Scarlet Fever Llanelli Rugby Sport Wales Tickets Homepage
Forum Home Forum Home > RUGBY > SCARLETS GENERAL
  New Posts New Posts
  FAQ FAQ  Forum Search   Register Register  Login Login


Some Statistical Analysis on Attendance

 Post Reply Post Reply Page  <12
Author
Message
Rich (Bris) View Drop Down
Veteran
Veteran


Joined: 23 August 2007
Status: Offline
Points: 1338
Post Options Post Options   Thanks (0) Thanks(0)   Quote Rich (Bris) Quote  Post ReplyReply Direct Link To This Post Posted: 17 April 2018 at 12:32pm
Thanks Rob
Yep just looking now on PC and much better (noggin me - although I turned the phone horizontally, I didn't scroll to the right enough to see all the interpretation). Good work Joni

Edited by Rich (Bris) - 17 April 2018 at 12:37pm
Back to Top
Sponsored Links


Back to Top
joni_bach View Drop Down
Groupie
Groupie
Avatar

Joined: 27 August 2015
Location: Llanelli
Status: Offline
Points: 237
Post Options Post Options   Thanks (0) Thanks(0)   Quote joni_bach Quote  Post ReplyReply Direct Link To This Post Posted: 17 April 2018 at 12:32pm
Originally posted by John John wrote:

Why are there fewer observations since regionalisation than since the move to PyS given that the first is longer ago than the second?

So the move to PYS increased attendances inspite of it occuring at a time of relatively ordinary performances? Saturday is the best day of the weekend and in view of next year's shift to only pay tv, it seems that there will be no increase in attendance. Its good to be able to quantify some of these variables although some, like the number of home wins and the number of internationals are interconnected.

Thank you for your hard work....



I've realised why there was missing data for the stats since regionalisation. The software dismissed the observations where I didn't have information about free-to-air TV. Here is the revised data for 2003/04 - present. Please note that when you look at the data over a longer time-period, it is more difficult to establish relationships. We're comparing different home grounds for a start. For example, this data shows that winning the domestic championship has no significant effect on next season's attendances. This is true for when we won in 2003/04 but not true for 2016/17. You can read into this what you want (i.e. fans more loyal in 2003/04, smaller ground capacity) but on average the "2003/04 effect" cancels out the "2016/17 effect". This is why if you want relevant statistics, it's probably best to look at the PYS data rather than the data since 2003/04.

Dependent Variable: ATTENDANCE
Method: Least Squares
Date: 04/17/18   Time: 12:06
Sample: 1 198
Included observations: 197
Variable Coefficient Std. Error t-Statistic Prob.   Interpretation
C 5363.654 534.3685 10.03737 0 5363 fans attend each game regardless of other variables
CHAMPIONS 411.3662 341.2282 1.205546 0.2295 (not significant) Being champions last season has no significant affect on attendance
KOBH 1630.371 705.3552 2.311418 0.0219 Kick off on a bank holiday increases attendance by 1630 (compared to Saturday night kick off)
KOFRI -338.9818 332.8834 -1.01832 0.3098 (not significant) Kick off on a Friday night has no significant difference to attendance
KOSATPM -363.15 332.7979 -1.0912 0.2766 (not significant) Kick off on a Saturday night has no significant difference (compared to Saturday night kick off)
KOSUN -1289.581 430.2235 -2.99747 0.0031 Kick off on a Sunday decreases attendance by 1290 (compared to Saturday night kick off)
MATCHEURO 1191.333 299.0743 3.983402 0.0001 1191 extra fans attend a game if it is a European game
MATCHOSP 5554.988 626.178 8.871261 0 5555 extra fans attend a game if it is against the Ospreys
MATCHPLAYOFF 3905.523 1199.374 3.256302 0.0013 3905 extra fans attend a game if it is a play-off
WALESINT 195.7043 51.51144 3.79924 0.0002 195 extra fans attend a game for every Welsh international playing
WINPC 12.29727 4.758578 2.584231 0.0105 For every 1% increase in home wins in the current season, 12 extra fans attend
R-squared 0.574593     Mean dependent var 7698.746
Adjusted R-squared 0.551722     S.D. dependent var 2370.594
S.E. of regression 1587.198     Akaike info criterion 17.63155
Sum squared resid 4.69E+08     Schwarz criterion 17.81487
Log likelihood -1725.707     Hannan-Quinn criter. 17.70576
F-statistic 25.12284     Durbin-Watson stat 1.660806
Prob(F-statistic) 0
Back to Top
joni_bach View Drop Down
Groupie
Groupie
Avatar

