What is Trump's actual national poll #s?

[weixiaobao]

I did came to the conclusion that most of the internet national polls are wrong.  By leaving out the particular 5 polls (Zogby, Reuters, Morning Consult, Gravis, and YouGov).  The #s became more collaborative with state polls.  Trump max ceiling after his Alabama rally was about 30%.  He keep losing and gaining about 4 to 5%, after and in between debates (respectively).  Because the #s of average without those 5 mentioned polls and the state polls are extremely similar.  I don't think those mentioned polls are correct.  When Trump reached his high, those polls sometimes are 5 to 10% higher than my average.  Same with when Trump hit his low points. 
 
---------------------------------------------------------------------------------
From the period of mid July to Aug 6 (when Trump start to lead on the polls to pre 1st debate)

The top 3 were

Trump: a) 24% ----------b) 26% c) 24% d) 25% e) 24%
Bush:---a) 12% ----------b) 14% c) 13% d) 15% e) 14%
Walker: a) 12% ----------b) 10% c) 10% d) 10% e) 11%

--> A is straight up average of national polls between that time period minus the 5 polls I mentioned
 
The it is compare to the various state polls.  What I do is that if a state has multiple polls, I add average them (though proportionally by sample size).  I do include Gravis, Morning Consult, Zogby, and YouGov state polls to reach some sort of consensus for the state.

b) is just the average of different state polls (though multiple polls of one state count as 1) from the same period of time.
c) is the average of different state polls but I branch out a bit to other polls at close to that period of time that I think the major events didn't change the #s too much.
d) is proportional via the district delegates that each states give.  A flaw system but I want to take into account that each state should contribute the same # of voters in term of an imaginary national primary.
e) proportional via district delegates but extend to other  polls at close to that period of time that I think the major events didn't change the #s too much.

-----------------------------------------------------------------------------

From the fall out of the 1st debate to before Trump's rally in Alabama when the news dominated by Trump's official immigration plan. (8/07-8/20)

The top candidates at the time were Trump, Carson, and Bush.  Then Rubio, Cruz, Fiorina, and Walker.  I only did the top 3 candidates and Fiorina/Walker (because those 2 had drastic change.

1) Trump: a) 24%---------b) 22% d) 23%
2) Carson: a) 10%---------b) 11% d) 11%
3) Bush:    a) 10%---------b) 11% d) 11%
4) Fiorina: a) 7%-----------b) 8%   d) 7%
5) Walker: a) 7% ----------b) 7%   d) 6%

Used the same keys.  So the conclusion is that the first Trump's peak where he dominated prior to the first debate.  He had a ceiling of 25%.  After the first debate, he drop very slightly.  And once again, the national average that left out the 5 polls I mentioned are extremely close to the combine state polls using my very flaw method.

----------------------------------------------------------------------------------------
From the time period of Trump's rally in Alabama to prior to the second debate (8/21-9/16)

1) Trump: a) 30%-------------b) 29% d) 29%
2) Carson: a) 17%-------------b) 18% d) 18%
3) Bush:   a) 9%---------------b) 8%    d) 9%
4) Cruz:    a) 6%---------------b) 6%   d) 7%
5) Rubio:  a) 6%---------------b) 5%   d) 5%
6) Fiorina: a) 4%--------------b) 5%    d) 5%

Using the same key, once again, the average of state polls leaving out those 5 national internet polls (but including Survey monkey and internet state polls) collaborate with each others.  This is Trump at his highest which is at a 30% ceiling.  I believe Reuters, YouGov, Gravis, Morning Consult went somewhere like 36% to 37% to even 40%.
------------------------------------------------------------------------------
Fall out of the 2nd debate.  Last half of September post debate 2 (9/16 to end of september)

1) Trump: a) 23%------------ b) 23% d) 22%
2) Carson  a) 17%------------ b) 17% d) 17%
3) Fiorina  a) 11%-------------b) 11% d) 12%
4) Rubio    a) 9%--------------b) 10%  d) 10%
5) Bush     a) 8% --------------b) 7%   d) 8%
6) Cruz     a) 6%---------------b) 7%   d) 6%

Once again, the national average is exactly the same as the state polls.
Trump dipped quite a lot here, slightly less than his core base of around 25%.
 -------------------------------------------------------------------------------------------
Period of 10/1-10/21.  After this period, Carson either pulled evenly with Trump and or leading Trump in show polls

1) Trump a) 28%---------------b) 29% d) 28%
2) Carson a) 21%---------------b) 19% d) 19%
3) Rubio   a) 10%---------------b) 11% d) 12%
4) Cruz     a) 7%----------------b) 7%   d) 7%
5) Bush    a) 7%----------------b) 6%   d) 7%
5) Fiorina  a) 6%----------------b) 7%   d) 7%

Once again, the average of national polls (without those 5 mentioned polls) are similar to what state polls collectively portrayed.  The big different here seemed to be the different of 2 points for Carson.  Of course, when not enough states are polled.  The result can be screwed on which regions that poll the most that favored a candidate.  But overall, it still look very very similar.

-----------------------------------------------------------------------------
Pre Paris attack (10/21-11/13)- Carson reached front runner status

1) Trump a) 25.0%----------------- b) 23.3% c) 24.8% d) 23.7% e) 24.5%
2) Carson a) 24.5%------------------b) 23.4% c) 22.3% d) 21.6% e) 21.2%

At some point in this period, Carson probably pull ahead.  But the beginning and the end of this period, Trump probably ahead.  So average out, Trump was slightly ahead during this period.

The next period after this is due to Thanksgiving, I don't think I have enough data.  So we will wait until more polls came out in December.  And we will see if Trump is indeed break his previous high in September.