Every seat-sharing negotiation between the Congress and the Samajwadi Party starts the same way. One side asks for more than it expects to get. The other calls the ask unreasonable. Both insist, in public, that the alliance is intact. This is the script playing out again now, as the Congress and SP work out how to divide Uttar Pradesh's 403 assembly seats for the 2027 election. A first round of formal talks, expected around November 17, was deferred, Rahul Gandhi and Akhilesh Yadav spoke informally instead, at Sonia Gandhi's book launch in New Delhi, with a fresh round expected in the city around November 23-25.
Congress has asked the SP for 181 seats. SP chief Akhilesh Yadav confirmed the figure himself at the HT Leadership Summit on November 14, when he was asked if the number was close to 180 and replied, "add one to that." SP's camp has called the number impractical. SP MP Shivpal Singh Patel told media he expects Congress to settle for "50 to 60 seats, or maximum 65." Other reports put the SP's eventual offer anywhere between 30 and 80 seats, depending on who in the party is talking.
None of this is new theatre. It happened before the 2017 assembly poll and again before the 2024 Lok Sabha election. What is worth asking, while the two sides argue over a number, is which number actually works for the Congress. Its own record in Uttar Pradesh gives a fairly direct answer, and it does not support contesting 181 seats.
Look at four elections where the arithmetic is public and final.
In 2017, when Congress and SP were allied for the assembly election, Congress contested 114 seats and won seven, a strike rate of 6.14 percent the ECI's final certified count. SP, contesting 311 seats, won 47, a strike rate of 15.1 percent.
In 2019, when the SP left the Congress out of its "Mahagathbandhan" with the BSP, Congress contested Lok Sabha seats in UP alone, fielding 67 candidates (it stood down in only 10 of the state's 80 seats to support the SP-BSP-RLD alliance). It won one, Rae Bareli, out of 67 contested, a strike rate of 1.5 percent which was far lower than the 6.25 percent earlier reporting suggested. SP, allied with the BSP that year, won five of 37 contested, a strike rate of 13.5 percent.
In 2022, Congress and SP were not allied at all. Congress contested the assembly election alone, in all but four of the state's 403 seats, and won two. That is a strike rate of 0.5 percent. It lost its security deposit in 387 of the 399 seats it contested, according to Election Commission data reported at the time. SP, running its own alliance with the RLD and others, contested 347 seats on its own ticket and won 111, a strike rate of 32 percent.
In 2024, back in alliance for the Lok Sabha election, SP gave Congress 17 seats. Congress won six of them, a strike rate of 35.3 percent, the best it has managed in any of these four elections. SP, contesting 62, won 37, a strike rate of 59.7 percent.
A useful way to read this is not the raw strike rate but Congress's strike rate as a share of SP's, in the same election. In 2017, Congress converted at 41 percent of SP's rate. In 2019, without SP, and fielding a far larger slate than earlier reports suggested, it converted at just 11 percent of the rate SP managed with the BSP, its second-worst showing, not its second-best, once the seat count is corrected. In 2022, going entirely solo, that ratio collapsed further to 1.6 percent. In 2024, on a small, SP-allocated slate of 17 seats, it climbed to 59 percent of SP's rate, easily its best showing of the four.
That last number is the real argument for a small, carefully chosen Congress slate rather than a large one. When SP picked the seats, Congress performed closer to SP's own level than it has at any point in a decade. When Congress picked its own 403-seat battlefield, it barely registered.
This is not a case that proves itself cleanly, and a seat-sharing committee would be right to press on the gaps. The four elections mix Lok Sabha and assembly polls, which pull different voters to the booth for different reasons. The 2024 sample is six seats, two of them Gandhi family strongholds, which is not enough to generalise with confidence.
The one time Congress was both allied with SP and given a large slate, in 2017, it still won only seven seats, so a bigger Congress allocation inside an alliance has not actually been tested at a size between 17 and 105. Congress also has a specific argument for why 2022 should not be the baseline: it says minority, Dalit and young voters have shifted toward it in the months since, a claim that is the party's own reading of ground reports. No independent, quantified survey confirming this shift was found; published reporting (The Hindu's September 2026 data piece on Dalit voters in UP; Hindustan Times' June 2026 piece on Congress's Dalit outreach) covers strategy and organisational moves, not polling evidence of an actual vote shift.
What the comparison does rule out, fairly cleanly, is the idea that a bigger Congress slate produces a proportionally bigger Congress result. The one large allied slate it has had, in 2017, delivered a strike rate worse than its small allied slate in 2024. The one time it contested almost the entire state alone, in 2022, it lost its deposit in 97 percent of the seats it fought.
Set aside the back and forth over 60 versus 80 versus 181. The sharper version of the question is this: what evidence, beyond its own reading of recent voter movement, shows that Congress can convert a meaningful share of 181 seats when it won single digits from a comparable-sized, SP-allied slate of 105 to 114 seats in 2017?
Nobody on either side expects the next round of talks to end with Congress at 181 or SP at 30. The number will land somewhere in between, the way it always does. But if Congress wants that number to be closer to its ask than to the SP's offer, its own decade of results in Uttar Pradesh is not the argument for it. The SP's caution about grassroots morale, uncomfortable as it is to hear from an ally, is backed by more recent and more consistent numbers than Congress's own case for a bigger share.