the document says 20, but it's actually 19.7, so it's not yet over my stated threshold.

if the next batch of sample comes in and the ndp is higher than 20%, then that's signal. if it flops around at 18-19%, then the signal is fading. and, if it falls down, it's just fluctuation around the mean - and likely explained by the long weekend, rather than the debate or a bradley effect.

the numbers are less weird, here, though. more to the point is that the number of undecideds is falling.

the actual numbers are:

liberals 28.0
conservatives 28.8
ndp 17.8
green 8.0
undecided 10.9

the undecideds were closer to 15% a few days ago. that's where the real movement is. the liberals and conservatives are each down a point or two, which is just static at this point; the real break seems to be from the undecided, which took a big leap to the ndp.

i guess they found singh's deep arguments at the debate to be compelling. "mr. delay and mr. deny". profound stuff. lol.

i'm going to give it one more day, but there's very close to almost being a signal. i'm baffled, but so be it.

i will say, though, that the movement from the undecideds is particularly suggestive of a bradley effect, but, you can't know that until it actually happens.

https://secureservercdn.net/198.71.233.47/823.910.myftpupload.com/wp-content/uploads/2019/10/2019-1445-CTV-Globe-ELXN-2019-10-12.pdf

the liberals are supposed to do better than this