The old David's Models

The margins reflected here are the median out of 250,000 simulations, and probabilities are calculated using margins.

	County	Rating	Median Margin	Dem % chance of winning	GOP % chance of winning
0	Appling	Safe R	-50.53	0.00	100.00
1	Atkinson	Safe R	-44.62	0.00	100.00
2	Bacon	Safe R	-62.76	0.00	100.00
3	Baker	Safe R	-9.75	1.25	98.75
4	Baldwin	Lean D	4.20	71.00	29.00
5	Banks	Safe R	-64.53	0.00	100.00
6	Barrow	Safe R	-31.10	0.00	100.00
7	Bartow	Safe R	-39.79	0.00	100.00
8	Ben Hill	Safe R	-22.05	0.00	100.00
9	Berrien	Safe R	-56.91	0.00	100.00
10	Bibb	Safe D	27.11	100.00	0.00
11	Bleckley	Safe R	-44.89	0.00	100.00
12	Brantley	Safe R	-69.98	0.00	100.00
13	Brooks	Safe R	-20.54	0.00	100.00
14	Bryan	Safe R	-24.70	0.00	100.00
15	Bulloch	Safe R	-20.50	0.00	100.00
16	Burke	Lean R	-2.37	38.15	61.85
17	Butts	Safe R	-34.30	0.00	100.00
18	Calhoun	Safe D	17.44	100.00	0.00
19	Camden	Safe R	-24.88	0.00	100.00
20	Candler	Safe R	-36.17	0.00	100.00
21	Carroll	Safe R	-30.99	0.00	100.00
22	Catoosa	Safe R	-43.79	0.00	100.00
23	Charlton	Safe R	-43.09	0.00	100.00
24	Chatham	Safe D	24.70	100.00	0.00
25	Chattahoochee	Lean R	-3.22	33.90	66.10
26	Chattooga	Safe R	-48.97	0.00	100.00
27	Cherokee	Safe R	-27.36	0.00	100.00
28	Clarke	Safe D	46.03	100.00	0.00
29	Clay	Likely D	7.53	87.65	12.35
30	Clayton	Safe D	73.69	100.00	0.00
31	Clinch	Safe R	-47.83	0.00	100.00
32	Cobb	Safe D	22.03	100.00	0.00
33	Coffee	Safe R	-35.50	0.00	100.00
34	Colquitt	Safe R	-41.49	0.00	100.00
35	Columbia	Safe R	-17.61	0.00	100.00
36	Cook	Safe R	-35.31	0.00	100.00
37	Coweta	Safe R	-24.67	0.00	100.00
38	Crawford	Safe R	-36.61	0.00	100.00
39	Crisp	Safe R	-22.91	0.00	100.00
40	Dade	Safe R	-53.77	0.00	100.00
41	Dawson	Safe R	-51.85	0.00	100.00
42	Decatur	Safe R	-16.99	0.00	100.00
43	DeKalb	Safe D	69.18	100.00	0.00
44	Dodge	Safe R	-39.88	0.00	100.00
45	Dooly	Likely R	-5.06	24.70	75.30
46	Dougherty	Safe D	42.15	100.00	0.00
47	Douglas	Safe D	35.00	100.00	0.00
48	Early	Likely R	-8.37	8.15	91.85
49	Echols	Safe R	-66.71	0.00	100.00
50	Effingham	Safe R	-36.12	0.00	100.00
51	Elbert	Safe R	-34.17	0.00	100.00
52	Emanuel	Safe R	-33.76	0.00	100.00
53	Evans	Safe R	-31.67	0.00	100.00
54	Fannin	Safe R	-51.26	0.00	100.00
55	Fayette	Lean D	4.72	73.60	26.40
56	Floyd	Safe R	-30.97	0.00	100.00
57	Forsyth	Safe R	-21.84	0.00	100.00
58	Franklin	Safe R	-59.70	0.00	100.00
59	Fulton	Safe D	50.84	100.00	0.00
60	Gilmer	Safe R	-50.69	0.00	100.00
61	Glascock	Safe R	-69.39	0.00	100.00
62	Glynn	Safe R	-18.84	0.00	100.00
63	Gordon	Safe R	-51.10	0.00	100.00
64	Grady	Safe R	-26.87	0.00	100.00
65	Greene	Safe R	-23.14	0.00	100.00
66	Gwinnett	Safe D	25.69	100.00	0.00
67	Habersham	Safe R	-51.23	0.00	100.00
68	Hall	Safe R	-33.27	0.00	100.00
69	Hancock	Safe D	40.38	100.00	0.00
70	Haralson	Safe R	-60.55	0.00	100.00
71	Harris	Safe R	-34.35	0.00	100.00
72	Hart	Safe R	-44.51	0.00	100.00
73	Heard	Safe R	-57.46	0.00	100.00
74	Henry	Safe D	33.45	100.00	0.00
75	Houston	Lean R	-4.03	29.85	70.15
76	Irwin	Safe R	-44.72	0.00	100.00
77	Jackson	Safe R	-44.82	0.00	100.00
