder would be executed at the IB price.If you have a MC-Strategy including Stoploss with marketorder this Stoploss get triggered by the (spiked-)price.Then MC send the marketorder to IB and get filled to the normal price.In this way there is no problem.
On the other side if you have a takeprofit-marketorder then MC "think" the profit is reached and send the marketorder for takeprofit with "perhaps 400 US$" loss! :wink:

aiti :smile:

bowlesj3
Posts: 2180
Joined: 21 Jul 2007
Has thanked: 227 times
Been thanked: 429 times

Re: Strange spikes in the e-mini from IB's TWS

Postby bowlesj3 » 06 Aug 2011

Like Aiti I also didn't receive that email - how did you get subscribed to that email list John? Perhaps MC Support can comment on this, because in the future I'd also like to receive these mails. I think it would be a great added form of 'customer support'.
Good question Josh. I have no idea.

Thanks aiti, you have confirmed what I suspected (suspected because of the wording of the email). I guess I could have checked the TWS chart but the first thing I thought of was to bring it to the forum since anyone placing orders from MC has a bigger problem and the more attention it gets the better. Even for people like myself it can scare you a bit as well as waste time in distraction, double checking the TWS chart and doing reloads.

So in view of the above paragraph, here are the most important questions. Since it is a really big issue for MC users, do the MC people have any estimates as to when will it be fixed . The first spike like this I saw was about 3 or 4 weeks ago. That one and this one are the only ones I have seen. For such a serious problem 4 weeks to get it fixed is pretty long. On a related note, will a TWS upgrade fix it? I may be a few versions back on TWS (build 916.5). An email of its fix and the required TWS build would no doubt be one email worth sending to all MC users.

John

bowlesj3
Posts: 2180
Joined: 21 Jul 2007
Has thanked: 227 times
Been thanked: 429 times

Re: Strange spikes in the e-mini from IB's TWS

Postby bowlesj3 » 09 Aug 2011

As an experiment, on Tue Morning, I tried a reload to get those Friday afternoon spikes out shown in the picture in the first post. I chose 3 days back (Mon, Sun, Fri). It did not work. Am I correct in assuming I need to use 4 days back (Mon, Sun, Sat, Fri) or would it be 5 days which would include the current day.

Thanks,

By the way, I saw another spike in the 10 second bars from late last night. I am surprised there are not a lot more people posting here. Maybe no one auto trades MC because it would be real hard to trust it with this problem going on (even if you are baby sitting it). Sure glad I don't. For me it is not a huge deal.

User avatar
Dave Masalov
Posts: 1712
Joined: 16 Apr 2010
Has thanked: 51 times
Been thanked: 489 times

Re: Strange spikes in the e-mini from IB's TWS

Postby Dave Masalov » 09 Aug 2011

Dear Sirs,

Let me clarify the situation. These spikes come in real-time and can be eliminated by using "Generate new tick if total volume changes" option as was described in the e-mail that we have sent to all registered users (if you have not received it please check your spam box or send us the updated contact information). This is IB API issue confirmed as a bug by IB API support. You may want to contact IB to speed up fixing this bug.

MC 6 did not have this spikes problem because initially IB did not send ticks through their API - so MC plotted ticks based on total volume change. As IB started to provide normal tick data, MC 7 gives you an option to plot ticks that IB provides or the ones based on total volume change.
I saw another spike in the 10 second bars from late last night. I am surprised there are not a lot more people posting here. Maybe no one auto trades MC because it would be real hard to trust it with this problem going on (even if you are baby sitting it). Sure glad I don't. For me it is not a huge deal.
John,

Have you tried to enable "Generate new tick if total volume changes" option? It should eliminate those spikes.
If the bug lays in the IB API, I'd assume this would not trigger a stop order, since I assume that the IB Order Server checks with the IB Historical Data Server if the price level is violated (and thus the order should be submitted).
The marketorder would be executed at the IB price.If you have a MC-Strategy including Stoploss with marketorder this Stoploss get triggered by the (spiked-)price.Then MC send the marketorder to IB and get filled to the normal price.In this way there is no problem.
On the other side if you have a takeprofit-marketorder then MC "think" the profit is reached and send the marketorder for takeprofit
Both market and price orders will be executed at normal IB prices.

bowlesj3
Posts: 2180
Joined: 21 Jul 2007
Has thanked: 227 times
Been thanked: 429 times

Re: Strange spikes in the e-mini from IB's TWS

Postby bowlesj3 » 09 Aug 2011

Hi Dave,

I have not tried then "Generate new tick if total volume changes" option. Question? What is better (seeing as I am a discretionary trader and do not get effected by this ticks because my order is only put on TWS and also only put there when I consider all the information to be valid)? Does the current default setting (that occasionally gives a strange price bar) generally have more accurate charts. If the charts are the same and only the volume differs I guess I should use the option you suggest. Will the MC people let us know when we can switch back?

Thanks,
John.

User avatar
Dave Masalov
Posts: 1712
Joined: 16 Apr 2010
Has thanked: 51 times
Been thanked: 489 times

Re: Strange spikes in the e-mini from IB's TWS

Postby Dave Masalov » 11 Aug 2011

What is better (seeing as I am a discretionary trader and do not get effected by this ticks because my order is only put on TWS and also only put there when I consider all the information to be valid)? Does the current default setting (that occasionally gives a strange price bar) generally have more accurate charts.


The charts look pretty much the same. Sometimes (very rare) they can look a little different because with this option on ticks are generated within MultiCharts on total volume change basis.
Will the MC people let us know when we can switch back?
It depends on how fast IB solves this "spikes" issue in their API.


Return to “MultiCharts”