MacroMatrix

Madurai final Claude model verdict in 12 seats DMK - 8 NDA - 2 Remanning two it is toss it will go either TVK or NDA. DMK voter will decide that seats winner. @TVKVijayHQ #TVKVijay‌HQ #DMK #NDA #BJP #ADMK

@Devarjanv1 @Ayyappan_1504 Annamalai

@MacroMatrix1 @Ayyappan_1504 Who is that "he"?

Annamalai is most likely to contest from Coimbatore North.

Kolathur Claude model predicted Stalin wins. Many might be wondering by seeing young crowd TVK will win in this seat. Model is showing some analysis based on past voting pattern history. In this seat more than 60% of people getting welfare from Stalin Govt and people will prefer more for star candidates than others. This is reality. But TVK surge is huge in this seat and pushing ADMK to third in first election is very big achievement. #TVKVijay‌ #TVKVijay‌HQ #Kolathur #Stalin #Vijay

@Devarjanv1 @Ayyappan_1504 If he contest he will have chance to win in that seat due to three way fight.

@MacroMatrix1 @Ayyappan_1504 அது அந்த அம்மா நினப்பு ஆனால் மேலிடம் வேறு மாதிரி நினக்கலாமில்ல!

@Harish24749455 @TVK_NewsTrichy @TVKVijayHQ It will give trend in every seats not accurate prediction.

@MacroMatrix1 @TVK_NewsTrichy @TVKVijayHQ Good one🔥

Perambur predicted by Claude. Yesterday I have mentioned this as well. Vijay will win this seats. Always people will vote more for star candidates. @TVKVijayHQ #TVKVijay‌HQ #DMK #Perambur

Below is the Claude response most of the people have asked how it will be predicted. Please note it will show trend in every seats not accurate prediction it can have error. 1. Why ML won’t work here: ∙TVK has ZERO electoral history no training data exists ∙You cannot train a supervised model without a target variable ∙Only 2 elections (2016, 2021) with current boundaries ∙not all constituency-level survey data exists to train on 2. What we actually built: ∙A vote transfer model — tracks where 2021 voters go in 2026 ∙DMK retains ~82% of 2021 vote, loses youth to TVK ∙AIADMK 2021 vote splits: ~55% stays NDA, ~10% goes TVK, ~5% goes DMK ∙MNM 2021 vote (~9.5% avg) transfers ~80% to TVK — this is TVK’s readymade base ∙New voters from turnout increase (60%→80%) split ~60% TVK, 25% DMK, 10% NDA 3. Data sources (all real, publicly verifiable): ∙2021 + 2016 ECI results for all 234 constituencies ∙CEO TN official district age-cohort data (38 districts × 8 age brackets) ∙50 independent vox pop sources totaling ~6,20,000 ground respondents ∙All candidate lists (DMK 164, AIADMK 169, BJP 27, TVK 234, NTK 234) ∙Alliance seat-sharing with transfer efficiency modeled per-party 4. What makes this model unique: ∙District-level age demography weighted by turnout assumptions (80%+ all ages) ∙Alliance transfer correction — when BJP/PMK/AMMK contests instead of AIADMK, NDA loses 3-8% ∙Star candidate personal vote adjustments calibrated from 2021 margins ∙Every seat has unique numbers no regional averaging 5. Validation approach: ∙Backtested against MNM 2021 as a new-party proxy — model correctly predicts MNM ~10% urban, ~3% rural, 0 seats ∙Cross-checked against all formal surveys and 50 vox pop independent sources ∙Two completely different methods (vote transfer model vs district-age-turnout model) converge on same result. This is Claude response of its model. It back tested this model against new entrant in TN election.

@MacroMatrix1 @TVK_NewsTrichy @TVKVijayHQ What model are you using and what data?

