Research paper + PPT

*Sateesh 11395519, you requested help with Content Development:   

The case that you’re analyzing is not clear. While you wrote about applying models to NFL power rankings, you did not describe the case itself, so readers may not easily understand your analysis. For example, a case study can focus on a particular workplace conflict. The paper will not discuss all types of workplace conflicts; instead, it will focus on one specific incident. For instance, an office has a new supervisor, and because of the new supervisor’s leadership style, the employees want to quit. That case will be described. Afterwards, the case study will then explain the problems with the new supervisor’s leadership style and recommend how the problems can be solved. You can add a paragraph that describes the case. First, are you examining real NFL rankings, or are you examining fantasy football rankings? Are you focusing on a particular season? If so, which season are you analyzing? What is the main issue with regards to NFL power rankings, or what is it about the rankings that you need to find out or determine? Describe the case first before you analyze it.

*Sateesh 11395519, you requested help with Introduction/Conclusion:   

Your first paragraph identifies different models, but your claims about them may not be accurate, Sateesh. The first sentence specifies the Bradley-Terry model. The next two sentences then seem to refer to other models, like the following:

Also, online ranking would help in getting models for performance of each team.

The sentence suggests that you’re going to use online ranking to find models to analyze each team’s performance. Readers can then get confused. How many models are you really using? You need to check your sources again to make sure you’re not using the term “model” inaccurately. For example, in one body paragraph, you mentioned “ESPN model,” but you may need the term “ESPN Football Power Index (FPI)” instead. Are you really looking for models, or are you getting your data from the ESPN FPI and then using the Bradley-Terry model to analyze that data? Make sure your terms are correct and that your introduction/first paragraph really reflects what you did in your case study.

Use of Resources 

In your body paragraphs, some of the examples are too broad, like the following:

For example, in this experiment, it is possible through classification of different teams (Tierney & Luke, 1997). There are many ways teams can be classified. How are the NFL teams in your particular case classified? You can specify the way the teams will be classified. You can then explain the example further. Why does that particular classification enhance openness? Look out for other similar issues in the body of your paper, Sateesh.

Summary of Next Steps: 

· Describe the case.  

· Make sure the claims in the first paragraph are correct.  

· Make the examples more specific.

Thank you for submitting your case study for a review, Sateesh. I enjoyed helping you with this step in the revision process. Have a good day! ~Katrina D.

You can find more information about writing, grammar, and usage in the Smarthinking Writer's Handbook.

_________________________________________________________________________________

Please look for comments [in bold and in brackets] in your essay below. Thank you for submitting your work to Smarthinking! We hope to see you again soon.

NFL Power Rankings through Bradley-Terry Modeling

Sateesh Matturi

University of Potomac

Data Analytics

12-1-2019

NFL Power Rankings through Bradley-Terry Modeling

Introduction

Use of games played in a week gives details of 208 games that would be used for Bradley-Terry model for entire results. Use of different model such as application of win/loss would help in identification of different ranks of teams from the total teams. Also, online ranking would help in getting models for performance of each team. However, some teams would have to apply play-offs.

Use of models such as Bradley-Terry models comes before application of modelling. Bradley-Terry models would fit best whereby availability of repeated points of a game are evident. For instance, Bradley-Terry model would best fit in areas with NFL (Tierney & Luke, 1997). The grouping s further subdivided into A and B to avoid conflict6 of number when handling games for the 14 weeks. After grouping, the aspects of Bradley-Terry modeling methods are applied. An assumption is made while working with this model, “every individual represented is a non-empty set, and individuals represented in second set beats individuals in the first set at least once.” After careful consideration, checking whether all teams are featuring in 14 weeks makes the experiment examinable (Hunter, 2004).

When using Bradley-Terry model, assumptions work to ensure that all partitions in any experiment are put into consideration. Enhancing openness is enhanced through various means. For example, in this experiment, it is possible through classification of different teams (Tierney & Luke, 1997). In addition, in order to put all players in consideration, the model checks to add all other key players. For example, all players who had participated in games are put into consideration such as key winners of the game from preceding years. With this form of grouping, it is expected that all individual groupings have satisfactory answer for issues of enhancing modeling using Bradley-Terry model.

