RogueXV wrote: Tickingover, thanks for throwing this up here and more importantly explaining some of what it represents. One question though, am I not reading the chart correctly or is it missing Friday and maybe even Thursday games? You mentioned there was no significant difference between Friday & Saturday games and when I went to see the comparison I didn't see a Weekday[T.Fri]. Are you a Saturday game day fan trying to hide something?

The Days are in comparison to Friday as that is the first, alphabetically, category in the list of days. It's the same for all the categorical variables....they are compared to another(first variable) in the list. so the teams are compared to Aironi, the Home and LastHomeResult to Draws(D), the time of day to Afternoons. So Mondays give us a 59.57% increase on the seasonal average in comparison to Fridays.....Saturdays 4.8% increase on Fridays but this isn't statistically significant as P>| t | value is quite large. There were no Thursday league games, that I could find. As I said there are probably quite a few errors in this as it was thrown together pretty quick.

interesting to see the dip for the 2010/2011 season for most teams, was that the bad winter weather we had affecting attendances?

tickingover wrote: See attached regression analysis of all Connacht Home Games in the league for the last 11 years, since the league went to single table format.

Any correlation between attendances and whether game was televised?

with the day thats in it i've upgraded this with 2013/2014 season and this season so far. Average attendances for pro12 games only. Postseason data not included.
Of note Munsters downward trend has been halted - presumably because musgrave park is not being used this year.

Tickingover, thanks for throwing this up here and more importantly explaining some of what it represents. One question though, am I not reading the chart correctly or is it missing Friday and maybe even Thursday games? You mentioned there was no significant difference between Friday & Saturday games and when I went to see the comparison I didn't see a Weekday[T.Fri]. Are you a Saturday game day fan trying to hide something?

sea_point wrote:

ummm, wrote: Incredible amount of stats-porn lovers on this site.

And you were drawn to this thread....why..? :whistle:

Now now Seapoint, Data is not the plural of Anecdote!

ummm, wrote: Incredible amount of stats-porn lovers on this site.

And you were drawn to this thread....why..? :whistle:

Incredible amount of stats-porn lovers on this site.

so basically in summary... Connacht fans like games late on Monday (bank holidays I reckon) against munster/leinster where we've either drawn the last home game or won a few home games on the trot. We're not that bothered by Ulster, hate the rain and don't mind the cold (TX,TG,TN are temperature max, medium, min)

salmson wrote:

tickingover wrote: I have the code to scrape the fixtures and the end of season tables written if you want it, would save doing it manually.

The problem with figuring out attendance deviation as the team drops out of contention would be that you'd have to figure out where the team sits after every match, which would mean you would have to figure out the points achieved for every team for every game and then recalculate the table.....not impossible but probably bloody difficult......

Not half geeky enough I'm afraid this was a very amateur model. .

Pretty impressive! Am I reading it right in saying that playing Munster, Leinster, or late in the evening are string positive factors on attendance, while rainfall is a strong (significant) negative factor?

Couple of questions:

Could you explain a bit more the last home result - is it a case of a last result loss being worse (less bad) than a last result win rather than the win being good (positive effect) - or am I misreading it just because the determination and departure values are both negative?

Also how did you break up time of match. Obviously 3pm starts are afternoon and 7.30 is evening, but the current untelevised start time of 5pm and the old Friday 6.30 kickoff are a bit of both.

Trying to figure out how to respond to bits of this but basically yes that's right about Leinster/Munster, time of day and rain.

The last homeresult is based on a comparision against the unlisted result being a Draw (D) so a loss drops our attendance (deviation from the average for the season) at the next game by 36% in comparison to a draw but a win drops our attendance by 33% in comparison to a draw. (not that many draws so may not be as significant)

timeofday is rough enough and subjective but basically based on the hour of the start time so 13.00 - 15.00 is afternoon, 16.00 - 18.00 is evening and 19.00-21.00 is night. but interestingly a linear increase in hour causes a negative impact...but not that significant.

again not that strong on the details on all this, a real statistician would piss themselves , I just like getting the data and letting the machine do the hard work.

One chunk of the Clan is in Katie Dalys now, trying to pursuade Seamus to stand his round. Another chunk is pouring over a serious regression analysis of match day attendances.

I wonder do IRFU/Pro 12 look at any of this type of hard data before they make their decisions?

If the 3 game and 7 game packages currently on sale count as "season tickets" then attendance figures will be bumped up for rest of season whether or not people show up

tickingover wrote: I have the code to scrape the fixtures and the end of season tables written if you want it, would save doing it manually.

The problem with figuring out attendance deviation as the team drops out of contention would be that you'd have to figure out where the team sits after every match, which would mean you would have to figure out the points achieved for every team for every game and then recalculate the table.....not impossible but probably bloody difficult......

Not half geeky enough I'm afraid this was a very amateur model. .

Pretty impressive! Am I reading it right in saying that playing Munster, Leinster, or late in the evening are string positive factors on attendance, while rainfall is a strong (significant) negative factor?

Couple of questions:

Could you explain a bit more the last home result - is it a case of a last result loss being worse (less bad) than a last result win rather than the win being good (positive effect) - or am I misreading it just because the determination and departure values are both negative?

Also how did you break up time of match. Obviously 3pm starts are afternoon and 7.30 is evening, but the current untelevised start time of 5pm and the old Friday 6.30 kickoff are a bit of both.

NO idea what Tickingover's analysis indicates, it's at a level beyond my nerdishness (congrats... )...
Mine has a pretty Graph (with Colours and squiggly lines)...... :whistle:

I have no idea what any of this stuff means but thanks for sharing!

I have the code to scrape the fixtures and the end of season tables written if you want it, would save doing it manually.

The problem with figuring out attendance deviation as the team drops out of contention would be that you'd have to figure out where the team sits after every match, which would mean you would have to figure out the points achieved for every team for every game and then recalculate the table.....not impossible but probably bloody difficult......

Not half geeky enough I'm afraid this was a very amateur model. .