As we head to the finish line in #nbpoli there are all sorts of guesses and interpretations of the polling we are seeing, and predictions. FWIW, here's what I'm looking at to get a sense of what to watch on election night.

1. First of all -- don't buy the posturing of those who say "polls mean nothing". If you use them right, you'll know more. What we remember as polling misses are usually (a) late trends that continued after polls went dark or (b) people mistaking the CW they remember for polls

2. A big part is to avoid getting caught up in any one poll. Sure, right now Libs love EKOS and PCs are all debunking it. But smart prognosticators look at the commonalities and averages -- any one poll can miss something, but they rarely all miss. NB polling has commonalities.

3. First, remember where we started -- last election tells us what a dead heat in seats looks like. The Libs won the popular vote by 6 points.But lots of those were running up the score in the North, and in seats it was a tiny (but significant) 22-21 PC edge.

4. That also means that when you see a party surge in pop vote, you need a theory of where it came from. A surge in the rural North or rural South may mean little -- just turns 40 point wins to 20 point wins. But in the cities, a 5 point swing can matter. So, what do we know?

5. We have had 6 polls in the elxn period. Two came later, post-debate. Only one pollster (Mainstreet) gave us 2 polls to compare. We have that. We also can bet parties have polls we don't see--but we can sometimes guess their assumptions by their actions and choices.

6. The PCs: in 5/6 polls, they fall into the high 30s/low 40s. The 2 late polls are lowest --37% and 33%. The good news? All of these mean they're UP from 2018, and they narrowly won that. The bad news is there's some sign of slippage. But how much?

7. Some slippage is normal. "Rally-round-the-flag" leads (where a leader handles a crisis competently) tend to be fragile. An election returns a leader to equal status w/others, and voters remember their issue differences. War leaders --Bush, Churchill-- lost post-crisis.

8. The good news for Team Higgs is that the one repeat pollster -- Mainstreet -- was their least bullish early poll. Comparing their best poll (44%) to EKOS (33) looks like freefall. Mainstreet's 38 to 37 looks like a stable vote in a fairly tame election, and the PCs want that.

9. The Libs have stability, too. 4/6 polls have them at 32-33%. Two more show slippage to the high 20s. Even EKOS,shows the PCs falling back to them, but no Lib gains. Good news-their floor is solid, even if Vickers has been seen as weak. Bad news - no growth beyond their base.

10. It is significant that the later the poll, the better the Greens do. Low teens early, late polls are pushing 20%. David Coon has quietly stuck to issues, and it's quietly helping him. There's consensus among pollsters on that.

11. PANB polls are hardest to read. Some had them near collapse at 2%, below the "Thomason Line". EKOS has them rebounding to 10% with some gains in Fton and rural south. The PC/PANB split is the most volatile among the pollsters.

12. Other factors -- I think incumbency and local riding assns will matter. In a pandemic w/few flashy issues, those years of bingo sheets and GOTV info matter. Incumbents (or those inheriting strong local machines) likely ID'ed their vote better&got them to the advance polls

13. If you're a Green in Moncton Centre and you're seeing a surge -- do you know who those voters are & can you find them on E-Day? Because, say, a Liberal in Moncton Centre will know VERY well where the Liberals are. Could save a few incumbents.

14. Also-regional breakdowns are VERY volatile due to small sample size. But in this race, WHERE the parties gain matters a lot. And again, looking at several polls and finding commonalities can help (but caution is really needed with regional numbers).

15. Here are some commonalities -- the PCs dominate Saint John's CMA. Moncton-last two polls both show modest but relevant PC gains. Fredericton is most volatile, pollsters are showing a late Green surge and Grit weakness, maybe a PANB recovery.

16. There's no sign of a PC recovery in the North, but the good news for them is that means the inefficient vote advantage is still there. The Greens HAVE gained as the northern alternative to the Libs, but that may not move many seats.