Bradley-Terry model has some raw data which is required for experimentation. [Your sentence claims that the model itself has raw data, but a model usually analyzes data. You can check your sources again. Where does raw data come from? What does the Bradley-Terry model with that raw data? Make sure your claim is correct.] For example, understanding which teams have to be put into consideration enhances clear data classification. When raw data is grouped together, there is an ease of understanding NFL (Tierney & Luke, 1997). For example, a column of raw data shows that there is availability of the data but has not been considered for final results. In our experimentation, classification of data is carried out with all information considered.

Use of power ranking comparison in analysis of data makes all work possible. For example, in our experiment, it is possible to have all data made useful (Tierney & Luke, 1997). If data is classified into good teams, highly scoring teams, losing teams and winning teams. In our experiment, it is evident that all teams are considered with winning teams shows and losing teams made available in classification. Maintaining ways of data analysis helps in understanding all areas of classification. Further, when all data is made available in experiment, there is classification according to Bradley-Terry model classification methods using the following method:

P(i>j) = Pi/Pi+Pj

With Pi representing a positive real number which is assigned to an individual i.

(i>j) shows a preference of one entity to another. For example, it can be read as entity I preferred to entity j.

Information on performance of different teams is evident when ESPN and EPL models are used. For example, it is easy to have individual information using the above formula. When such information is substituted in the formula, it becomes easy to determine each individual team performance. Further, Bradley-Terry model is well represented in ESPN and NFL models (Tierney & Luke, 1997). However, there is no any form of order that is preferred to the other. Information distribution is made possible when all data is acquired through various ways but when it comes to substitution, it is easily done for all games with equity. However, some minor aspects of team’s performance may not be captured in this data modeling. For example, it may not be easy to have all team’s key players put into consideration. Much information is left out since no definite way of substituting it (Hunter, 2004).

Also use of play-off method in data analysis helps in analyzing data at a different measure. Some aspects of data may not be easy to put into consideration. For example, it may not be necessary to use an eye test to have all data handled and considered. For instance, eying on play-off simulations would be done better through Monte Carlo methods. [What are Monte Carlo methods? You can briefly explain or define them to provide readers with background information and at the same time show your understanding of the methods.] When using Monte Carlo methods of simulations, it is necessary to have a consideration to Bradley-Terry models (Hunter, 2004). When data is available for many teams, it becomes easy to have a different estimation to data. For example, when all other data is well simulated, it becomes easy to have a remaining teams substituted. A win can be predicted using a playoff of data.

Bradley method is one of the best methods of factoring data. When all information is put together, it figures out most important parts of data and leaves the rest. For example, in example given above, raw data is put by itself while other data is left for factoring (Hunter, 2004). Information with Bradley-Terry model is clear on where each individual data belongs. With such information about analysis of data, it becomes easy to figure out all aspects of data analysis and work out to all data irrespective of its usefulness. Data factorization is an important aspect of data analysis whereby information must be taken into account as t is from original source. Bradley-Terry model qualifies its usefulness in different aspects of data analysis since most areas of data analysis methods require proper analysis of data.

Conclusion

In conclusion, results obtained from this kind of model are satisfactory. With consideration of different sources of data, the methods acquires information from different areas. [The verb “acquires” is singular, but its subject “methods” is plural. A subject and verb should agree in number. A plural subject needs a plural verb, so what is the plural form of “acquires”?] For example, information may be from NFL which makes available data to be from a different classification. Also, Bradley-Terry methods has put other methods of data analysis into account. Some aspects considered are play-off methods of data analysis. However, the method is a good method for refining data in terms of predictions for future fate of information systems. Some advantages of data analysis may have a wide view of information management.

References

Hunter, D. R. (2004). “MM Algorithms for Generalized Bradley-Terry Models”. The Annuals of Statistics, Vol. 32, No. 1, 384-406. http://sites.stat.psu.edu/~dhunter/papers/bt.pdf

Tierney, Luke. (1997). “Fitting a Bradley-Terry model”.http://homepage.divms.uiowa.edu/~luke/xls/glim/glim/node8.html http://www.pro-football-reference.com/