JBookTrader release 6.08

[nonlinear5]

Latest release download: http://jbooktrader.googlecode.com/files/JBookTrader-6.08.zip
Release Notes: http://code.google.com/p/jbooktrader/wiki/ReleaseNotes
Historical data (requires a password): http://jbooktrader.googlecode.com/files/ES-Jan2007toNov2008.zip
Historical data (free sample): http://jbooktrader.googlecode.com/files/ES-January2009.zip

Changes in this release:

-- Upgraded third party libraries: ibapi from 9.51 to 9.6, jcommon
from 1.0.14 to 1.0.15, and jfreechart from 1.0.11 to 1.0.12

-- Better handling of disconnect/reconnect events, order placement,
and executions. In the previous releases, under certain (rare)
circumstances, it was possible to get a "stuck" order, if it was
placed after the disconnection. This latest release uses a new IB API
which provides means to gracefully handle these situations.

-- Market depth detected and processed up to level 10. You don't need
to make any changes to your code, except possibly change the strategy
parameters.

-- Fixed a problem in the "do not log/report/send duplicate messages"
logic


Since this release uses a new IB API, testing is recommended if you
are trading live. As a reminder, the strategies themselves are best
tested using a real account and the "forward test" mode. In this mode,
real market data is used, but trades are simulated. To test order
placement, trade execution, connectivity events, and all other aspects
of live trading, use a simulated (paper trading) account and run the
strategies in the "trade mode".

On the last note, JBT has approached maturity. The bugs have been
shaken out, the releases are less frequent, and the code changes are
less and less dramatic. This is all good news, as we seem to be
settling on the stable and well tested platform. Please keep up your
testing efforts and do not hesitate to report problems, even if you
only suspect there are any. It's better to report too much than too
little. Thanks again to everyone for participating in this project.

[Crichton McCutcheon]

Eugene,
Any thoughts on how the 10 level book may effect strategy params? Since
we won't have much if any backtest data that was derived from 10 level book
recording, we can't optimize to it.

[nonlinear5]

> Any thoughts on how the 10 level book may effect strategy params?

My guess is that the range of depths balances will be smaller, when
all 10 levels are taken into account. That probably means that the
strategy params would need to be smaller as well, to trigger the same
amount of trades.

[Javier]

Hi Eugene,

I've been testing this release (6.08) since last week and I've found
some problems with live quotations after IB servers restart.

When I run more than one strategy at the same time with different FX
symbols all is working fine until IB servers restart. After that, only
one strategy keeps live quotations, the others remain freezed with the
same price all the time until you restart manually JBT.

My EventLog file report, as always after IB servers restart:

"1102: Connectivity between IB and TWS has been restored - data
maintained."

but nothing more regarding the problem I've commented.

[Eugene Kononov]

When I run more than one strategy at the same time with different FX
symbols all is working fine until IB servers restart. After that, only
one strategy keeps live quotations, the others remain freezed with the
same price all the time until you restart manually JBT.

Which FX symbols? I'd like to try to reproduce this problem.

[Javier]

I used EUR.USD and GBP.USD.

I downgraded to v6.07 and all is working fine, maybe it's a problem
with the handling of disconnect/reconnect events of v6.08

[Eugene Kononov]

I downgraded to v6.07 and all is working fine, maybe it's a problem
with the handling of disconnect/reconnect events of v6.08

Can you please post or send me your EventReport.htm for the day when the problem occurred?

[Crichton McCutcheon]

Has anyone observed any differences in the "book feed" with the
10 level book ? I'm still using the 5 level book.

[Javier]

> Can you please post or send me your EventReport.htm for the day when the
> problem occurred?

