|
|
Post by jamesdoyle on Mar 19, 2023 21:42:49 GMT
GWBWI Con +53 SNP +9 Lab -5 LDm -5 Grn -18 ASV Con +1.1 SNP +0.1 LDm +0.1 Lab +0.0 Grn -0.5 Cons positive - and on top of the rankings. Good result in Stirling, and the good vote increase in SCambs even meant they scored better (+13) than the LibDems who won the seat (+10) Not a good week for the Greens, Stirling being a little worse than SCambs. LDms matched Lab for GWBWI score (-5), but matched SNP for ASV (+0.1). Partly down to rounding, partly because they held the SCambs seat. I absolutely accept that it was the Tories' week, but -5 is a bit harsh on Labour, whose vote rose in Dunblane/Bridge of Allan, and who didn't do badly at all in Cottenham either. Perhaps you are basing your expectations on opinion polls, but it would have been unreasonable to expect Labour to be anywhere near winning either of these particular seats although it would perhaps have been different in the mid to late 90s. I'm not basing my 'expectations' on anything, because I don't have 'expectations'. I developed an algorithm which takes a number of factors into account and scores the results based on that. I don't put my thumb on the scale. You might disagree with the factors I use, or the results they give, but then you are free to work out something similar, or different, and make your case for that. All you are saying is that your opinion is that Labour did all right. Which is precisely what everyone does through the prism of their own expectations: mark up some results and mark down others. My system doesn't. The system has been in place for nearly a year now and to assume that I designed it a year ago to downscore Labour this week is rather ludicrous. For what it's worth, I'm disappointed that my first and second preferences, Grn and LDm, didn't do all that well. But I still report it, and I didn't nudge the figures to suit my preferences. |
|
|
|
Post by jamesdoyle on Mar 20, 2023 9:51:55 GMT
I absolutely accept that it was the Tories' week, but -5 is a bit harsh on Labour, whose vote rose in Dunblane/Bridge of Allan, and who didn't do badly at all in Cottenham either. Perhaps you are basing your expectations on opinion polls, but it would have been unreasonable to expect Labour to be anywhere near winning either of these particular seats although it would perhaps have been different in the mid to late 90s. I do think that the primary benefit of James’ system is showing that creating such a metric is basically impossible, but to be fair (well … sort of) the fault here may be his incomprehensible interpretation of STV. Nicely phrased - almost sounds like a compliment! I accept your comments about STV by elections. In my defence, I felt there were so few of them (9 out of 189 so far this year, I think) that it wasn't worth the effort of working out a fix - but I think I'll put the work in during the fallow period over the next few weeks. |
|
maxque
Non-Aligned
Posts: 8,984 
Member is Online
|
Post by maxque on Mar 20, 2023 12:44:22 GMT
I do think that the primary benefit of James’ system is showing that creating such a metric is basically impossible, but to be fair (well … sort of) the fault here may be his incomprehensible interpretation of STV. Nicely phrased - almost sounds like a compliment! I accept your comments about STV by elections. In my defence, I felt there were so few of them (9 out of 189 so far this year, I think) that it wasn't worth the effort of working out a fix - but I think I'll put the work in during the fallow period over the next few weeks. The fix is easy. As each elector only has 1 vote, you just add the candidates together. If there is 4000 valid votes and 1500 gave their 1st preference to SNP candidate Smith and 500 and their first preference to SNP candidate Collins, SNP has 50% of valid first preferences (2000). |
|
|
|
Post by jamesdoyle on Mar 20, 2023 13:40:39 GMT
Nicely phrased - almost sounds like a compliment! I accept your comments about STV by elections. In my defence, I felt there were so few of them (9 out of 189 so far this year, I think) that it wasn't worth the effort of working out a fix - but I think I'll put the work in during the fallow period over the next few weeks. The fix is easy. As each elector only has 1 vote, you just add the candidates together. If there is 4000 valid votes and 1500 gave their 1st preference to SNP candidate Smith and 500 and their first preference to SNP candidate Collins, SNP has 50% of valid first preferences (2000). You haven't seen the spreadsheet! Essentially there is a very long formula which branches off two options - all out elections or single seat elections - and now I've got add a third, which will then add treatment for every sub-ordinate condition. It's going to take a while. |
|
