|
PureBytes Links
Trading Reference Links
|
Greetings Everyone,
I just received this notice from a friend of mine, and I thought that the group might be interested, especially for you mechanical systems traders out there.
As of the start of this week, on Monday, October 5, 2009, the CME has started "unbundling" some large transactions that were previously sent as a single "bundled" trade tick. The net consequence of this is that while the volume remains the same as it would have been, the number of trade ticks to be transmitted, stored, and processed have approximately doubled. CME claims this is an improvement, and will get you more information faster. They're right on target about the "more information" part - it will take more bandwidth to send it, more disk space to store it, and more CPU to process it. The day it was implemented, the total number of trade ticks to transmit, store and process on CME approximately doubled over the previous day from 165K to 315K (see cumulative tick volume for the day in the bottom pane). The computational performance impact is most severe for "update by tick" trading systems, which could see increases in their latencies as a result of this increased amount of data to be processed in order to get the same total volume as before.
CME's notice of this change is quite technical and some say makes this less than transparent what they've done
Quote:Please note: as a result of the reduced message response times, bandwidth utilization in these futures markets may ultimately increase by as much as 20%.
To the extent that the CME was bundling trades together that didn't originate together from the same source at the same time (if it was doing this) then unbundling them would be generally good, and provide greater granularity for instance for use in building N contract bars accurately, and this is an argument they've presented for making the change. However, there is an important flip side - to the extent that the CME is breaking up ("unbundling") larger block trades now, there could be less clear visibility of something important - one of the key ways of distinguishing key large players from the masses. Some might have said that leaving the transactions bundled was giving away too much good information, hence, not exactly a lot of institutional resistance to making this change which makes it harder for retail traders to keep up with what's happening - both because of reduced visibility in the data itself and because of once again doubled data requirements to transmit, store, and process.
Important: if you are using "tick" bars to do your analysis, after the change the same number of ticks will represent a shorter period of time because of this difference (therefore, there will be more tick bars per day). Bandwidth requirements will have increased, and loading historical data will take longer. The average size per trade will have gone down in exact coincidence with the increase in the number of trade ticks, because the total volumes remain the same. If your analysis relies on analysis of specific trade sizes, especially for larger trades the CME may previously have transmitted together as a bundled transaction, this will need to be carefully re-examined in light of these new practices at the exchange.
This change affects basically all data feeds in a similar way, in that it's a change at the CME exchange, and is being made before the data is passed on to the data feeds for redistribution.
Having the ability to see clearly what's different and respond to it effectively can be quite important during this transition time as the market place is absorbing this change, and not everyone understands yet what's happened.
In summary, this is the new reality - you can still extract good information from the new data, but it's important to recognize what it is exactly that has changed and when it changed, so that you can understand the steps you may need to take for your specific trading methodologies, and to make sure you are understanding what it is you're seeing in these increased tick volumes.
__._,_.___
![]()
__,_._,___
|