Manfredi, Albert E wrote: >Mark Schubin wrote: > > > >>But here are some year-to-date figures I am permitted >>to release for what CEA calls "digita television": >> >>Through the 43rd week - DOWN 1.3% >>Through the 42nd week - DOWN 0.4% >>Through the 41st week - UP 1.2 % >>Through the 40th week - UP 1.5% >>Through the 39th week - UP 2.9% >>Through the 38th week - UP 3.6% >>Through the 37th week - UP 4.4% >>Through the 36th week - UP 5.2% >>Through the 35th week - UP 5.6% >>Through the 34th week - UP 6.7% >> >>Do you not see a trend there? >> >> > >Yes, I see a single-percent-digit INCREASE for "digital TV" >products until the 42nd week, followed by a slight downturn >on the 42nd and 43rd weeks. Agreed? So not much change, week >to week, for sales of so-called "digital." > > > >>Well, here are the year-to-date numbers for analog for >>the same period. I'll use Brinkley's (and CEA's) figures >>without the combo sets, because those percentages are more >>readily available: >> >>Through the 43rd week - down 13.2% >>Through the 42nd week - down 12.5% >>Through the 41st week - down 12.3% >>Through the 40th week - down 12.3% >>Through the 39th week - down 12.7% >>Through the 38th week - down 12.6% >>Through the 37th week - down 12% >>Through the 36th week - down 12.3% >>Through the 35th week - down 13.5% >>Through the 34th week - down 13.6% >> >> > >And here I see a double-percent-digit, steady drop in sales >of analog sets, between 12 and 13.6 percent, from the 34th >through the 43rd week. A steady and significant erosion of >sales of analog sets. > > > >>Now, maybe you have some fascinating mathematical tool that >>will come to a different conclusion, and I eagerly await >>your sharing it with us. But, using simple arithmetic, >>I'd say "digital television" has been steadily declining >>over the past ten weeks, while analog, give or take a >>hiccup, has been roughly steady. >> >> > >Perhaps I don't understand what up and down mean. If "digital" >is steady, more or less, while analog goes down significantly, >one would expect analog to reach zero faster than digital? > >The year-to-date drop of digital by 1.7 percent, compared with >last year, seems to confirm this. The analog drop, compared to >last year at the same time, was much higher than 1.7 percent. >The analog drop was 23.4 percent. So what am I missing? > >Bert > Bert Mark is probably pulling his hair out on this one. Have you ever taken a math class or statistics? Do you know what a trend is? The analog sales figures show basically a yearly decline to day of 13%, not much of a trend. You would have to look at a number of years to see a trend for those figures. But the digital sales show a "trend" this year. That trend is down week by week. First it is a trend down in the positive but in the last two weeks it has gone negative. As Mark suggest at this rate digital could be declining faster than analog soon. The average weekly change in sales for digital is .88% for the 10 weeks listed. That is the average. A better statistician would discover an increase in rate of decline I suspect but if we just go on the average, the decline in sales, if this "TREND" continues would be 9.22% for the 52nd week of this year. 7.92% cumulative plus the 1.3% already in place for week 43. A simple, statistically insignificant, way to look for a rate of increase would be to add up the differences for the first five weeks in sales of digital, compute an average and compare to the average of the last five weeks. Those figures are .76% for the first five weeks and .98% for the last five weeks compared to the 88% average for the combined 10 weeks. That says the rate of decline in sales is increasing. If we accepted the .98% as the rate for the rest of the year we would have a decline in sales in week 52 of 9 X .98% + 1.3% or 10.12%. Of course since the rate is increasing .98% is low so the real decline in week 52 would be more than 10.12% or could as Mark suggest top the average, no trend, drop in analog sales of 13%. But then a real statistician might detect that the rate of change is not just increasing but that it is accelerating and then all bets are off. We could be down digitally 50% for Christmas. Bob Miller ---------------------------------------------------------------------- You can UNSUBSCRIBE from the OpenDTV list in two ways: - Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org - By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word unsubscribe in the subject line.