IndiaVotes YouPredict: An Elections Game

Published May 9, 2024

1

Background

I was recently thinking about how to combine multiple of my interests: elections, email (E3), and prediction markets. And that’s where the idea of an email-based election results predictions game came up. I should probably have thought about this a few months earlier, but it’s never too late! I don’t know whether it will work out or not, but it’s too good an idea not to be tried.

Let’s start at the beginning with IndiaVotes. Here is the story from a previous essay The Battle for 2024: “I like doing election analysis. I had created an election data website, IndiaVotes, as early as 1998. I had even created an “Indian Political Stock Exchange”. Both those were when I was running IndiaWorld. After I sold the business to Sify in late 1999, my direct interest in elections data waned. It was only in 2009 that I got interested again. In 2011, I wrote a blog post suggesting a path that would get BJP to a majority in the 2014 elections. Later that year, I bought the domain IndiaVotes.com, and (re-)launched it as a website with in-depth data on all Lok Sabha elections since 1952 and all State elections since 1977. While the site probably needs a refresh, it remains one of the best sources for data and some basic analyses on Indian elections.” Incidentally, traffic on IndiaVotes is at an all-time high. (Given the interest, maybe I will get an AI-first redesign done after the Lok Sabha elections.)

Next comes email. I have been writing about how to popularise AMP in Email and make email a more powerful marketing channel in India.  In a recent essay An Email 2.0 Newsletter for All, I wrote: “The key to making Email 2.0 as the next new channel for marketers for their customer engagement messaging is by bringing attention to the email inbox with a service which becomes part of our lives. What is needed is an Email 2.0 newsletter which we all want to subscribe to – and read as a daily habit. This is the mountain which needs to be climbed to make Email 2.0 a commercial success. WhatsApp grew thanks to its easy personal and group messaging capabilities on the mobile, which started replacing the clunky SMS which never grew up. Email 2.0 needs a “killer app” – something which gives us a reason to open our inboxes, and then share with our network. What can such a service look like?” I then went on to describe the solution: “The Email 2.0 newsletter is built on the foundation of three key ideas: the interactivity AMP brings in, the incentives and gamification that Atomic Rewards enables, and the ease of content creation and personalisation that Gen AI can power. These together are the pillars of creating a new type of daily email service delivered to the inbox, garnering attention, and demonstrating to marketers (as consumers of such a service) the infinite possibilities of what Email 2.0 can do for their brands.” Could a short-term elections “newsletter” for a game do the trick?

Which brings us to prediction markets. In Mu and Prediction Markets, I wrote: “I have been fascinated by prediction markets for some time. The idea of using the wisdom of crowds (or a subset) to predict future outcomes as opposed to relying on a few experts is quite compelling… While opinion polls and surveys are the norm globally for political predictions, prediction markets could provide an interesting alternative…While we can rely on experts (or even the grey market – the satta bazaar) to get some answers, the “wisdom of crowds” can provide an alternate (and perhaps better) view on outcomes.”

IndiaVotes YouPredict brings the three ideas together into an email-based predictions game for the ongoing Indian elections.

2

The Rules

Here is how the game can be played. Participation is open to everyone with an email address. Like I wrote, it is an experiment by which I hope to popularise AMP in Email and demonstrate how building a B2C property does not necessarily need a website or app; it can all be done inside an email.

The IndiaVotes YouPredict game is about predicting which party will win each of the 543 Lok Sabha seats. (It has been kept at the party-level rather than alliance-level because in Indian politics, there is no guarantee that today’s pre-poll partner will continue post-poll!) Players can predict outcomes for any number of seats. Winners will be determined based on a simple scoring system for the 543 questions: +2 for every right answer, -2 for a wrong answer, and 0 if no selection was made. (It is a bit like our scoring system for quizzes. Try QShots in case you haven’t.) Answers can be changed till the close of the last round of polling on June 1. Players can share their predictions on social media or keep it private. The daily emails to subscribers will have the aggregate predictions.

Players subscribe to the game with their Gmail or Yahoo ID. (AMP does not work with other email domains.) All actions after that are done in the email itself. (The email needs to be opened in the Gmail or Yahoo apps, or the browser; Apple’s iOS native mail client does not support AMP.)

There are no prizes, just bragging rights after the election results come out on June 4. I know that for a true predictions market, participants need skin in the game and therefore needs money. In India, the rules (and taxes) with real-money gaming are complicated so we decided not to go down that path.

What is novel is the fact that the entire game is to be played within the email. From selection of a state, then a seat, and eventually a party – everything is done in-place. All the data is stored on a server and the email interacts with the backend in real-time to create an engaging and mesmerising experience.