78	Jasper	Safe R	-43.92	0.00	100.00
79	Jeff Davis	Safe R	-55.44	0.00	100.00
80	Jefferson	Likely D	5.89	79.45	20.55
81	Jenkins	Safe R	-25.92	0.00	100.00
82	Johnson	Safe R	-38.99	0.00	100.00
83	Jones	Safe R	-25.03	0.00	100.00
84	Lamar	Safe R	-33.65	0.00	100.00
85	Lanier	Safe R	-35.25	0.00	100.00
86	Laurens	Safe R	-23.29	0.00	100.00
87	Lee	Safe R	-32.87	0.00	100.00
88	Liberty	Safe D	28.87	100.00	0.00
89	Lincoln	Safe R	-36.32	0.00	100.00
90	Long	Safe R	-20.44	0.00	100.00
91	Lowndes	Safe R	-11.33	0.00	100.00
92	Lumpkin	Safe R	-45.94	0.00	100.00
93	Macon	Safe D	23.88	100.00	0.00
94	Madison	Safe R	-41.61	0.00	100.00
95	Marion	Safe R	-19.76	0.00	100.00
96	McDuffie	Safe R	-16.86	0.00	100.00
97	McIntosh	Safe R	-17.63	0.00	100.00
98	Meriwether	Safe R	-16.02	0.00	100.00
99	Miller	Safe R	-42.52	0.00	100.00
100	Mitchell	Safe R	-9.27	3.65	96.35
101	Monroe	Safe R	-34.05	0.00	100.00
102	Montgomery	Safe R	-43.53	0.00	100.00
103	Morgan	Safe R	-34.79	0.00	100.00
104	Murray	Safe R	-61.24	0.00	100.00
105	Muscogee	Safe D	28.79	100.00	0.00
106	Newton	Safe D	20.38	100.00	0.00
107	Oconee	Safe R	-24.95	0.00	100.00
108	Oglethorpe	Safe R	-30.11	0.00	100.00
109	Paulding	Safe R	-15.13	0.00	100.00
110	Peach	Tossup	-0.47	47.65	52.35
111	Pickens	Safe R	-51.37	0.00	100.00
112	Pierce	Safe R	-64.98	0.00	100.00
113	Pike	Safe R	-59.27	0.00	100.00
114	Polk	Safe R	-46.34	0.00	100.00
115	Pulaski	Safe R	-30.45	0.00	100.00
116	Putnam	Safe R	-33.13	0.00	100.00
117	Quitman	Safe R	-10.17	0.00	100.00
118	Rabun	Safe R	-42.64	0.00	100.00
119	Randolph	Safe D	11.70	100.00	0.00
120	Richmond	Safe D	39.66	100.00	0.00
121	Rockdale	Safe D	50.60	100.00	0.00
122	Schley	Safe R	-48.05	0.00	100.00
123	Screven	Safe R	-16.55	0.00	100.00
124	Seminole	Safe R	-35.86	0.00	100.00
125	Spalding	Safe R	-10.22	0.00	100.00
126	Stephens	Safe R	-49.36	0.00	100.00
127	Stewart	Safe D	20.54	100.00	0.00
128	Sumter	Likely D	7.49	87.45	12.55
129	Talbot	Safe D	22.26	100.00	0.00
130	Taliaferro	Safe D	23.94	100.00	0.00
131	Tattnall	Safe R	-41.39	0.00	100.00
132	Taylor	Safe R	-20.64	0.00	100.00
133	Telfair	Safe R	-30.53	0.00	100.00
134	Terrell	Safe D	9.83	99.15	0.85
135	Thomas	Safe R	-16.58	0.00	100.00
136	Tift	Safe R	-30.10	0.00	100.00
137	Toombs	Safe R	-40.34	0.00	100.00
138	Towns	Safe R	-48.21	0.00	100.00
139	Treutlen	Safe R	-33.29	0.00	100.00
140	Troup	Safe R	-15.03	0.00	100.00
141	Turner	Safe R	-17.41	0.00	100.00
142	Twiggs	Lean R	-3.51	32.45	67.55
143	Union	Safe R	-49.17	0.00	100.00
144	Upson	Safe R	-26.83	0.00	100.00
145	Walker	Safe R	-47.94	0.00	100.00
146	Walton	Safe R	-36.59	0.00	100.00
147	Ware	Safe R	-34.19	0.00	100.00
148	Warren	Likely D	8.70	93.50	6.50
149	Washington	Lean D	3.23	66.15	33.85
150	Wayne	Safe R	-47.90	0.00	100.00
151	Webster	Safe R	-14.46	0.00	100.00
152	Wheeler	Safe R	-31.93	0.00	100.00
153	White	Safe R	-53.33	0.00	100.00
154	Whitfield	Safe R	-35.90	0.00	100.00
155	Wilcox	Safe R	-41.26	0.00	100.00
156	Wilkes	Safe R	-10.34	0.00	100.00
157	Wilkinson	Likely R	-7.56	12.20	87.80
158	Worth	Safe R	-39.69	0.00	100.00