I think if Vijay have campaign all constituencies it might have more effect on this election. Especially in rural areas it might have helped to convert few seats. Unfortunately it is too late for him. Politics is different in ground reality and in micro level every constituencies there are some big heavy weight is there from both the big party. If you have seen my earlier post from Claude model madurai prediction you might understand how difficult to win single seats. At least he can focus on Urban cities and try to convert into some seats. #TVKVijay‌ #TVKVijay‌HQ @TVKVijayHQ

Below is the Claude response most of the people have asked how it will be predicted. Please note it will show trend in every seats not accurate prediction it can have error. 1. Why ML won’t work here: ∙TVK has ZERO electoral history no training data exists ∙You cannot train a supervised model without a target variable ∙Only 2 elections (2016, 2021) with current boundaries ∙not all constituency-level survey data exists to train on 2. What we actually built: ∙A vote transfer model — tracks where 2021 voters go in 2026 ∙DMK retains ~82% of 2021 vote, loses youth to TVK ∙AIADMK 2021 vote splits: ~55% stays NDA, ~10% goes TVK, ~5% goes DMK ∙MNM 2021 vote (~9.5% avg) transfers ~80% to TVK — this is TVK’s readymade base ∙New voters from turnout increase (60%→80%) split ~60% TVK, 25% DMK, 10% NDA 3. Data sources (all real, publicly verifiable): ∙2021 + 2016 ECI results for all 234 constituencies ∙CEO TN official district age-cohort data (38 districts × 8 age brackets) ∙50 independent vox pop sources totaling ~6,20,000 ground respondents ∙All candidate lists (DMK 164, AIADMK 169, BJP 27, TVK 234, NTK 234) ∙Alliance seat-sharing with transfer efficiency modeled per-party 4. What makes this model unique: ∙District-level age demography weighted by turnout assumptions (80%+ all ages) ∙Alliance transfer correction — when BJP/PMK/AMMK contests instead of AIADMK, NDA loses 3-8% ∙Star candidate personal vote adjustments calibrated from 2021 margins ∙Every seat has unique numbers no regional averaging 5. Validation approach: ∙Backtested against MNM 2021 as a new-party proxy — model correctly predicts MNM ~10% urban, ~3% rural, 0 seats ∙Cross-checked against all formal surveys and 50 vox pop independent sources ∙Two completely different methods (vote transfer model vs district-age-turnout model) converge on same result. This is Claude response of its model. It back tested this model against new entrant in TN election.

@MacroMatrix1 Can you explain how this model works

Madurai South Claude model is predicting it is three way fight and slight edge to TVK due to DMK ally weakness. Watch out for this seat. #TVKVijay‌HQ #DMK #BJP #ADMK #NDA #MDMK #TNElection2026

@vetriox03 @georgeviews Claude Opus 4.6 model

@MacroMatrix1 @georgeviews Which app is this?

Mark my words. The combined votes of DMK's R. D. Shekar and PMK's Thilagabama would still fall way short of Thalapathy Vijay's votes in Perambur! He is heading for a historic win. 🧨

@Ahmedshabbir20 This model is before alliance consider if DMK symbol stand there. Now it will be going to be tough fight for VCK vs TVK.

VCK Chief Thirumavalavan MP to contest from Kattumannarkoil constituency.

Below is the Claude response most of the people have asked how it will be predicted. Please note it will show trend in every seats not accurate prediction it can have error. 1. Why ML won’t work here: ∙TVK has ZERO electoral history no training data exists ∙You cannot train a supervised model without a target variable ∙Only 2 elections (2016, 2021) with current boundaries ∙not all constituency-level survey data exists to train on 2. What we actually built: ∙A vote transfer model — tracks where 2021 voters go in 2026 ∙DMK retains ~82% of 2021 vote, loses youth to TVK ∙AIADMK 2021 vote splits: ~55% stays NDA, ~10% goes TVK, ~5% goes DMK ∙MNM 2021 vote (~9.5% avg) transfers ~80% to TVK — this is TVK’s readymade base ∙New voters from turnout increase (60%→80%) split ~60% TVK, 25% DMK, 10% NDA 3. Data sources (all real, publicly verifiable): ∙2021 + 2016 ECI results for all 234 constituencies ∙CEO TN official district age-cohort data (38 districts × 8 age brackets) ∙50 independent vox pop sources totaling ~6,20,000 ground respondents ∙All candidate lists (DMK 164, AIADMK 169, BJP 27, TVK 234, NTK 234) ∙Alliance seat-sharing with transfer efficiency modeled per-party 4. What makes this model unique: ∙District-level age demography weighted by turnout assumptions (80%+ all ages) ∙Alliance transfer correction — when BJP/PMK/AMMK contests instead of AIADMK, NDA loses 3-8% ∙Star candidate personal vote adjustments calibrated from 2021 margins ∙Every seat has unique numbers no regional averaging 5. Validation approach: ∙Backtested against MNM 2021 as a new-party proxy — model correctly predicts MNM ~10% urban, ~3% rural, 0 seats ∙Cross-checked against all formal surveys and 50 vox pop independent sources ∙Two completely different methods (vote transfer model vs district-age-turnout model) converge on same result. This is Claude response of its model. It back tested this model against new entrant in TN election.

