Hello here is my problem.

I am analizying some tennis matches, and I am able to see the highest odd (bookmaker's odd) reached by that tennis player in N-matches (N is number of matches)

Then I take all the highest odds reached by that tennis player during his N-matches , and I extrapolate the simple mean, to get the ''average mean'' of the highest odd reached by that tennis player during N-matches.

Ok where is the problem?

Not all the the opponents have the same value: let's suppose tennis A is a good player. On Monday he will have a match against tennis player B, maximum odd reached by tennis A is 2

On thurstday Tennis player A plays against tennis Tennis player C, tennis player A reaches a maximum odd of 1,40 during his match, BUT tennis player C is more talented than tennis player B, so the macth against tennis player C has been more difficult for tennis player A.

Here is my problem; I want to build up a weighted mean, so I need to add some ''quantity'', some ''value'' to each addend in order to have a weighted mean

This ''weight'' I want it to be the different difficulty of playing against one tennis player, rather than oneother. So, in our example, tennis Player C will have ''more weight'' in our weighted average than Tennis player B, BUT HOW MUCH? I was thinking to build up my personal classific from 1 to 8 and assign to every match a ''weight'' in order to calculated the final average weighted mean of the maximum odd reached from the A tennis player, but how much ''weight'' to give to one tennis player, rather than another one... how to calculate this? Based on what parameter?

A personal subjective parameter?

Is this a personal subjective model?

Can I run a regression model for various tennis player in order to get the ''weighted'' value (from 1 to 8), but it will always be a ''subjective'' parameter?

Can you give me a way to STANDARDIZE the ''value'' of each tennis player compared to one other? Let's suppose I want to base my model on the ranking position so

B has 34 ranking,

C has 10 ranking positin

I have to standardize their value (to put in the weigheted mean). Well the standard values range from 1 to 8.

I am analizying some tennis matches, and I am able to see the highest odd (bookmaker's odd) reached by that tennis player in N-matches (N is number of matches)

Then I take all the highest odds reached by that tennis player during his N-matches , and I extrapolate the simple mean, to get the ''average mean'' of the highest odd reached by that tennis player during N-matches.

Ok where is the problem?

Not all the the opponents have the same value: let's suppose tennis A is a good player. On Monday he will have a match against tennis player B, maximum odd reached by tennis A is 2

On thurstday Tennis player A plays against tennis Tennis player C, tennis player A reaches a maximum odd of 1,40 during his match, BUT tennis player C is more talented than tennis player B, so the macth against tennis player C has been more difficult for tennis player A.

Here is my problem; I want to build up a weighted mean, so I need to add some ''quantity'', some ''value'' to each addend in order to have a weighted mean

This ''weight'' I want it to be the different difficulty of playing against one tennis player, rather than oneother. So, in our example, tennis Player C will have ''more weight'' in our weighted average than Tennis player B, BUT HOW MUCH? I was thinking to build up my personal classific from 1 to 8 and assign to every match a ''weight'' in order to calculated the final average weighted mean of the maximum odd reached from the A tennis player, but how much ''weight'' to give to one tennis player, rather than another one... how to calculate this? Based on what parameter?

A personal subjective parameter?

Is this a personal subjective model?

Can I run a regression model for various tennis player in order to get the ''weighted'' value (from 1 to 8), but it will always be a ''subjective'' parameter?

Can you give me a way to STANDARDIZE the ''value'' of each tennis player compared to one other? Let's suppose I want to base my model on the ranking position so

B has 34 ranking,

C has 10 ranking positin

I have to standardize their value (to put in the weigheted mean). Well the standard values range from 1 to 8.

Last edited: