969 replies to this topic

[johnmhinds]

The race took place at the same time as last year actually.

Last years race was rain delayed though so was on the air longer and at a better viewing time for Europe.

[Crusoe]

Last years race didn't have a Brit on pole.

[DutchCruijff]

jakehumphreyf1 Jake Humphrey
This weekend a peak audience of 7.2m people tuned into the first ever Indian Grand Prix. In TV terms those figures are the size of Everest.

Anything to back this up?

[D.M.N.]

That'll be combined live + re-run peak. Should mean for some good looking averages though...

[D.M.N.]

Indian GP second most watched race of the entire season... 5.55m average for race, 3.42m for Qualifying...

Indian Grand Prix - Ratings Roundup
2011
Qualifying - 2.35m (30.8%) and 1.17m (12.0%)
* peaks: 2.74m @ 10:25 and 1.31m @ 13:50
* combined peak: 4.05m
* most-watched Qualifying of the year
* highest Qualifying rating since the 2009 Brazilian Grand Prix
Race - 4.14m (39.6%) and 1.41m (10.7%)
* peaks: 5.55m @ 09:35 and 1.79m @ 15:55
* combined peak: 7.34m
* second most-watched Race of the year

Stunning ratings for the weekend, huge figures there.

Thanks to rzt for the ratings.

Edited by D.M.N., 02 November 2011 - 15:38.

[Ross Stonefeld]

Combined peak is a bullshit figure unless you(not you specifically DMN) can prove no one watched both broadcasts.

And 5.5 seems high for a morning race?

[D.M.N.]

Combined peak is a bullshit figure unless you(not you specifically DMN) can prove no one watched both broadcasts.

And 5.5 seems high for a morning race?

This may seem an odd comment to make... but I wonder if there's a large (or substantial) Indian population that live in England that tuned into watch the race specifically because it was coming from India? The numbers are high.

And yes, combined figures are quite a crap figure, however given the reduction in output for 2012, I can see why Jake would want to quote the big figures this year.

[arknor]

This may seem an odd comment to make... but I wonder if there's a large (or substantial) Indian population that live in England that tuned into watch the race specifically because it was coming from India? The numbers are high.

And yes, combined figures are quite a crap figure, however given the reduction in output for 2012, I can see why Jake would want to quote the big figures this year.

probably all the people buying brtish VPN's so they can watch the race on iplayer

[engel]

probably all the people buying brtish VPN's so they can watch the race on iplayer

not unless they also bought a BARB people meter along with their VPN ;)

[NissanGTR]

Remarkable figures, truly shows that F1 is our national sport now not football.

A Sunday morning of 9.30 where most of the country is hungover from binge drinking the night before for 5m to be turning on the tv to see a race is nothing short of remarkable.

All the more remarkable when there is not a British driver in contention for the world title

[dank]

This may seem an odd comment to make... but I wonder if there's a large (or substantial) Indian population that live in England that tuned into watch the race specifically because it was coming from India? The numbers are high.

And yes, combined figures are quite a crap figure, however given the reduction in output for 2012, I can see why Jake would want to quote the big figures this year.

I'd put it down to the fact it was on at a time of day when there's bugger all else to do (apart from Steve Wright's Sunday Love Songs, of course...). Grands prix tend to get in the way of Sunday lunch and other chores at their regular time.

[Andy865]

Irritating considering the race was crap. I shall now have to endure f1 is rubbish banter at the public house.

[Risil]

And 5.5 seems high for a morning race?

9.30/10.30am isn't that early though. Can't think of many Grands Prix that have gone out at that sort of timeslot, at least not for Britain. But Brazil has no problem getting TV audiences for what is presumably a 10am live broadcast for a lot of the country.

Edited by Risil, 02 November 2011 - 17:09.

[dank]

Brazil's a 5pm start here in the UK. Again, a more convenient time of day for Joe Public to do what they need to do and settle down to watch the telly for two or three hours.

Edited by dank, 02 November 2011 - 17:33.

[Risil]

Brazil's a 5pm start here in the UK. Again, a more convenient time of day for Joe Public to do what they need to do and settle down to watch the telly for two or three hours.

I meant, Brazilians watch European races at 10am, and I don't think that's resulted in problems for live broadcasts.

[Ensign]

I suppose 10am (11am on the continent) isn't a bad time slot as most people are still in the house at that time and, more importantly perhaps, it's before the football matches begin all across Europe.

[D.M.N.]

15-minute breakdown:

08:30 - 2.8m (34.5%)
08:45 - 3.3m (36.6%)
09:00 - 3.8m (38.7%)
09:15 - 4.4m (41.7%)
09:30 - 5.5m (47.8%)
09:45 - 5.3m (46.8%)
10:00 - 5.2m (45.6%)
10:15 - 5.2m (45.2%)
10:30 - 5.0m (43.5%)
10:45 - 5.0m (42.9%)
11:00 - 4.5m (40.5%)
11:15 - 3.4m (33.0%)
11:30 - 2.6m (27.3%)
11:45 - 2.1m (22.1%)

[D.M.N.]

Abu Dhabi Grand Prix - Ratings Roundup
2009
Qualifying - 2.54m (27.3%)
Race - 4.8m (37%)
* peak: 5.9m (44%) at 13:30

2010
Qualifying - 2.71m (27.3%)
Race - 5.78m (41.4%)
* peak: 7.35m (46.3%) at 14:40

2011
Qualifying - 2.90m (29.7%)
* peak: 3.62m at 14:00
Race - 4.57m (36.4%)
* peak: 5.52m at 14:35

Lowest of the 3 Abu Dhabi races which doesn't really surprise me considering it wasn't the final race (2009) or a title decider (2010). Qualifying rated well, though.

Edited by D.M.N., 16 November 2011 - 16:14.

[HuddersfieldTerrier1986]

I'd also guess that the fact it's been such a pathetically dull race in the 2 years previous to this didn't help, as some people are likely to think "this race is boring as absolute hell, I'd rather watch a frog jump around in mud than watch the Abu Dhabi GP".

[D.M.N.]

Some sports ratings, including F1, from yesterday courtesy of MG: http://www.guardian....k-webber-f1-win

BBC One
15:00 - F1: Brazilian Grand Prix Live: 4.8m (27.2%)
* peak: 6.1m
23:15 - F1: Brazilian Grand Prix Highlights: 700k (7.8%)

Sky Sports 1
15:30 - Live Ford Super Sunday: 2.0m (10.5%)
* peak: 3.2m

Rather tame rating for the F1 to go out on considering the timeslot and the lowest Brazilian Grand Prix rating since 2005. Of course, as average as that is, it's still double a high profile football match on Sky Sports 1...

Will do a full F1 post in the next few days.

[Ross Stonefeld]

I think this is the first GP of Brazil since 2004 where the championships were already decided?

[Schumacher7]

Of course, as average as that is, it's still double a high profile football match on Sky Sports 1...

But what you seem to be forgetting is that a football match contains only two teams and as such people that are not fans of those two teams are unlikely to watch (barring very enthusiastic neutrals) where as in an F1 race that's all you get, everyone competes at the same time, it's not broken up into ten different matches so if you are a fan of F1 you'll probably be watching it. Although then also you factor in the overall number of fans of both sports and other things, it gets quite complex, what I'm trying to say in a long and convoluted way is that you can't really compare viewing figures across sporting events like this (although as I said earlier in the thread the viewing figures are purely estimates (I believe) and therefore could be completely different to actual viewing patterns).

[D.M.N.]

I think this is the first GP of Brazil since 2004 where the championships were already decided?

Yes, I think you're right, 2005, 2006, 2007, 2008 and 2009 were all decided at Brazil, 2011 wasn't decided at Brazil but was up for grabs.

But what you seem to be forgetting is that a football match contains only two teams and as such people that are not fans of those two teams are unlikely to watch (barring very enthusiastic neutrals) where as in an F1 race that's all you get, everyone competes at the same time, it's not broken up into ten different matches so if you are a fan of F1 you'll probably be watching it. Although then also you factor in the overall number of fans of both sports and other things, it gets quite complex, what I'm trying to say in a long and convoluted way is that you can't really compare viewing figures across sporting events like this (although as I said earlier in the thread the viewing figures are purely estimates (I believe) and therefore could be completely different to actual viewing patterns).

Viewing figures are more than just 'purely estimates' as you put it and are based on the viewing habits of 5,100 homes meaning that the margin of error is ~3% I think. As for your football example, not sure I agree - yesterday's game was Liverpool vs Man City, so I think fans of other clubs such as Man Utd, Chelsea and Arsenal would be watching as well as neutrals given that Man City are leading the table.

[D.M.N.]

Brazilian Grand Prix - Ratings Roundup
2008
Qualifying - 2.99m
Race - 8.8m (41%)
* 15-minute peak: 12.5m (49%) at 18:45
* 5-minute peak: 13.1m

2009
Qualifying - 4.4m (24%) [before switch to BBC2]
Race - 6.6m (35%)
* 15-minute peak: 8.9m (39%) at 18:30

2010
Qualifying - 2.4m (17.2%)
Race - 5.1m (27.5%)
* 15-minute peak: 5.9m

2011
Qualifying - 2.77m (19.4%)
* 15-minute peak: 3.4m (22.3%) at 16:30
Race - 4.8m (27.2%)
* 5-minute peak: 6.1m

-------------------------------------

With all the figures in for this season, the unofficial averages are:

- Race: 4.55m
* highest since 1998
- Qualifying: 2.73m
* highest since circa. 1996



Note: There is not much data available for Qualifying sessions between 1994 and 1998 and as thus the averages may be inaccurate.

Also, as no race on BBC has had more than 8.8m, it means that the 2008 Brazilian Grand Prix remains the most watched F1 race in the UK ever.

Edited by D.M.N., 30 November 2011 - 13:02.

[Schumacher7]

Viewing figures are more than just 'purely estimates' as you put it and are based on the viewing habits of 5,100 homes meaning that the margin of error is ~3% I think.

There are 25 million homes in the UK, taking 5100 as a sample size isn't nearly large enough, a good sample size should be at least 10% of the total group you are studying which would be 2.5million. I'm sorry but 5100 homes does not convince me at all and it is highly probable that these viewing patterns do not reflect the viewing patterns of the country as a whole ( although they could but it is unlikely).

[Grayson]

With all the figures in for this season, the unofficial averages are:

- Race: 4.55m
* highest since 1998
- Qualifying: 2.73m
* highest since circa. 1996

Those figures are truly amazing when you consider how much less fragmented the media was in 1996 and 1998. Back then, the two main channels put together would have had around 60-70% of all TV viewing between them, whereas the figure would be around half that today.

There were also three fewer Grands Prix back in 1998. I could be wrong, but it feels to me like it's easier for the average ratings for each GP to be very slightly higher if there are fewer in a year.

[engel]

There are 25 million homes in the UK, taking 5100 as a sample size isn't nearly large enough, a good sample size should be at least 10% of the total group you are studying which would be 2.5million. I'm sorry but 5100 homes does not convince me at all and it is highly probable that these viewing patterns do not reflect the viewing patterns of the country as a whole ( although they could but it is unlikely).

Statistical science begs to differ, lookup a sample calculator and you 'll understand you can get pretty small margin of error with reasonably high confidence from a 1:5000 sample. In fact it's pretty much the standard globally for TV viewership, it's reasonably accurate without being prohibitively expensive

[Amphicar]

There are 25 million homes in the UK, taking 5100 as a sample size isn't nearly large enough, a good sample size should be at least 10% of the total group you are studying which would be 2.5million. I'm sorry but 5100 homes does not convince me at all and it is highly probable that these viewing patterns do not reflect the viewing patterns of the country as a whole ( although they could but it is unlikely).

Actually 5,100 is a very large sample size - opinion polls typically use a sample size of 500 to 1,000 with a plus or minus 3% margin of error.

[Juan Kerr]

I cannot believe how people can be such short-sighted idiots. 'TV ratings have gone up due to the racing being closer' what a stupid thing to say by a brainless idiot, is that the only factors he could think of why ratings might go up ? What about all the other factors ?? No wonder no-one can judge anything properly anymore.

[Ross Stonefeld]

Id say graph the ratings against the performance of drivers from that country...

[Juan Kerr]

Well yes exactly, other factors like that, to just think its down to one is absurdly stupid.

[Schumacher7]

Actually 5,100 is a very large sample size - opinion polls typically use a sample size of 500 to 1,000 with a plus or minus 3% margin of error.

That's why opinion polls very often do not reflect the opinions of the majority of people however when they do it is because what the people are saying is common sense on a subject most agree on due to human values, TV programs however are purely down to the individuals personal tastes. Are you trying to tell me that 0.000204% is a high enough sample size?

[midgrid]

Looking at that graph makes me feel sorry for ITV.

[Ross Stonefeld]

Well they came in the year after Damon won the title(but cunningly signed for Arrows) and then had to wait a decade for Lewis.

[wrcva]

There are 25 million homes in the UK, taking 5100 as a sample size isn't nearly large enough, a good sample size should be at least 10% of the total group you are studying which would be 2.5million. I'm sorry but 5100 homes does not convince me at all and it is highly probable that these viewing patterns do not reflect the viewing patterns of the country as a whole ( although they could but it is unlikely).

as long as the "sample" is randomly drawn -- that is, there is no systematic bias of any sort in the way those 5100 were selected or the monitoring devices are distributed/deployed -- n=5100 is more than sufficient enough (even too big) to make projections about the "population" ==> country. The key assumption is the randomness of the sample drawn ...

Engel and Amphicar are spot on... Schumacher7's concern is valid as well that if the polling company is not paying attention to sampling process you'll have GIGO.

[hillfan]

Looking at that graph makes me feel sorry for ITV.

Serves them right for putting adverts in the races!

[madraykin]

as long as the "sample" is randomly drawn -- that is, there is no systematic bias of any sort in the way those 5100 were selected or the monitoring devices are distributed/deployed -- n=5100 is more than sufficient enough (even too big) to make projections about the "population" ==> country. The key assumption is the randomness of the sample drawn ...

Engel and Amphicar are spot on... Schumacher7's concern is valid as well that if the polling company is not paying attention to sampling process you'll have GIGO.

It's supposed to be a demographically representative sample I believe, so not random.

[hotstickyslick]

Serves them right for putting adverts in the races!

That is a joke, right?

[wrcva]

It's supposed to be a demographically representative sample I believe, so not random.

randomness will still apply, even with demographics based stratification (within cell). Simplified example...
They usually start with some sort of "representative" list that contains say million records -- could be taxpayer list, voter registration so on...no names but demographic variables such as household income, education of the head of the household and geo designation (street address, postal code so on).
You pull 51000 records (randomly) out of this list based on the most recent census proportions to maintain desired level of representativeness (city, state/province, region, or national). In other words, to recruit 5100 households you may have to start with 50-60K records or more (1:10 is an arbitrary example because not everyone will consent to install the monitoring/ratings device)...

Example:
- your list has 9% of the households at [income range of 100-125K and education college or above]=cell, but
- the census says that "cell" proportion is 7%
so you have to weight down (or up) the number of records you pull to make sure your final list of 5100 has the correct proportional representation as you randomly select sufficient number of records from the input list. In some countries (with no telephone penetration) may draw (randomly) street names and target all odd or even numbered buildings, to pick 1 from household from each building... it all depends on what you are dealing with...

Alternative method (if you do not have a list) will be to randomly generate the last two digits of the telephone numbers (from known working numbers, depending on country's dialing plans and geo representation in tel no), call that number, and first ask address, income, and education...and, try to reach your national quota (5100 HH) this way ... and subsequently apply the census proportions.

either way, you have to maintain randomness and worry about your assumptions that may impact representation. Example: if you only call land lines to construct your sample, you will not be covering households that use cell phones only and have no land lines... if you use online recruitment, you may not represent older folks or rural areas type of thing...

then, they typically ask households (selected by the methods above) to see if they are willing to install the rating device to their TV(s). Device reports the TV viewing habits of that household without names but just demographic variables of interest...by channel selected, time, duration with some sort of unique id... now, it could be that you are sleeping with the TV on channel X ... then, they may sometimes apply correction factors as such after surveying those households to assess what % might have that type of habit...

well, this is a very basic explanation, but if you do your sampling right, you can make pretty good projections about that population ...

[Snic]

Hmm I dont want to make this thread into some sort of statistics exam but doesnt the fact that the households know that they are being recorded effect the legitimacy of the study? Surely knowing you are being recorded may alter your viewing habits, for example you may not be watching Men & Motors at 1am (Not that I know whats on that channel at that time!) or even worse, be caught watchign The Jeremy C*** show. And if the same people are always used the fact they are being part of a study may alter the significance.

Anyway next years viewing figures will be very interesting... For example the 2005 ashes free-to-air on channel 4 had an average viewing figure of 4.7 million. The 2009 ashes excl. to Sky had an average of 856,000. If Sky's figures are that low that would surely drive advertising out of f1 fast. Or perhaps the BBC races will be watched by even more people as it is a once monthly thing and may feel like more of an "event". Any predictions on whatll happen?

Edited by Snic, 04 December 2011 - 20:52.

[D.M.N.]

Overnight figures suggest* that the Australian Grand Prix Qualifying rating for the highlights was the lowest since 2004:

Australian Grand Prix Qualifying Highlights Ratings
2004 - 1.48m (+0.41m live)
2005 - 1.73m (+0.41m live)
2006 - don't have figures
2007 - 1.72m (+0.37m live)
2008 - 1.70m (+0.38m live)
2009 - 2.53m (+0.80m live)
2010 - 1.93m (+0.81m live)
2011 - 1.96m (+1.08m live)
2012 - 1.60m (+???)***

Source for 2012 - Jake Humphrey on Twitter.

*** I should note that 1.60m may not include the BBC HD channel, which could change the figure drastically compare to 2010 and 2011 and may well move it in front of both years. If it does include BBC HD, then that's not good for BBC. Either way, it looks like the combined number is down, unless BBC HD added big numbers.

I think, once BBC HD and Sky F1 are factored in though (assuming the BBC2 rating excludes BBC HD), then it will only be the lowest since 2008. I can't see it getting near the 2.74m from 2010 and higher in 2009 and 2011 unless Sky F1 performed above expectations.

[dank]

How difficult will Sky's viewing figures be to come by?

[D.M.N.]

How difficult will Sky's viewing figures be to come by?

I think they will be reported with it being the first weekend. Hoping to post figures tomorrow evening. If not, we will have to wait until BARB officially updates a week on Monday with the official data taken over 7 days (ie, everyone that watched the 4:30am showing live, recorded, Sky+ etc over a 7 day period).

[D.M.N.]

Right, time to attempt a potentially invalid comparison, let's see how far we get!

Australian Grand Prix - Ratings Roundup
2008
Qualifying - 2.08 million (0.38m + 1.70m)
Race - 4.56 million (0.81m + 3.75m)

2009
Qualifying - 3.33 million (0.80m + 2.53m)
Race - 4.98 million (2.13m + 2.85m)

2010
Qualifying - 2.74 million (0.81m + 1.93m)
Race - 4.51 million (1.95m + 2.56m)

2011
Qualifying - 3.04 million (1.08m + 1.96m)
Race - 4.15m (2.13m + 2.02m)

2012
Qualifying
Sky Live - ???
BBC Highlights - 1.48m (13.4%) inc HD
* peak: 1.71m at 13:45
- combined lowest Australian Grand Prix Qualifying highlights figure since 2004
Total = ??? million

Race
Sky Live - 617k [weighted average based on the below between 05:00 to 08:00, comparable with above figures as 60 minutes pre-show; 30 minutes post-show]
* entire programme average - 526k (16.68%)
* 04:30 to 06:00 - 267k (25.39%)
* 06:00 to 07:40 - 874k (30.91%)
* 07:40 to 09:00 - 384k
* one of the above percentages is wrong, but not sure which - there's no way the average can be 16.68%
* peak: 1.02m (22.15%) at 07:35

BBC Highlights - 2.73m (22.9%)
* peak: 3.28m (26.15%) at 15:20
Total = 3.35 million
* combined peak: 4.30m
- lowest Australian Grand Prix figure since 2004 and possibly the second lowest since 1995

Note: Sky Sports F1 figures for Qualifying have not been reported.

Sky and Jake can spin it all they want, but in reality, the figures are abysmal.

[Ross Stonefeld]

So from an average of about 4.5 over the years to 3.3?

[D.M.N.]

So from an average of about 4.5 over the years to 3.3?

Yes, here's a full rundown:

1999 - 4.95m
2000 - 4.02m
2001 - 4.42m
2002 - 3.41m
2003 - 4.02m
2004 - 2.96m
2005 - 3.67m
2006 - 3.67m
2007 - 3.50m
2008 - 4.56m
2009 - 4.98m
2010 - 4.51m
2011 - 4.15m
2012 - 3.35m

The figures are more disappointing because we had a British 1-3, I could have seen a low figure if Vettel ran away.

[OwenC93]

That's not as bad as I thought but it's still moving money away from sponsors (Teams, tracks, organisers etc) and into Bernie's pocket .

[Dalton007]

3.35 million isn't bad, two million more than SKY. Let's remember, majority of the sky audience probably didn't tune into the BBC. It's the first race of the season, let's see what happens at the next race.

[kar]

I think this is to be expected with the move to Sky.

It will be very interesting to see the numbers for the British Grand Prix though.

[Wiggy]

What sort of ratings does the average football game on sky sports get? Will Sky really be happy with those figures?

Maybe more people will buy sky after realising what the BBC have done, still even last week people I spoke to who don't follow all the Internet talk etc, thought that the race would be live on the BBC.