This is a very different approach from surveys and opinion polls which asks a random sample about who they are likely to vote for. In this game, the aim is to predict who will win. So, it is about collective intelligence – wisdom of crowds. In that sense, it is like prices in a market – all the info is baked in. Of course, since there is no money involved, results can be skewed by simply guessing, but that’s something to be addressed in future versions.

3

Enhancements

There are some enhancements which I am hoping we can do in this election, and then there are many ideas for the future.

To make the game a success, it will need to spread virally. I am hoping that happens! These games only work well when there are many participants. This may not be easy for the first iteration since there are no rewards to be won, other than a bit of fame post elections. Maybe we can offer Atomic Rewards (Mu) for referrals.

One possibility is to have Mu as a token/currency for participation. I had discussed a similar idea in Mu and Prediction Markets: “As consumers own (“mine”) Mu for their attention, data and other brand-incentivised actions, prediction markets can be a good potential redemption (and additional earning) mechanism. It brings out the latent wisdom each of us have in our interpretation of future events. By removing the need for real money, MuCo’s Prediction Markets also eliminate the need for regulation. The use of crypto tokens ensures that MuCo cannot debase or devalue the tokens, and there are rules that govern it. Google helps us search the past, Twitter the present, and Mu-powered Prediction Markets (MPM) can help us with the future… The working of MPM would be very much like prediction markets today which are fuelled by real money. People could launch their own public or private bets and contracts, and invite others to compete against them. A contract needs two parties. MPM would, in the initial days, work as a market maker to ensure liquidity. There should be clearly defined outcomes to determine payoffs.”

Another idea is to create a “league” – where people who know each other can share more detailed predictions and later compare their scores. (Future: there could even be a “league” entry.)

We could show how predictions are changing over time.

We could give greater weightage (or credence) for those predicting outcomes for the seat where they are voting. Presumably, they will know the situation better in their own constituency than elsewhere. For this, we will need to allow players to pick their “Home” constituency.

We can adapt the game for the state elections. Elections keep happening every few months in India!

We could also ask players for more personal information, so predictions can then also be shown by profiles (gender, age, income levels, urban/rural, and so on).

One big idea for the future is to bring in payments. While the “Free” version of the game can continue, there is a “Paid” version – where players pay Rs X for every prediction they make, and then the winners share the kitty. (Taxation will eat away a chunk as has happened with real-money games. Presumably, this will be seen as a game of skill, rather than chance.)

Another idea in the future is to create the equivalent of a real-time predictions market like the Iowa Electronic Markets for the 2024 US elections.

In future, predictions could be expanded to many other areas: economic indicators, government policies, box office outcomes, rainfall, food production, sports, current events, company announcements, and so on. See Polymarket for the possibilities.

So, a fascinating new world awaits! Go, play, and spread the word.

4

AI Inputs – 1

I asked the AIs what could be done to make the game better. Here are the inputs.

ChatGPT

1. User Experience and Interface Enhancements:

1. • Improve Visual Design: Ensure the email layout is visually appealing and intuitive. Use interactive elements like buttons and sliders that are easy to use within the email environment.
   • Mobile Optimization: Given the high mobile usage in India, ensure that the game interface is fully optimized for mobile platforms, particularly for Gmail and Yahoo apps.

2. Incentivization and Engagement:

1. • Introduce Rewards: While real-money incentives are complicated, consider virtual rewards like badges, levels, and leaderboards to motivate participation and retention.
   • Referral Programs: Implement a referral program that rewards users with points or virtual currency for bringing new players into the game.

3. Community and Social Interaction:

1. • Social Sharing Features: Enhance the ability for players to share their predictions and scores on social media. This could help in virally marketing the game.
   • Create Groups or Leagues: Allow players to form groups or leagues where they can compete against friends or colleagues, adding a social layer to the game.

4. Data Utilization and Transparency:

1. • Real-time Updates: Provide real-time updates on predictions and results, making the game more dynamic and engaging.
   • Detailed Analytics: Offer players insights into their performance and the overall trends in the game, which could help in strategizing their predictions.

5. Gamification Elements:

1. • Progressive Difficulty Levels: Introduce levels of difficulty with progressively harder predictions that offer greater rewards.
   • Milestone Achievements: Set up milestones for participation, such as completing predictions for all constituencies in a state, and reward players upon achieving these milestones.

