[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Are You Seeing INDU Glitches?



PureBytes Links

Trading Reference Links

For more than a year (don't know when it started) I have noticed occasional
intraday glitches in the value of INDU.  I am on BMI/cable and one other
BMI/cable subscriber has reported similar glitches in one or more individual
stocks.  I would like to hear from some other subscribers, both BMI/cable
and other different feeds, to determine where in the data pipeline these
glitches are occurring.

The normal behavior of INDU is that the average move from one tick to the
next is less than a point, but when a glitch occurs a single tick jumps up
(occasionally down, but more often up) by 10, 15 points or more and the next
tick returns to the former value.  This is probably only noticeable on a
one-tick line chart and on such a chart the anomalous behavior is strikingly
apparent.  (On even a one-minute bar chart, the glitch will be somewhat
obscured by actual short-term changes in the index.)  On a one-tick INDU
chart the line will trace across the chart with just small random
up-and-down jiggles, but a glitch will appear as a sharp one- or two-tick
spike which stands WELL above the random jiggly noise.

Such glitches seem to appear in bunches.  Yesterday, for example, the index
behaved normally from open until 11:45 EST.  During that 11:45 minute a
single tick jumped up nearly ten points and the following tick returned to
the previous value.  Then during the next 22 minutes about twenty such
glitches occurred and then behavior settled back to a more normal state for
about an hour.  Also, during that period from 11:45 till 12:07 even the
"normal" intervals between glitch spikes seemed much more erratic than truly
normal behavior.  In other words, a moving average of tick-to-tick
differences would probably show a significant rise during that period, even
with the glitches filtered out.

Then, at 13:00 an even more abnormal incident occurred - the index dropped
(in a single tick) from 10584 to 10555, traced a normal-appearing pattern at
that lower level until 13:02, then jumped (in a single tick) back up to
10584.

Incidentally, I have changed the data precision for INDU in the Server from
1/100 to 1/10 so that is not a factor.

Please let me know what you show for yesterday, 1/8/2001.

Thanks.

Carroll Slemaker