Joined: 27 August 2015
Location: Llanelli
Status: Offline
Points: 237
Post Options Post Options   Thanks (0) Thanks(0)   Quote joni_bach Quote  Post ReplyReply Direct Link To This Post Posted: 17 April 2018 at 12:37pm
Originally posted by Rob o'r Bont Rob o'r Bont wrote:

Not sure what the Bank Holiday stats are telling us.  These would be  the local derbies and so we would expect attendances to be up in any case.  Or does the software take account of this?

You're right, you can't read into it and say "attendances are higher because it's a bank holiday", it's more likely that it's because they are local derbies against the Blues or Dragons. When I have time, I'm going to have a look to see if I can control for them too. It's just tricky because the more variables you include, the less robust the analysis is but I'll work on it. 
Back to Top
John View Drop Down
Veteran
Veteran


Joined: 15 August 2004
Status: Offline
Points: 4131
Post Options Post Options   Thanks (0) Thanks(0)   Quote John Quote  Post ReplyReply Direct Link To This Post Posted: 17 April 2018 at 12:39pm
Thanks for that. So now there is no significant difference in friday night attendances compared to saturday. And all the posts about that on here. All the moaning and groaning- much of which I did! But there is alot of interesting data here. Thank you for it.
Back to Top
John View Drop Down
Veteran
Veteran


Joined: 15 August 2004
Status: Offline
Points: 4131
Post Options Post Options   Thanks (0) Thanks(0)   Quote John Quote  Post ReplyReply Direct Link To This Post Posted: 17 April 2018 at 12:41pm
Originally posted by joni_bach joni_bach wrote:

Originally posted by Rob o'r Bont Rob o'r Bont wrote:

Not sure what the Bank Holiday stats are telling us.  These would be  the local derbies and so we would expect attendances to be up in any case.  Or does the software take account of this?

You're right, you can't read into it and say "attendances are higher because it's a bank holiday", it's more likely that it's because they are local derbies against the Blues or Dragons. When I have time, I'm going to have a look to see if I can control for them too. It's just tricky because the more variables you include, the less robust the analysis is but I'll work on it. 


But the fewer variables you take account of, the less robust the analysis is also!! Big smileWink
Back to Top
joni_bach View Drop Down
Groupie
Groupie
Avatar

Joined: 27 August 2015
Location: Llanelli
Status: Offline
Points: 237
Post Options Post Options   Thanks (0) Thanks(0)   Quote joni_bach Quote  Post ReplyReply Direct Link To This Post Posted: 17 April 2018 at 1:54pm
Originally posted by John John wrote:

Thanks for that. So now there is no significant difference in friday night attendances compared to saturday. And all the posts about that on here. All the moaning and groaning- much of which I did! But there is alot of interesting data here. Thank you for it.

Well, if you look at the more relevant data from seasons at PYS, Friday night kick offs lose about 500-odd fans and Sunday kick-offs lose close to 1000 but it's the time on a Saturday (e.g. 3pm v 7.35) that makes no difference. Still, I think we're still allowed to have a moan about what time on a Saturday we'd prefer.
Back to Top
 Post Reply Post Reply Page  <12
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 11.04
Copyright ©2001-2015 Web Wiz Ltd.

This page was generated in 0.031 seconds.