ETR is looking for a mathematician with a passion for fantasy golf and/or golf betting to help with a six-month project. After this project, there is potential for an ongoing part-time or full-time role with the company depending on skills and availability. The role is fully remote. We expect the workload to be significant during this six-month period (at least 20 hours per week, on average), but there is no specific expectation for what time of day you need to be available. We can work around prior commitments for exceptional candidates, should they exist. Compensation is TBD depending on a candidate’s skill and availability.
This role will be focused on planning and executing improvements to our golf projection system. At a high level, this means predicting how many fantasy points each golfer will score in a given tournament. More specifically, this could mean predicting a distribution of outcomes for every golfer on each hole of a particular course/tournament. For example, on hole 1 at Augusta, how often does Patrick Rodgers shoot a 1, 2, 3, 4, 5, etc.
While ETR will provide support, candidates will be expected to execute a significant portion of this project on their own. Candidates must be self-motivated and have significant experience in coding, data analysis, and predictive modeling. Preference will be given to candidates who have:
- Extensive coding experience (preferably in Python and/or R)
- Experience collecting data from scratch (e.g., scraping from web pages, gathering from APIs). Even better if this is golf data.
- Experience organizing/aggregating data
- Concrete examples of previous projects they have completed
- Experience working in the sports domain
To apply, please send your résumé and a comprehensive answer (including code) to at least one of the questions below to [email protected]:
- Simulate the distribution of scores for the average professional golfer (or a specific golfer of your choice) on a Par 4 hole (i.e., they complete the hole in two strokes 1% of the time, three strokes 20% of the time, etc.). Preferably, this is done using flexible/dynamic values for any input parameters you deem appropriate (e.g., hole distance, golfer skill).
- Related to question No. 1 directly above, list the input parameters you think are key to accurately simulating a golfer’s distribution of scores on a given hole. Describe why you think they are important. Provide code that derives specific estimates of at least one of these parameters (e.g., calculate golfer skill ratings for every golfer in The Masters, calculate a hole difficulty rating for every hole at Augusta).
- Describe how you would predict/simulate a distribution for scoring outcomes for each golfer in The Masters without simulating their detailed hole-by-hole scores. It is sufficient to describe the steps you would take (rather than implementing all of the actual code), but your answer should include some code.