POLL TABLE
Polls are weighted in the model using 538 rating, recency, and whether or not they're an internal.

	Pollster	Rating	Date	Internal	Party	Margin
0	Fabrizio, Lee & Associates/Impact Research	NaN	11/17/22	NaN	NaN	4.0
1	Frederick Polls	B/C	11/26/22	NaN	NaN	0.0
2	Phillips Academy	NaN	11/27/22	NaN	NaN	-1.0
3	SurveyUSA	A	11/30/22	NaN	NaN	3.0
4	Emerson College	A-	11/30/22	NaN	NaN	2.0
5	SSRS (CNN)	C	11/29/22	NaN	NaN	4.0
6	SSRS (CNN)	C	11/29/22	NaN	NaN	7.0
7	Patriot Polling	NaN	12/02/22	NaN	NaN	1.7
8	University of Massachusetts Lowell	B+	11/28/22	NaN	NaN	5.0
9	InsiderAdvantage	B	12/04/22	True	REP	3.0
10	Data for Progress	B	12/05/22	NaN	NaN	2.0
11	Trafalgar Group	A-	12/05/22	True	REP	3.7
12	Landmark Communications	A	12/04/22	NaN	NaN	6.0
13	Mitchell Research & Communications	C-	12/04/22	NaN	NaN	5.0

RATING CATEGORIES

Safe Republican - A <=95% chance of voting Republican

Likely Republican - A <=75% chance of voting Republican

Leans Republican - A <=55% chance of voting Republican

Tossup - A >=55% chance of voting for either party

Leans Democratic - A <=55% chance of voting Democratic

Likely Democratic - A <=75% chance of voting Democratic

Safe Democratic - A <=95% chance of voting Democratic

METHODOLOGY

This model is made up of two components, two which everyone that's a politics junkie knows well - the fundamentals and the polls.

The model fundamentals are made up of three components in of themselves - a presidential composite using

5% 2012 presidential election results
15% 2016 presidential election results
and 80% 2020 presidential election results.

On top of that, the presidential composite is adjusted for state trends over the last 2 years, which altogether makes a partisan lean index. However, a partisan lean index is on its own not a good way to forecast elections, so that's why I've also used;

a fundraising index using the dollar spending difference between the Democratic and Republican candidates. SPECIAL THANKS TO ROB PYERS/ADIMPACT!
a candidate index (using 2020 results for current incumbents and candidate experience (i.e. whether a candidate is a mayor, state representative, etc) for non-incumbents)
ratings from the Cook Political Report with Amy Walter , Inside Elections from Nathan Gonzales , Sabato's Crystal Ball from the UVA CFP , and Split Ticket .
and a demographic index based off AP votecast (although demographics has very little weight.)

As for the polls, it's a little more simple; polling data simply uses a weighted average of every poll based off of 538 rating, recency and whether the poll is an internal or not. Poll margins are also edited based off if they're a Democratic or Republican internal polls, since internal polls are usually 5 points more biased toward the surveying party.
After the fundamentals and poll margins are calculated, they're weighted together using the amount of polls that have been produced, which produces an exact margin. Once an exact margin is created, a randomizer of +2 to -2 points is ran 250,000 times, and the median margin is used as the neutral year forecast margin. The neutral year forecast margin is then added to the state level environment (which is calculated from the national environment), which produces the official acctuallydavid.com forecast margin for the state.

Special note because the Georgia US Senate runoff model has a county model attached to it - the county model compiles a partisan lean index for each county, and adds the statewide margin to the partisan lean index, and just adjusts for turnout to make the county model accurate to the statewide margin.