> 3. I have the feeling that if one closes JBT (or if it crashes for any > reason) then all the data is lost i.e. if a strategy was running then it is > not possible to carry it on. My point is how to handle strategies that haave > overnight positions? Do we have to let JBT run permanently and hope it won't > crash? >
To adress this issue for my personalized setup, I am considering giving anything that needs it an initial value. The easiest way is to just hardcode it, and make it all a bunch of final variables at the top of your strategy class. This creates a couple of challenges though. First you need to get values to initialize it with, this is a challenge in its own right. You also need to be able to get the data into JBT. I recommend keeping the datapoints as few as possible. For indicators that run multiple days deep, consider using just the closing price as an input to that indicator. They are easy to look up, and much easier to enter than 370,000 price snapshots. All of this, of course, requires heavy modding. JBT wasn't designed with these operating conditions in mind, but I think it can be tweaked for whatever you personally desire. You have to be willing to get your hands dirty. On Tue, Aug 11, 2009 at 5:15 AM, nonlinear5 <[email protected]>wrote: > > > 1. It looks like the strategy is run by a scheduler and not triggered by > > the market events (tick update - orderbook change...). Why this choice? > How > > could I tweak it so that the strategy reacts on every ticks? > > > > JBT is driven by changes in market depth, not by changes in price. For > liquid instruments, market depth changes about 50 times per second, > and each change may involve multiple bid and ask levels. During a > single day, the total number of changes is literally millions. That is > way too much data to handle, especially for optimization. To make it > manageable, JBT uses 1-second snapshots of market depth. The snapshot > averages and normalizes the market depth changes which occurred in the > last second. These snapshots is what JBT uses for all its modes of > operation, which are backtesting, forward testing, trading, and > optimization. If you want tick-based trading, JBT is probably not for > you. > > > > 2. In the scheduler it looks like the code is not optimal (speedwise) but > > there must be a reason I guess. > > In what way is it not "optimal"? > > > >Why do we write the data and then only run > > the strategy code? > > I don't understand this question. > > > > 3. I have the feeling that if one closes JBT (or if it crashes for any > > reason) then all the data is lost i.e. if a strategy was running then it > is > > not possible to carry it on. My point is how to handle strategies that > haave > > overnight positions? Do we have to let JBT run permanently and hope it > won't > > crash? > > > > JBT is meant to be used for day-trading only. > > > > 4. I think some of you purchased CME data. What kind of information is > there > > (how many order book levels and which contracts)? Is it possible to "take > a > > share" in this data or future data? > > The CME data that we purchased with the project funds is 5-level deep. > IB has recently changed the depth to 10 levels, so the CME data > probably should not be used. CME is planning to publish 10-level > historical depth data soon, and we may purchase it then. > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "JBookTrader" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jbooktrader?hl=en -~----------~----~----~----~------~----~------~--~---
