Turkey General election - November 1st 2015

[Zanas]

So, things pretty much unchanged, Others a bit down, AKP slightly up, HDP up a little more. The AKP going on Kurds as terrorists may win them a few moderates, but may also trigger a solidarity vote in the Southeast. HDP could well end up raising their vote share and losing a few seats in Istanbul, while not gaining as much in the Southeast.

[Zanas]

Two other polls are pretty recent and of some use (another one is a CHP internal, not very outlier-y, but outlier-ish) :

AKAM, released August 24th :

AKP 39.7    
CHP 28.5    
MHP 12.6    
HDP 14.8    
Others 4.4

They also give the AK Party at 250, the CHP at 144, HDP at 88 and MHP at 68.

Gezici, released August 18th :

AKP 39.2    
CHP 26.4    
MHP 16.2    
HDP 14.1    
Others 4.1

If we average those three recent polls, we get (with variations from the actual June result) :

AKP 40.2 (-0.7)
CHP 26.8 (+1.85)
MHP 14.8 (-1.46)
HDP 14.5 (+1.41)
Others 3.6 (-1.1)

Also interesting, the AKAM poll asked who was to blame for the coalition talks' failure, with multiple answers possible :

President Erdogan : 79% (!!)
AK Party: 65.3
CHP: 40.2
MHP: 45.3
All: 15.9

31% of MHP voters disagreed with their party's stance in the talks, compared to 4% of HDP's, 3% CHP and 14% AKP. Hence their (relative) drop in support I guess.

[Zanas]

An average of two recent September polls (one having a 15,000 sample !) and another 2-week old poll that seem in line with these gives the following figures :

AKP 43.5
CHP 27
MHP 14.3
HDP 11.7
Others 3.5

So the HDP actually loses support, and AKP regains a bit, but these figures are actually pretty close to what polling seemed like a week before the June election. And HDP could still end up with more seats than MHP with these.

[Zanas]

(Bloomberg) -- Support for Turkey’s governing AK Party drops in latest poll by Sonar Arastirma ahead of Nov. 1 repeat election, according to Gazeteport.
Results:
AKP 38.2% (vs 41% at June 7 election)
CHP 28%
MHP 18.2%
HDP 13.7%

Yeah, but another poll today (Andy-Ar) doesn't tell the same story :
AKP 42.9    
CHP 27.0    
MHP 15.9    
HDP 12.8    
Others 1.4

That's with undecideds proportionally redistributed, though, so to be taken with a grain of salt.

Looking a bit back, it seems Sonar consistently overestimates the MHP.

So far, none of the Big Four has made successful inroads. The war on the HDP launched by Erdogan these last days (a hundred of their offices nationwide have been attacked by AKP frindly thugs) seems to have no effect on the voting intentions.

[Zanas]

Gezici tensd to underpoll AKP, and Erdogan isn't amused.

Overall, the latest polls don't show any significant shift from the June results.

[Zanas]

A bombing just occurred during a left-wing rally in Ankara, causing 30+ deaths. HDP have suspended all their election rallies for now. Not sure how all this situation will play out. Could HDP call for a boycott ? Granted, the bombing is not directly the work of Erdogan or the AKP, but the hostile atmosphere that they have instilled these last months is in part responsible, at least.

[Zanas]

Well, not really. We've had polls ranging AKP from 38% to 44% throughout August, September and October, with no discernable trend. Same goes for CHP (25% to 28%, one outlier at 30%), for MHP (13% to 18%, mostly 14 or 15), and for HDP (11% to 14%, mostly 12-13), and again throughout most of August, September and October.

I think it is therefore safe to say that that one only poll showing AKP at 47% when they've never been higher than 44% since early August is a) an outlier, b) a manipulation. Oh, surprise ! A&G haven't polled the November race yet ! They're coming with a final prediction without a track record. From June, I can only find their seat-count prediction, not their voting intention one.

There are five final polls : ORC, Metropoll, Gezici, A&G and Konda. That's excluding a Kurd-Tek poll from Oct 21, which obviously overpolls HDP a bit at the expense of AKP. That's also excluding a mega AKAM poll from Oct 20 with a sample of 24,500 !!! So the arithmetic mean of these 5 final polls gives this :
AKP 43.7
CHP 26.5
MHP 14.3
HDP 12.8
Others 2.7

The Konda one seems a bit too favorable to the opposition, with AKP at 41.7, CHP at 27.9 and HDP at 13.8. The A&G one has the AKP at 47.2. If you exclude those two, then the three remaining ones tell a quite consistent story (with variation from June result in brackets) :
AKP 43.2 (+2.5)
CHP 26.5 (+1.4)
MHP 14.6 (-1.9)
HDP 12.6 (-0.4)
Others 3.2

I'd hardly call that a surge for AKP. Probably a bunch of tactical voters from MHP switching to AKP, plus a small boost in legitimist turnout for the party in power. The most recent seat prediction based on a poll resembling something like this is the Betimar poll from Oct 20, which had the AKP and HDP at one point higher than these numbers, and CHP and MHP one point below. This Betimar projection gave the AKP a 6-seat absolute majority, and the HDP (nearly tied with the MHP) 18 seats more than the MHP. So if we took our above 3-poll mean, we could be on a razor thin majority, or just below. Should be interesting.

Anyway, the dice are cast now, wait and see.

[Zanas]

CHP is pretty much at its polling level. MHP clearly took a blow, I'm guessing vastly in favor of AKP. A number of June HDP voters were scared away. The only poll at 47% in the end was indeed right in feeling the trend. HDP gets pretty much the worst result they could : below 10% they would have pleaded voter fraud, with a certain level of plausibility. Now they're left with nothing to really plead.

Terrible for Turkey and reason in general.