6. Educational Components:

1. • Informational Content: Provide background information on constituencies, historical voting patterns, and key candidates to educate players and help them make informed predictions.
   • Tutorial and Guides: Offer tutorials on how to make predictions and how the scoring system works, which could be particularly useful for new players.

7. Integration of Advanced Technologies:

1. • Artificial Intelligence (AI) Insights: Use AI to provide predictive insights based on historical data and current trends which players can use as a reference.
   • AMP and Interactivity Enhancements: Continue to explore and implement Advanced Mobile Pages (AMP) features that enhance interactivity within the email, such as dynamic forms and real-time result display.

8. Diverse Participation Encouragement:

1. • Localization: Ensure the game content is available in multiple regional languages to broaden its appeal across different linguistic demographics in India.
   • Accessibility Features: Implement features that make the game accessible to people with disabilities, such as screen reader compatibility and alternative text for images.

Claude

1. Incentives/Rewards: Since there are no prizes currently, introducing incentives or rewards could increase participation. You mentioned using Atomic Rewards (Mu tokens) for referrals or as a currency within the game. Offering prizes (even small ones) for top predictors could drive more engagement.
2. Social Features: Adding social sharing and league/group features could make it more viral and competitive. Allowing people to create private leagues with friends/family and see comparative predictions could boost interest.
3. Enhanced Analytics: Providing data visualizations on how predictions are changing over time, by demographic groups, geographic regions etc. could make it more insightful and sticky for participants.
4. Credibility Weighting: Giving higher weightage to predictions from people’s home constituencies, as you mentioned, could make the aggregate predictions more accurate.
5. Expanding Scope: Adapting it for state elections and expanding to other domains like economic indicators, policy predictions etc. after this pilot could make it an ongoing, sustainable platform.
6. Payments/Real Money: Introducing a paid version where users have skin in the game with real money predictions could increase strategic participation, like other prediction markets.
7. Improved UX: An AI-first redesign enhancing the user experience, making it more interactive and visually appealing within emails could drive better engagement.
8. Marketing Push: Promoting it effectively across social media, news websites and with influencers could help get a critical mass of initial participants for better crowd wisdom.

Gemini

• Offer rewards for participation: The game currently offers no prizes, relying solely on bragging rights. The author suggests using Atomic Rewards (Mu) for referrals, where users can earn Mu for referring others to the game.
• Introduce a league system: Create a league system where people can compete against each other and compare scores. This can be done by allowing users to join groups or “leagues” with friends or colleagues.
• Track prediction changes: Implement a feature to show how predictions change over time. This will allow users to see how popular opinion sways towards different candidates or parties.
• Weigh predictions by location: Give weightage to predictions made by users for their own constituencies. The assumption is that users will have better knowledge about the local political situation and candidates in their area.
• Adapt the game for state elections: Since elections in India happen frequently, the game can be adapted for state elections as well.
• Allow users to share profiles: Collect optional demographic data from users such as gender, age, income level, and location (urban/rural). This will allow for segmenting predictions by these profiles.
• Introduce a paid tier with real money rewards: Create a paid version of the game where users pay a fee to participate and the winners share a prize pool. This would require navigating legal and tax implications.
• Develop a real-time prediction market: The author envisions creating a real-time prediction market similar to the Iowa Electronic Markets for the US elections.
• Expand predictions to other areas: Broaden the scope of predictions beyond elections to include economic indicators, government policies, and other events. The author references Polymarket as an example for exploring various possibilities.

5

AI Inputs – 2

I then asked the AIs: “What other email-based prediction markets could I create for India?”

ChatGPT

1. Bollywood Box Office Predictions:

1. • Participants could predict the opening weekend numbers, total box office collection, and longevity of major Bollywood releases. This could also extend to regional film industries like Tollywood, Kollywood, and others.

2. Cricket Match Outcomes:

1. • Cricket is a major passion in India. An email-based prediction market could focus on outcomes of IPL matches, international cricket series, and other domestic leagues. Predictions could include match winners, top scorers, and tournament standings.

3. Economic Indicators:

1. • This market could involve predictions about key economic indicators such as GDP growth rates, inflation, and stock market indices. Participants could predict quarterly or annual economic outcomes based on current events and policies.

4. Monsoon and Weather Predictions:

1. • Given the significant impact of monsoon on agriculture in India, a prediction market focusing on rainfall levels, timing, and impact on crop yields could engage a wide audience, including those in the agricultural sector.

5. Government Policy Outcomes:

1. • Participants could predict the outcomes of proposed government policies or legislative actions. This could include predictions on the passage of key bills, policy implementation success, or even outcomes of government initiatives.