@MacroMatrix1 @VinothTvk06 Hello, I have a curious question: What data sources does your model analyze? For example, does it use X (Twitter), Facebook, Instagram, or other social media platforms?

Final kongu region Claude model predicts DMK - 7 NDA- 18 TVK-4 More than 6-8 seats can go for toss to other party., Most of the seats are very close fight. #TVKVijay‌HQ #DMK #ADMK #BJP #TNElection2026

You are right but it will happen in urban cities. Every region is very different and TN is most polarised state election in India. You can see my Madurai seat predcition what Claude is finding out and how every party is selecting candidates. In every seats both two parties have some big heavy weight. That is why even after huge anti incumbency ADMK scored 60+ seats. You will understand the dynamics of politics how it will work.

@MacroMatrix1 It is not about gen-z vote they are the influencers in their family they might convert minimum 1 member in their family that's what he covers them

Today’s Vijay speech in Perambur he mentioned GenZ and youngster. Even my Claude model shows that he is gaining more vote share in younger section. If younger section turn out comes more for this election, little 2-3% can swing some seats to TVK favour in urban cities. He knows his data where vote is coming from. @TVKVijayHQ #TVK #TVKVijay‌HQ

@DevanandRaghul Sure this is predictive mode just shows the trend not accurate prediction. To minimise error in estimation need crore of money to get some what accurate prediction.

@MacroMatrix1 Oh thanks much ur model is the most interesting one and different among others so excited too see the results

@AlthafShaikh7 @TVKVijayHQ I have mentioned to more than 50 people bro. Check my past replies I might have pasted the Claude response to this question.

@MacroMatrix1 @TVKVijayHQ Just curious what kind of data are you using to build this model and prediction

@DevanandRaghul I am running every region wise bro. Even it will be having limit to analyse, it is not unlimited. I have used all. If I want to top up it will be more than 12000 rupees. It is high cost bro. I will try to run as much I can in two weeks.

@MacroMatrix1 can you run ur model and predict the total number of seats tvk is winning

This model will show trend not accurate prediction. Margin can vary with some error. Even TVK gain 3-4% Stalin will win. Almost more than 50% of people in Kolathur is getting welfare. Model is indicating that. And star candidate in the local region people will give more importance. Model analysed voting behaviour of star candidate by using past all historical election. Kolathur is safe for Stalin.

@MacroMatrix1 @TVKVijayHQ Will it change in near future. Today u saw the crowds for TVK and DMK in Kolathur.

@NarayanaPrasat8 Yes which shows BJP weakness and NDA alliance is not jelling there. This is what I can infer from the data.

@MacroMatrix1 In 21 it was not lotus if am not wrong. In 24 it is lotus without admk.

Surprisingly in Mylapore Claude is predicting TVK will win. #Vijay #TVKVijay‌HQ #Mylapore

@DilipvAL @TVKVijayHQ Stalin is winning and TVK is coming second.

@MacroMatrix1 @TVKVijayHQ Wat Abt kolathur

@VamsiChandran Melur model predict it is comfortable NDA wins. Usilampatti tight race between DMK alliance Vs NDA. Due to OPS faction halo effect.

Melur and Usilampatti are 2 seats where Dmk is structurally very very weak, they have never won Melur and Usilampatti they have won only once, getting such seat as part of alliance is no big achievement. In place of getting 1 good winnable seat, Cong have got 2 scape goat seats in Madurai, spoiling their strike rate. Same mistake done in Coimbatore as well.