17. What's it all mean? Well, a lot of safe incumbents in the North and rural South. And a few key ridings could mean the difference between Premier Higgs getting a majority and a scenario where Mr. Vickers backs into a coalition with Green support.

18. (On that note, remember that some races matter more than others if traditional alliances hold. A PC taking Fton-York from PANB, or a Lib comeback in Tantramar doesn't change the balance. Libs winning Miramichi or PCs winning Fton North...matters a LOT).

19. The most vital ridings list has shifted a bit, as I read the polls. I'd upgrade Moncton East & South as vital for the PCs. If they have gained in that area as the last 2 polls suggest (and the gain isn't coming from Dieppe seats) then they need Mssrs Turner and Allain to win.

20. The next most likely seats to take the PCs from 21 (their 2018 baseline w/Shippagan reassigned) are SJ Harbour (usually tight but BIG regional surge) and Carleton-Victoria (demographically Tory but an excellent Lib organization) and Fton North (Lib collapse but a Green rise)

21. There are 2 PANB seats they could convert as well (as stated above flipping Lib seats is more vital for keeping power, but they'd like a majority w/o the purple chaperone, I'm sure). Fton-York should be theirs, and you saw Austin name check Minto a LOT. That would be 28.

22. The big Liberal problem -- down from their 2018 numbers, what seat can they flip? Miramichi is a must, and Vickers is likely to win but nervous. He shouted out a LOT in the last debate, too, and seeing his manager attack Conroy's EA on Twitter suggests their polling is close.

23. They have a great candidate in Oromocto-Lincoln, but their Fredericton numbers and organization is collapsing. When you run a candidate you banned last election in Hanwell, where Cindy Miles almost won, you know you're slipping in Fton. Greens are taking votes & talent away.

24. Even if EKOS is right, the Lib scenario is (a) hold what you got, somehow, (b) flip the Chi (c) catch a vote split surprise somewhere in an urban centre and (d) be nice to the Greens and hope red + Green = 25. And this is NOT impossible. Vickers can still be premier.

25. (At some point, spotting the Libs 15 Northern seats will hurt the Tories. It just may not be this point.)

26. Green incumbents look safe. The student vote drop won't bother Mitton - party vote is up & she's a good riding MLA. If Fton surge is real, Luke Randall will be in good shape (if early polls showing the PCs above 40 hold, then Jill Green is a good candidate).

27. The northern Green gains may not amount to much, given there's no vote split and many Libs have machines and 40 point leads to start. But Charles Theriault, like Kevin Arseneau in 2018, has built an organization & has a real identity. He may overperform and win.

28. The late movement in polls suggest that Kris Austin has put himself back in the race. But his schedule and rhetoric suggest Roy Wiggins has a shot. At this point, a good night for them would be holding 2 of 3 seats. Rick Desaulniers is going to need a vote split miracle.

29. Last point - there WILL be a surprise. Demographics and buzz were on the side of a few winners last time -- Andrea Anderson Mason, Michelle Conroy, and Kevin Arseneau -- but they all won handily and pundits weren't predicting that. Here are a few scenarios that are PLAUSIBLE:

30. Those late Green urban numbers have a high MoE, but there might be a 6th seat in there for them. I'd watch, in order, Moncton Centre, Fton York, Oromocto-Lincoln, Moncton East, Fton Hanwell, or SJ Harbour. No one of those is >50%, but that among them there could be a shocker.

31. Other sleepers -- there's PA buzz in Carleton, and the EKOS numbers are just enough to hint at a sleeper vote. The Green northern surge could lead to a PC northern seat w/a good candidate -- I'd watch the Northwest. There's a non-zero scenario where Austin loses, Conroy wins.

32. Well, that's it. Hope you enjoyed the rant. As always, we have a lot of good people running. They can't all win, so take time to express your gratitude to those who do. It's scary putting yourself out there and risking a public loss. That good people do it anyway matters.

33. Oh yeah....my predictions:

PC 26, Lib 16, Greens 6, PA 1