Post Hectic Environment Period
After the turbulent period of June to mid-August 2007, as mentioned in the previous article, the explosive growth in real-time data has retreated some what. Here is an update about the current situation.
Before and After
Before June 2007, the emini S&P (for those not familiar with the instrument – it is the benchmark of trading volume in terms of the number of trades and the number of closest bid/ask updates in the US market ever since its introduction) has about 300K updates, combining trades and bid/ask updates. During June 2007, it surges to around 500K a day. During July 2007, it breaks its own record again by having more than 750K updates a day. In mid-August 2007, we have seen more than 1.1 million updates within a day.
By the end of the week ending on Friday August 24, 2007, the number of updates dropped back down to the range of 600K to 700K a day.
What Has Changed
From our observation, for most of the common stocks, index futures, etc., the old golden ratio of having 2 to 3 bid/ask updates per trade in real-time no longer holds. The new ratio stands at 5 to 6 bid/ask updates per trade during July and early August, and that still holds true now.
What does this increase in bid/ask updates per trade imply? The price discovery process in real-time is getting less efficient and that there is an increase in uncertainty to commit to a particular price point to complete a transaction.
Volatility Induces More Volatility
One old saying among traders is that “Volatility Induces More Volatility”. What it means is that once a market started to get more volatile, the volatility would keep increasing until some form of exhaustion occurs. Then volatility would slowly reduce over time, not drop dead all of a sudden like what we have seen in the past week.
Since we have seen volatility increased so much since June, it is not likely that it would stop abruptly now. For a trader, you should better prepare yourself for more volatility until it gradually decreases over time.
Well Known Trading Tool Affected – Tick Index
The well-known TICK index has been rendered useless during this hectic environment due to the explosive growth in number of trades in real-time. Every few mintues, we saw another extreme reading in TICK index. In the old days, TICK index could provide important clues in real-time trends. Now, TICK index readings are more like pure noise.
Our TICK16 indices based on various baskets of index components are affected too. Although TICK16 indices can still maintain some form of trends, they are getting much less useful in terms of identifying change in trend. That, we will resolve in upcoming release of NeoTicker to adapt to this new hectic environment.
In retrospect, the environment change makes a lot of sense. The predecessor of TICK16 was TICK10, which I invented back in 1990s, was good enough at that point in time and worked very well for over a decade. Then, TICK16 was invented to deal with the trading environment since year 2000. As we enter a new era of trading, it is obvious that we will need something better to deal with the current situation.
Adjustments to My Trading Computer
To preserver CPU power for handling the massive amount of real-time data, I have reduced the real-time update frequency across all my charts with higher time frames (e.g. 45-min, hourly, daily, etc.) to update every 1 or 2 seconds only. The indicators on these charts are calculated on every tick arrival, but the visual part, is done in lower frequency so that the computer can use most of its power on processing the real-time data.
One thing I discovered during my quest to improve efficiency on my trading computer is that my quote windows all have row filtering formulas enabled by default. I do not use filtering at all myself in those quote windows! That eats up some CPU load as I have many symbols tracked in my quote windows. After turning that off, I see significant improvement in the responsiveness of these quote windows.
Good luck trading!