First of all, before beginning this new journey, I would like to say sorry for my complete absence from this section post T20 Simulator. I was excited, was building and then work hit me hard. I was struggling both financially as well as personally since I was completing my Masters and Working full time. Being a Software Engineer is not easy and its really difficult to be in front of computer screen 24 hours. When you are young (not that I am old), you can spend as much time on it but after a certain period of time 8 hours in front of computer is all you can take.
I would like to apologize to @LBW. since he was my partner and I did not make him aware as well. I feel terrible about it. I am sorry.
Next moving on from apologizes, its time to look forward and start fresh. I have lost all my previous code since a lot has changed since last January. My wife spilled water on my mac, I completely destroyed it to a point that I could not even recover my data. I even lost my thesis work on that and had to start things from scratch.
Not to bore you any more, I will be working on a new cool project called AI Cricket. I made a mistake by promising a lot of things previously so will no longer be doing that. I will start small, release those small little things and move from there.
What is AI Cricket?
That is a good question. A little bit of background, I recently completely my thesis where I used various machine learning algorithm to predict which city in New Jersey lacks transportation and needs one. Back in January 2019, I had this idea where I could use these algorithms and techniques and can actually let AI predict true outcomes based on historical data.
For people that are non tech friendly, it is basically reading all the stats and predicting outcome based on that stats. For example, Virat Kohli scoring 100, based on his recent innings AI can learn his pattern given few different conditions such as opponent, weather, conditions, etc.
So basically AI cricket will be a simulator that would actually work based on learning things from already existing data and with new data.
What is your goal?
Well, my goal is simple this time. As mentioned I am going to take small baby steps and achieve those and then move on. I am listing my few immediate goals
- Toss Mechanism - I want to start with toss mechanism. Gone are those days where generating a random number will decide whether you bat or bowl first. I want to create a dataset where given condition, stadium, pitch type and tournament type, predict to either bat or bowl.
- For example, given World Cup Final, overcast condition, good batting surface at Lords... Team winning the toss would like to bat first and put pressure on the opponent
- Compared to a bilateral match, overcast condition, good batting surface at Lords.. Team winning might incline towards bowling first to utilize the overcast condition in bowling
- Simulating an over - After the toss mechanism is complete, I would want to extensively test simulating an over. A lot of factors to take into consideration with even a bigger dataset. More details at a later stage.
I want to know the release dates
For the toss mechanism, I would like to allow public access tentatively by 10th of May 2019. I have already started working on the dataset.
Once again, thank you always having my back despite my repeated failures. You learn a lot from your failures.
Looking forward to hear your thoughts.
Thanks
Parth