T20 Fantasy Team: Gaming the Game
As my near ones could tell you, I used to follow cricket passionately but with one exception. I had no interest in the IPL or T20 in general for the best part of last decade.
All through the decade, mobile phones got more functional, network coverage improved and thanks to Motabhai, data charges crashed. This led to online gaming gaining traction and soon enough, middle aged Indians were playing Ludo online in Mumbai locals, on their way back home while teenagers were nonchalantly swiping their parents’ cards to make purchases in Pub G, sometimes racking up enormous bills.
With some favorable rulings, money based Fantasy gaming began to take roots as Dhoni told us to play “dimaag se” and Indians found a platform where they could show off AND monetise their knowledge in a sport which they live, breathe, discuss and dissect every living moment, from boardrooms to rest stalls. Admittedly, I am a bit of a snob about my own knowledge of the game and consider myself, if we are being modest, above average*. So just like others around me, the combination of IPL and Fantasy gaming offered me the perfect dopamine hit.
- Exhibit: Ganguly, the considered by some to be India’s best captain ever, had won a grand total of 1 final out of the 14 he led the team in. The same one in which he turned the Lord’s balcony into a small retail store located somewhere in the bylanes of Kolkata at the peak of summer by taking his shirt off; a moment etched in every Indian cricket fan’s memory.
To validate my delusions, I began by participating in free mega contests just to see how good I was. I wasn’t. Most times, I’d finish in the 40–70 percentile, performance which would lose me miniscule amounts of money over a season if I played money games. I also started participating in small friends and family groups of 4–8 players wherein I consistently lost money. So clearly, I wasn’t good at it. That made me question my own existence and I entered a bottomless spiral of alcohol and gambling, losing everything in the process.
No I did not.
It did make me think though. I also rewatched Moneyball at that point. The similarities between Cricket and Baseball are well known. (The similarities between Brad Pitt and I, although significant, are not that well known). Both the sports essentially involve players with separate skill sets to form a team and try to win by hitting the ball more frequently and better than the other team. I wondered if I could get rid of my emotions and biases and reduce living, breathing players to skill sets and turn myself into a mean, logic driven machine. For the next six months, I sat about daydreaming what I would achieve if I had my own Jonah Hill.
As time went by and another IPL season approached, whatsapp groups began heating up with cricket news, predictions, proposals to form own fantasy leagues. In short, with people attempting to show off their cricketing knowledge. In one of those groups, one that had people who have an exceptional understanding of the game, a friend, Chinmay, mentioned about how he had got hold of IPL data dating back to the beginning of time. Voila! The moment I had been daydreaming about. Over the course of next IPL, I played high roller matches, won huge amounts and retired to a Caribbean island well before hitting 40.
No, I did not.
In reality, it was a shocker. I was a dog chasing a car who had managed to catch up to it and did not know what to do next. This was a very complete data set. It had every detail about every IPL match, including the ground, the batting position and the bowling style. I shared my vision on what I thought I could achieve with this data and Chinmay, a data genius, readily agreed to partner with me on this. It was the perfect match too. I had the illusion of cricketing knowledge while he had actual knowledge about cricket AND SQL; something essential to our plans.
For the first few days, I kept asking him to pull out random numbers in the hopes of hitting a jackpot. With time, it dawned upon us that this database only had IPL numbers, not Bitcoin keys and so, some sanity was in order.
At this point, my experience as a Product Manager kicked in and I turned a fun exercise into a boring work thing by forcing ourselves to answer some questions; “What are we looking to prove? How will we prove that? What would a fantasy team, based on data look like? What are the criteria for success? What would we do with the millions that we are about to make?”
And then we clearly answered all of these questions:
We are trying to prove that a data based approach to fantasy gaming gives you an edge over the regular player and you are will not lose money over the long run”. We deliberately chose “not losing over “winning” as the key word since you need to consistently keep recovering the money to get a shot at winning big.
“We will prove this by entering small sized contests” (The idea of winning mega contests was dropped at this point. I chanced upon a study which showed that consistent wins can only happen in small groups. In large groups with long tail, there will always be a monkey picking the perfect team.)
Data based team: this was the most difficult question to answer. We examined many combinations and finally settled on skill set as the primary driver, with some weight to players’ personal record and the conditions. Conditions were unlikely to impact an individual player’s selection but in deciding the team combination: more batters or bowlers.
Once the answers were in place, Chinmay patiently put in the extra work of extracting data for a week or so as we tried to finetune the rules and even tested them by picking rule based teams in past matches and checking our standing. We even decided to keep tweaking the rules based on new evidence, running up to one week in to the tournament.
Armed with this knowledge, we began playing our respective leagues. As the tournament went on and results came in, we observed that we had been reasonably successful in gaming the game with more wins than losses. And then we repeated it for every subsequent IPL.
No, we could not.
Covid struck. The IPL was first abandoned and then moved to another country with quite different conditions. This completely paid put to our data models and we began losing in the second leg right from the get go. In the middle of it though, we changed tack and without swaying from the basic rules, decided to tweak the skill set a bit by picking skills which were demonstrably more effective in UAE than they were in India. Or skill which were useful in India but not so much in UAE. One example of this is a bowler, whom I regularly picked as a VC in the Indian leg, was an absolute disappointment in the second leg.
As the final begins tonight and we place our last IPL bets for the year, even if I lose, I’d be ending the season in green. The formula however, was struck by a Covid named train in the middle of the season and although quite successful already, it will be retweaked and reapplied next year before I can guarantee its success and sell it for millions.
PS: I generally play in two groups, one consisting of college friends and colleagues who are sane and logical in their picks. The other is a set of cousins who “feel” a major knock or a hat trick incoming. My findings are that it’s easier to win against sane, predictable people than against a bunch of wackos.