EventReport file (v6.08):

01:00:03 04/13/09 CEST Email notification sent
02:00:02 04/13/09 CEST Email notification sent
03:00:02 04/13/09 CEST Email notification sent
04:00:02 04/13/09 CEST Email notification sent
05:00:02 04/13/09 CEST Email notification sent
05:58:57 04/13/09 CEST 317: Market depth data has been RESET. Please
empty deep book contents before applying any new entries.
05:58:57 04/13/09 CEST Market depth data has been reset.
05:58:57 04/13/09 CEST 2103: Market data farm connection is
broken:cashfarm
05:59:00 04/13/09 CEST Email notification sent
05:59:08 04/13/09 CEST 2104: Market data farm connection is
OK:cashfarm
06:00:02 04/13/09 CEST Email notification sent
06:00:55 04/13/09 CEST 317: Market depth data has been RESET. Please
empty deep book contents before applying any new entries.
06:00:55 04/13/09 CEST Market depth data has been reset.
06:00:55 04/13/09 CEST 2103: Market data farm connection is
broken:cashfarm
06:00:57 04/13/09 CEST Email notification sent
06:02:26 04/13/09 CEST 2104: Market data farm connection is
OK:cashfarm
06:04:28 04/13/09 CEST 317: Market depth data has been RESET. Please
empty deep book contents before applying any new entries.
06:04:28 04/13/09 CEST Market depth data has been reset.
06:04:28 04/13/09 CEST 2104: Market data farm connection is
OK:cashfarm
06:14:16 04/13/09 CEST 1100: Connectivity between IB and TWS has been
lost.
06:14:18 04/13/09 CEST Email notification sent
06:14:34 04/13/09 CEST 1102: Connectivity between IB and TWS has been
restored - data maintained.
06:14:36 04/13/09 CEST Email notification sent
07:00:02 04/13/09 CEST Email notification sent
08:00:02 04/13/09 CEST Email notification sent
09:00:02 04/13/09 CEST Email notification sent
10:00:02 04/13/09 CEST Email notification sent
11:00:02 04/13/09 CEST Email notification sent
12:00:02 04/13/09 CEST Email notification sent
13:00:04 04/13/09 CEST Email notification sent
14:00:03 04/13/09 CEST Email notification sent
15:00:02 04/13/09 CEST Email notification sent
16:00:02 04/13/09 CEST Email notification sent
16:54:00 04/13/09 CEST Msg ID: 1239595718 Msg Type: 1 Msg:
====================================================== To TRADEWEB
traders: Mon Apr 13 10:52:44 2009 EST TRADEWEB is currently
unavailable for trading due to technical problems at the exchange.
Exchange: TRADEWEB
17:00:02 04/13/09 CEST Email notification sent
18:00:02 04/13/09 CEST Email notification sent
19:00:02 04/13/09 CEST Email notification sent
20:00:02 04/13/09 CEST Email notification sent
21:00:02 04/13/09 CEST Email notification sent
22:00:02 04/13/09 CEST Email notification sent



MarketData for EUR.USD (v6.08):

041309,055657,7,1.31655
041309,055658,7,1.31655
041309,055659,-5,1.31652
041309,055700,-2,1.31652
041309,055701,0,1.31652
041309,055702,0,1.31652
041309,055703,0,1.31652
041309,055704,0,1.31652
041309,055705,0,1.31652
041309,055706,0,1.31652
041309,055707,0,1.31652
041309,055708,0,1.31652
041309,055709,0,1.31652
041309,055710,0,1.31652
041309,055711,0,1.31652
041309,055712,0,1.31652
[...]


As you can see in MarketData file fragment, after 05:57:00, all the
entries are the same:

[date],[time],0,1.31652

[nonlinear5]

> As you can see in MarketData file fragment, after 05:57:00, all the
> entries are the same:
>

There was a series of resets between 05:57:00 and 06:14:34. Did the
market data resume for your FX symbols after that?

[Javier]

> There was a series of resets between 05:57:00 and 06:14:34. Did the
> market data resume for your FX symbols  after that?

No, the market data has the same value during the rest of the day.

[Eugene Kononov]

OK, I'll see if I can reproduce this problem at the Sunday/Monday reset (at midnight EST).

[nonlinear5]

> OK, I'll see if I can reproduce this problem at the Sunday/Monday reset (at
> midnight EST).
>

I started a forward test with 3 strategies: EUR.USD, GBP.USD, and
USD.JPY. They are running normally, with regular price updates. We'll
see what happens when the market data is reset at midnight EST.

[nonlinear5]

I was able to reproduce the problem. After the midnight reset, the
book data for 2 of the 3 FX symbols appeared stale. I'll investigate
why this is happening. In the meanwhile, the short term workaround is
to restart JBT sometime after the IB reset.