6. Television Ratings and Reality Show Winners:

1. • With a vibrant television industry, predictions could be made about ratings of popular TV shows or winners of reality shows like “Bigg Boss,” “Indian Idol,” or “Dance India Dance.”

7. Technology Adoption Rates:

1. • Predictions could focus on the adoption rates of new technologies such as electric vehicles, solar power installations, or digital payment systems in India.

8. Real Estate Market Trends:

1. • Participants could predict housing market trends in various cities, including price increases or decreases, demand in commercial versus residential sectors, and the impact of new infrastructure projects.

9. Social Trends and Public Opinion:

1. • Markets could be created around major social changes and trends, such as shifts in public opinion on critical issues like education reforms, privacy concerns, or public health initiatives.

10. Educational Outcomes:

1. • Predictions about results of major examinations like the UPSC, JEE, or NEET, or the performance of educational institutions in national rankings.

To implement these email-based prediction markets effectively:

• Leverage AMP in Email: To make the prediction experience interactive directly within the user’s inbox.
• Incorporate Real-Time Data: Ensure that the data used for making predictions is up-to-date and reflects the latest trends and information.
• Engage with Experts: Partner with industry experts to provide insights and analysis that can help participants make more informed predictions.
• Promote Community Engagement: Encourage discussions and community engagement around predictions to build a more active participant base.
• Ensure Compliance: Be mindful of the legal implications, especially with topics that might involve sensitive data or privacy concerns.

Claude

1. Economic Indicators: You could run prediction markets around key economic indicators like GDP growth rates, inflation rates, interest rates, forex rates etc. There is a lot of interest and speculation around these numbers.
2. Corporate Earnings/Events: Allow people to predict quarterly revenues, profits, major announcements etc. for leading Indian companies across sectors like IT, banking, automotive etc.
3. Sports Events: Cricket is a national obsession in India. You could have prediction markets for outcomes of major tournaments like IPL, World Cups as well as individual match/series results.
4. State Elections: As you mentioned, running prediction markets for state assembly elections that happen every few months could sustain interest.
5. Annual Budget: Every year’s Union Budget is widely discussed. You could let people predict tax rates, spending allocation, fiscal deficits and other key budget numbers.
6. Monsoon: Rainfall Given the importance of monsoons for Indian agriculture, predicting rainfall levels across regions could have wide interest.
7. Policies/Legislation: There could be prediction runs on whether certain economic policies or legislations get approved by the government or not.
8. Blockbuster Movies: The Indian movie industry could lend itself well to predicting opening week/lifetime collections for major film releases across languages.
9. Reality Shows: Popular television reality shows with a voting process like Bigg Boss, Indian Idol etc. could leverage prediction markets.
10. Startup Unicorns: In the dynamic Indian startup environment, you could let people predict which startups will attain billion dollar valuations or get acquired.

The key would be identifying areas of wide public interest and information gaps where the “wisdom of crowds” prediction markets could provide unique insights compared to traditional expert forecasts. Frequent event-based cycles can help sustain consistent engagement.

Gemini

Current Events and Pop Culture:

• Cricket Matches: This is a huge potential market in India. You could offer predictions on match winners, top scorers, or specific events within a match.
• Movie Box Office Performance: Generate predictions on opening weekend collections or total box office gross for upcoming movies.
• Reality TV Show Winners: Capitalize on the popularity of reality shows by creating prediction markets for winners or specific outcomes within the show.
• News and Events: Allow users to predict the outcome of ongoing news stories or upcoming events, like policy changes or celebrity actions.

Business and Finance:

• Stock Market Movements: While regulations might prevent real-money trading, you could offer a prediction market for stock price movements within a specific timeframe.
• Interest Rate Changes: Let users predict the direction of upcoming Reserve Bank of India (RBI) interest rate decisions.
• Company Announcements: Create markets for users to predict upcoming company announcements, mergers, or product launches.

Other Ideas:

• Weather Forecasts: Allow users to predict seasonal rainfall or extreme weather events.
• Food Production: Generate predictions on crop yields or food price fluctuations.

General Considerations:

• Focus on Specific Events: For each market, choose a well-defined event with a clear outcome to avoid ambiguity.
• Data and Information: Provide users with relevant data and news articles within the email itself to aid informed predictions.
• Engagement Features: Consider incorporating features like leaderboards, discussions, or social sharing to boost user engagement.

Remember, the success of these markets will depend on attracting a large user base.

• Partnerships: Collaborate with media outlets, sports teams, or other relevant organizations to reach a wider audience.
• Social Media Integration: Promote the game on social media platforms to attract new users.