Jump to content

[Suggestion] Improve transparency of star rating system


Foot
 Share

Recommended Posts

I think the silver/gold star system is a great way to abstract the ratings/reputation of various players and staff.

An issue that shows up for users is that star ratings change, and it is difficult to understand why that is:

- Example 1: Thread on reddit wondering why player ratings change after they are signed

- Example 2: Thread on SI Forums where a user wonders why youth intakes are always 'average', with input from SI Staff

- Example 3: Request on SI Forums for alternatives to the star system with input from SI Staff

Users intrinsically understand the meaning of a star system. Yet the examples above, as well as any other discourse on star ratings, show that user's struggle to understand why they change, what factors influence them, or how they should interpret stars in different contexts.

My understanding of the star system is that it assesses multiple factors, including the perception of the staff assessing players, to generate a single score on a scale (0-100) which is then represented as a star rating. If I have a player that is rated as 2 stars in year 1, then get a promotion, he might be rated as 0.5 stars in year 2, inspite of player and staff abilities being unchanged. This is due to the relative nature of the star system, that one year a player with a score of 53 might be 2 stars, and a score of 53 might be 0.5 stars the next year.

Suggestion: Provide transparency, or user-control, as to what '0 stars' and '5 stars' represent.

Transparency:

- I do not mean "show the entire calculation", as it isn't fun to see how the pie is made

- I do mean, if the general meaning of 0 star and/or 5 star changes, notify me as to why.

- For example for broad changes in how staff are applying the star scale, "Hey {manager}, since promotion our benchmark for a 0 star player has risen from {Guy who actively scores own goals} to {Guy who is okay, but I wouldn't put him on my bench}"

- Or for specific changes, "Hey {manager}, we just downgraded players {A}, {B}, {C} from 2 star to 1 star because you signed player {D} who made us realise that there are strikers who can use both feet"

User-control:

- The dynamic nature of the rating means that the user has to adjust to changes in the rating, which are not immediately obvious (are not announced, but rather require the user to notice that the scale has shifted)

- To counter-act this, the user could provide input to staff to say, "Hey, rate players as being 0 star if they'd be the worst in {my present league}, and 5 star if they'd be the best in {my present league}", or ,"Hey, rate players as being 0 star if they're the worst in {the world}, and 5 star if they'd be the best in {top league in my country}".

- This would enable players to control some of the variance, and adjust the scale based on their needs/aspirations, as well as feeling like they have more agency in how the ratings are being computed.

 

I think a combination of the suggestions for transparency and user-control here could go a long way to making the star rating into something that players can actively use, without necessarily giving too much transparency that players can start gaming the ratings (i.e. providing control to enable/disable particular inputs, so that they can backwards hidden attributes in game).

Link to post
Share on other sites

BTW, I'm the one from your example 3. I have recently thought this issue a lot. I can't say that I would know what is the best system, but it isn't this one we have now.

And, of course, the better the scout is, the better (or more accurate) results he/she would be able to give us. Also faster.

 

My results. So far.

1) First one question. There are that view with "Scouted" players. Should those results be comparable or not? Should each position be reviewed on its own, or all players be reviewed together (on league level basis)? Or both?

       - If each position is reviewed on its own, then every time the best of any position would get 5 stars aso.

       - If reviewed all together, then maybe none of position X would get 4 or 5 stars if there aren't any good ones in the league...

2) The easiest way to get the best star ratings scale for any view is to remember what was the assignment that you gave to your scout. And, yes, this assignment has to be mentioned in the view.

    - If you ask anything like "Who is the best ... in our team?", then the worst of your players gets 0,5 stars and best 5 stars. This would give the most variation.     Most important, "in our team" is mentioned in that assignment.

     - Using this way we could order our scout to find better player that we have now, but that doesn't give any idea for us, what is his level in our league. Like "fits first team". Signing a new player WOULD change the whole scale, if the new player is worse/better than anyone else in our team.

3) This is the one for me. Scouting is on league level basis, and that league level can be changed (at least -1...+1). It is mentioned in the view, that assignment is to find player "for our current league level."

    - If assignment is anything like "...for this league", then scout would have to do extra research, which would take more time.

    - 0,5 stars would mean that player isn't good enough to this league level, and 5 stars would mean, that player belongs to the next level. Three stars would be "the average player".

    - This could happen - in a promoted team we have only 1,5 stars playmaker, we have lost a few games on a row, and then manager gives an assignment "get at least average playmaker in this league" would mean finding at least three stars playmaker.

               - Here you could also take a view where only playmakers are chosen, and from there you would see the CA levels of the other teams' playmakers.

   - When during the spring you realise that you're getting promoted or being relegated you could start finding new players for the next season.

   - When you have won a place in one of UEFA leagues (or...), there could also be an option to have those opponent teams' players in comparison.

             - You might want to find a Striker that could have any chance against those defenders.

 

BTW, I voted for this. And if anyone who reads this, thinks that something needs to be done, then VOTE. That is the best way of getting something to be done.

Link to post
Share on other sites

  • 2 weeks later...

Thanks for the good response and for the support, rristola. All very good points. I appreciate it.

Reading your post I think one 'easy' solution that jumped to mind was giving star ratings a tooltip that would say something like "Player A is rated 2.5 stars by SCOUT D among all players in our league" or "Player A is rated 5 stars by Assistant Manager among CB players in our senior squad".

 

Link to post
Share on other sites

  • 3 months later...
Guest
This topic is now closed to further replies.
 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...