Talk Proposal - Fractalized Tennis Bets

Hello, we had submitted the following brief to PyMCon a few months back. We have made some progress since then with the WIP app. We released a teaser blog with some video walk throughs here..

We would appreciate the opportunity to present where we have reached so far and get feedback on what our roadmap should look like.

Short biography (to post on the series website) *

Founder of www.fractal-data.com, a collaborative workspace for curious minds to use technology to build intuition about real world decision making problems.

What is the title of your proposed event? *

Risk Management in Sports Betting - Betting on Tennis using PyMC

Please describe the format of your proposed event *

Online webinar, we will provide a demonstration of a WIP Streamlit App which we are building for the above use case. We intend to use this use case as part of our training program to onboard fresh collaborators.

2 Likes

I couldn’t find the video but read over this.

This seems interesting for a couple of reasons potentially

  1. You mention betting so there must be some sort of conversion into currency which is always great
  2. Showing how to a bayesian streamlit app works and why its useful

As far as the analysis details I haven’t seen much, so perhaps its good but I dont know!

For PyMCon I suggest this being a two part event. One being an async event where folks get to try a webapp, and then a seminar like you mentioned that provides more details on the implementation of the model and infrastructure.

Even outside of PyMCon it would be awesome if you put this together. Would be a great contribution to the online corpus of knowledge!

Thanks @RavinKumar the videos are linked to from the blog.

Blog: Tennis Enthusiasts – Do you possess an edge over betting markets? – Fractal Dataminds
First Video: Loom | Free Screen & Video Recording Software | Loom (parts 2 and 3 are embedded in there)

At this stage its just a preview and the real analysis will come through from v0.2 onwards. A 2-part event sounds like a good idea. Main question is at what stage of completeness should we plan to present. I’m guessing we should only do so with a working end to end workflow, no point going in without that.

Regarding outside PyMCon - yup we do want to ungate at least some parts and make it available to the broader community.

Hi @RavinKumar, as asked by @Gireesh_Ramji I would be taking forward this discussion along with him. Any further thoughts on the proposal after looking at the videos, and the analysis related discussion in the last comment?