1. State House 2022 and 2024, contested D-vs-R races on the current lines: 227 races
   battleground             n= 39  Democrats won 18 (46%)  mean |margin| 7.3   (run of 2026-10-18 as persuasion n=37 17 D 46% |m| 7.0 + turnout n=2 1 D |m| 12.5; classes merged 2026-10-19)
   lean_D                   n= 23  favored side won 22 (96%)  mean |margin| 12.7
   lean_R                   n= 32  favored side won 28 (88%)  mean |margin| 13.5
   safe_D                   n= 78  favored side won 78 (100%)  mean |margin| 33.6
   safe_R                   n= 55  favored side won 55 (100%)  mean |margin| 29.5
   by |Partisan EAI| band (share won by the side the EAI favors):
     [0, 3)     n= 19  favored won 79%   State House margin vs Partisan EAI: r 0.60, vs raw vote: r 0.67
     [3, 5)     n= 10  favored won 60%   State House margin vs Partisan EAI: r 0.85, vs raw vote: r 0.77
     [5, 7)     n= 10  favored won 80%   State House margin vs Partisan EAI: r 0.61, vs raw vote: r 0.62
     [7, 10)    n= 31  favored won 84%   State House margin vs Partisan EAI: r 0.79, vs raw vote: r 0.80
     [10, 15)   n= 24  favored won 100%   State House margin vs Partisan EAI: r 0.91, vs raw vote: r 0.89
     [15, 20)   n= 24  favored won 100%   State House margin vs Partisan EAI: r 0.94, vs raw vote: r 0.94
     [20, 100)  n=109  favored won 100%   State House margin vs Partisan EAI: r 0.97, vs raw vote: r 0.97
   all races: State House margin vs Partisan EAI r 0.954; vs raw vote r 0.954; mean |margin - raw vote| 8.0; mean |margin - Partisan EAI| 6.5
2. Registration margin, active vs active + inactive (Nov 2024):
   1,000+ votes       n=199  mean active - all -0.9  sd 2.1  r 0.998
   100-300 votes      n= 86  mean active - all -1.9  sd 3.4  r 0.994
   300-1,000 votes    n=166  mean active - all -2.0  sd 3.1  r 0.994
   under 100 votes    n= 53  mean active - all -0.9  sd 6.2  r 0.987
   largest differences: Kingman Twp -29.4, Limestone -19.4, Frye Island +13.5, Castine +13.4, Stockholm -12.6, Brighton Plt -12.0
3. Relative lean, 2010-2016 vs 2018-2024 (equal weights), towns with 1,000+ two-party votes in 2024: n=199, r 0.918, rank r 0.909, moved more than 50 ranks: 13, mean |change| 9.1 points
4. Races with non-major share between 10% and 20% (the 15% rule decides):
 year     office district           dem              rep  nonmajor_share           largest_other  qualifies
 1994 U.S. House        2 John Baldacci  Richard Bennett          0.1356       John Michael 8.8%       True
 1996  President      NaN  Bill Clinton         Bob Dole          0.1762        Ross Perot 14.2%      False
 2002   Governor      NaN John Baldacci Peter Cianchette          0.1138    Jonathan Carter 9.3%       True
 2014 U.S. House        2    Emily Cain   Bruce Poliquin          0.1115 Blaine Richardson 11.1%       True
5. EAI weighting grid search copied: 492 schemes (audit script b5.py)
