In order to give as objective an answer as possible, I am looking at two years worth of daily data.
Today AMD closed at
-.85 of the Bollinger 50 (where lower band is -1, MA50 is 0)
-.65 of the Bollinger 20
There are 38 days in the past two years where AMD closed
in the range: Bollinger 50: -1.05 to -0.6
and Bollinger 20: -0.85 to -0.4
Only one of these days (4/15/2005) was within 15 days pre earnings. Usually we are much higher in both the BB20 and BB50 prior to earnings. The only exception is after the Jan 2005 Flash warning. So we are unusually low in the bands prior to earnings.
Of these 38 days:
The mean change in 5 days is: +1.1%
The mean change in 10 days is +3.9%
The mean change in 15 days is +5.9%
Within the next 15 days:
15/38 40% gained >10% before losing 10%
11/38 29% gained 5-10% before losing 5%
3/38 8% stayed within +/-5%
8/38 21% lost 5-10% before gaining 5%.
1/38 3% lost more that 10%
So, there is only a 1 in 4 you will be able to
buy back 5% lower than today's close.
Within the next 30 days:
24/38 61% gained >10% before losing 10%
5/38 13% gained 5-10% before losing 5%
3/38 8% stayed within +/-5%
5/38 13% lost 5-10% before gaining 5%.
1/38 3% lost more that 10%
The worst drop was 7/22/2004 of -13.8% (after falling 10% and rebounding)
The next worst was 4/15/2005 dropping 9% in 10 days
The best Gain was 1/24/2005, up 20% in 15 days
Next best gain was 1/19/2005, up 16% in 15 days.
Ticker Quote 15 days Price Price
Date Pre Earn- Change Change
5 days Fwd Fwd Fwd
ago 5days 10days 15days
AMD 4/3/2006 No 6.5
AMD 3/31/2006 No 9.0
AMD 3/30/2006 No 3.5
AMD 3/28/2006 No 1.0
AMD 3/23/2006 No - 1.9 - 3.3
AMD 3/22/2006 No 2.2 - .2
AMD 3/21/2006 No 1.6 - 1.0
AMD 3/20/2006 No - .4 3.5 - 2.8
________3/17/2006 No 7.9 6.4 - 2.3
AMD 5/3/2005 No 1.8 4.9 9.1 12.3
AMD 5/2/2005 No 2.7 5.1 9.4 11.4
AMD 4/29/2005 No 2.6 4.7 8.8 12.4
AMD 4/28/2005 No 5.9 4.1 8.1 12.0
AMD 4/27/2005 No 1.0 1.2 4.0 9.0
AMD 4/26/2005 No 4.5 - 1.8 3.1 7.2
AMD 4/25/2005 No 4.0 - 2.6 2.4 6.6
AMD 4/22/2005 No 5.8 - 2.5 2.0 6.0
AMD 4/21/2005 No 8.5 - 5.5 - 1.7 2.1
AMD 4/19/2005 No 12.8 - 4.3 - 6.0 - 1.4
AMD 4/18/2005 ?? 11.4 - 3.8 - 6.3 - 1.5
______ 4/15/2005 Yes 10.6 - 5.5 - 7.9 - 3.6
AMD 3/29/2005 No 2.1 2.8 11.7 - .9
______ 3/28/2005 No 3.9 3.1 10.0 - 1.2
AMD 1/25/2005 No - 1.6 3.3 12.0 12.3
AMD 1/24/2005 No .0 4.2 13.6 20.3
AMD 1/21/2005 No - 2.1 1.4 12.9 16.9
AMD 1/20/2005 No - 4.0 3.1 6.6 13.5
AMD 1/19/2005 No - 4.1 3.1 7.5 15.7
______ 1/18/2005 No 29.9 1.6 5.0 13.8
AMD 8/16/2004 No - .6 9.1 3.1 - 2.4
AMD 8/13/2004 No 1.8 8.2 8.1 - 1.5
AMD 8/12/2004 No 8.4 7.8 8.4 7.2
AMD 8/10/2004 No .7 - 1.2 3.7 - .8
AMD 8/9/2004 No 10.0 .6 9.8 3.8
AMD 8/6/2004 No 10.9 - 1.8 6.3 6.2
AMD 8/3/2004 No .1 - .7 - 1.9 3.0
AMD 7/29/2004 No 6.8 - .3 - 7.9 - .8
AMD 7/22/2004 No 2.6 - 6.3 - 6.6 - 13.8
Looking at all 2 years again (no filter on the Bolinger).
There are 156 days where the MA20 < MA50.
There is a 65% chance of 5% or greater gain in 15 day compared to a 35% of a 5% or greater loss.
Mean 15day price change is +4%
Filter on MA20<MA50, MA20 negatively sloped:
get the same thing. 100 days.
Reset
Filter on days where the MA50 slope just turned negative:
Surely the kiss of death: The slope went negative about 5 days ago. So filtering on days 0 to 20 days from a noseover...
55% for a >5% gain in 15 days vs 40% for >5% loss.
Expected value is +2% gain. Not a kiss of death.
Reset
Filter on days where in the prior 15 days we are within 1% of the low and had been as much as 10% higher.
29 Days.
41% for a >5% gain in 15 days, 58% for >5% loss.
Expected value change is -2% in 15 days.
Today the MACD is negative, downward sloping, the histogram is negative (bearly) and downward sloping. 43 days.
44% for >5% gain in 15days, 46% for >5% loss.
Expected value is +2%
But we are very close to a positive cross.
When the MACD goes positive, upward sloping with positive and growing histograms:
80% for >5% gain, 50% for >10% gain, 22% for > 5% loss,
Expected value gain is 9.5% in 15 days: (82 days).
There are 61 days where the day's low is lower than all previous 9 days:
It is about 50/50 whether a >5% gain or >5%loss,
it is 25% vs 30% for a >10% gain or >10% loss.
but Expected value is -2% in 15 days.
Of these 61, those with a higher close (17 days):
55% vs 45% for >5% gain vs loss in 15 days
30% vs 30% for >10% gain vs loss in 15 days
Expected value is -2%.
One last one:
Days where the Low < Bollinger 50 Low (like today)
55% vs 37% for >5% gain vs loss in 15 days.
Worst case: 7/14/2004 for a 18% loss in 10 days
Best case: 1/24/2005 for 20% gain in 15 days and
7/28/2004 for a 10% gain in 2 days (with a fall off later).
And if it closes higher after tagging the BB50Low (like today)
Then it is 70% vs 23% for >5% gain vs loss in 15 days.
and 46% vs 15% for >10% gain vs loss.
Expected value is 6.7% in 15 days from 13 cases.
Conclusion:
I think the odds favor it going up, but not immediately to the moon. The expected values are positive but risk is better that 2 to 1 for a gain.
Based upon the place in the Bollingers, it is unlikely to fall another 5% in two weeks (prior to earnings).
The case can be made however, to wait until the MACD goes positive and then pile in.
My take on statistical technical analysis.
How do you like them entrails?
-Magrathea |
