Welcome to the Hardcore Husky Forums. Folks who are well-known in Cyberland and not that dumb.
UW Doog Bot's Season Wins Probability Distribution Calculator Executive Summary
UW_Doog_Bot
Member, Swaye's Wigwam Posts: 15,977 
A. Objective
The purpose of this exercise was to poll a select set of “Hardcore Husky” fans on the probability of individual matchups of the Huskies 2018 opponents. This info would then be aggregated to develop a model to predict the most likely scenarios for the Huskies 2018 season.
B. Methodology
Users of the bored were polled on each matchup to determine, in their opinion, the likelihood of Washington’s football team winning the game against each individual opponent. The poll featured seven options of varying win likelihood. Responses were compiled and respectively assigned values approximately corresponding to normal curve values as seen in the following table.
Significant outliers were removed from the responses to account for respondents’ sense of humor and irony. Other, more standard weighting methodologies could be used but were beyond the scope of this exercise.
Values were aggregated to create an average win probability across the season using the standard binomial equation. Comparing results of this method to those of a more rigorous Poisson binomial distribution yielded non-significant differences in values for the purposes of this exercise. Using the standard equation significantly reduced the time necessary to calculate the results. Poisson discrete values may be of interest at the end of the season for comparison’s sake to improve the methodology of this exercise in the future.
From the standard binomial equation results a win distribution probability was developed. Further results were developed using the same methodology given the loss to Auburn, AU=0, and the likely win over North Dakota, If ND=1.
C. Results
D. Summary
The UW Huskies football team, even before a loss to Auburn, was highly unlikely to finish the season unbeaten. The most likely scenario for the Huskies with the given win probabilities was to finish 10-2 with most likely losses to Auburn and Utah. With the loss at Auburn the Huskies are now most likely to finish 9-3 overall and 7-2 in the conference with Pac 12 losses to Utah and Stanford being the most likely. Even before losing to Auburn the Huskies had a high probability of losing two or more PAC 12 games with the given probabilities if only looking at the conference games on the schedule.
The purpose of this exercise was to poll a select set of “Hardcore Husky” fans on the probability of individual matchups of the Huskies 2018 opponents. This info would then be aggregated to develop a model to predict the most likely scenarios for the Huskies 2018 season.
B. Methodology
Users of the bored were polled on each matchup to determine, in their opinion, the likelihood of Washington’s football team winning the game against each individual opponent. The poll featured seven options of varying win likelihood. Responses were compiled and respectively assigned values approximately corresponding to normal curve values as seen in the following table.
Significant outliers were removed from the responses to account for respondents’ sense of humor and irony. Other, more standard weighting methodologies could be used but were beyond the scope of this exercise.
Values were aggregated to create an average win probability across the season using the standard binomial equation. Comparing results of this method to those of a more rigorous Poisson binomial distribution yielded non-significant differences in values for the purposes of this exercise. Using the standard equation significantly reduced the time necessary to calculate the results. Poisson discrete values may be of interest at the end of the season for comparison’s sake to improve the methodology of this exercise in the future.
From the standard binomial equation results a win distribution probability was developed. Further results were developed using the same methodology given the loss to Auburn, AU=0, and the likely win over North Dakota, If ND=1.
C. Results
D. Summary
The UW Huskies football team, even before a loss to Auburn, was highly unlikely to finish the season unbeaten. The most likely scenario for the Huskies with the given win probabilities was to finish 10-2 with most likely losses to Auburn and Utah. With the loss at Auburn the Huskies are now most likely to finish 9-3 overall and 7-2 in the conference with Pac 12 losses to Utah and Stanford being the most likely. Even before losing to Auburn the Huskies had a high probability of losing two or more PAC 12 games with the given probabilities if only looking at the conference games on the schedule.
Tagged:
Comments
Utah is a TUFF game.
Oregon may take UCLAs place as the other most feared road game.