- Losing votes: all votes for non‑winners.
- Winner surplus: votes above the minimum to win (runner‑up votes + 1).
- Wasted share: wasted_votes ÷ total_votes (reported as %).
- NDA converted votes to seats more efficiently than INDIA.
- INDIA’s total wasted votes are larger despite fewer seats.
- Others (parties outside NDA/INDIA) wasted the vast majority of their votes.
| coalition | total_votes | wasted_votes | wasted_share_% | seats | wasted_per_seat |
|---|---|---|---|---|---|
| NDA | 275,171,127 | 140,904,289 | 51.21 | 289 | 487,558 |
| INDIA | 260,009,278 | 164,811,581 | 63.39 | 227 | 726,042 |
| Others | 110,182,126 | 101,073,334 | 91.73 | 26 | 3,887,436 |
| coalition | seats | mean_votes_polled | median_votes_polled | mean_margin | median_margin |
|---|---|---|---|---|---|
| NDA | 289 | 1,226,188 | 1,238,318 | 185,183 | 146,089 |
| INDIA | 227 | 1,167,621 | 1,142,509 | 136,366 | 103,554 |
| Others | 26 | 1,038,559 | 1,089,479 | 129,007 | 84,073 |
Interpretation. NDA tended to win larger seats and by bigger margins (which should raise its winner‑surplus waste). Yet NDA still shows lower total wasted votes and a lower wasted share than INDIA. The dominant driver is loss‑side waste: INDIA accumulated many more votes in seats it did not win (lots of seconds and some big trailing totals), which swamps NDA’s surplus in its victories.
- BJP: wasted_share ≈ 52.91%; 239 seats. Large winner with moderate waste.
- INC: 69.85%; 99 seats. High waste; many seconds/close losses.
- TDP: 32.90%; 16 seats. Very efficient conversion.
- SP: 46.00%; 37 seats. Efficient, concentrated base.
- AITC: 45.86%; 29 seats. Efficient, concentrated base.
- YSRCP: 83.81%; 4 seats. High waste vs seats won.
- CPI(M): 87.51%; 4 seats. High waste.
- BSP: 100%; 0 seats. All votes wasted.
- IND (Independents): 89.96%; 7 seats. Scattered efficiency.
- FPTP rewards geographically concentrated support and penalizes broad‑but‑shallow support. Two parties can have similar national votes yet very different seat counts.
Candidate‑level results (one row per candidate–constituency) were aggregated. For each seat, all losing votes were counted as wasted; for the winner, wasted = max((winner − runner‑up) − 1, 0). Wasted votes were summed by party and by coalition; wasted share = wasted ÷ total. Coalition buckets reflect contemporary alignments (NDA, INDIA, Others).
- Efficiency ≠ popularity. A lower wasted share indicates stronger vote‑to‑seat conversion, not higher overall support.
- Contest scope matters. Parties contesting widely risk more loss‑side waste; selective contesting can look more efficient.
- Coalition mapping is an assumption. Different mappings can shift the coalition totals.