Broken Bat - Help: Rating and Fan Mood

Forum >> Help >> Rating and Fan Mood

Post ID	Date & Time	Game Date

#73053	04/22/2020 3:58:57 pm	Apr 27th, 2045
Ken_Kennilworth Joined: 11/26/2019 Posts: 417 Charleston Hawks Legends	Re: The calculation of Fan Mood Way back on 6/11/2012 the Admin made this post about ratings and fan mood in this thread: Every team started with a 100 rating and has moved up or down from that based on wins and loses. I guess you can say 100 is the average. You can see the total rankings here. Rankings used to be used to determine cup inclusion, but now everyone is included. Fan mood is kind of the same thing. If you’re above 100, your fans are generally happy. Below 100, you fans are angry or losing interest. Steve I am led to believe by the above that Fan Mood changes are zero sum, meaning the team winning a game gains an amount equal to the reduction experienced by the team losing the game. If so, the average fan mood for all team should stay at 100. Is that correct, or are there other factors involved in calculating Fan Mood; factors I can influence? Because as of 4/21/2020 I believe the average Fan Mood for all 756 is not 100. No, it is 99.5 per my spreadsheet work. I copied the Rankings page containing all 756 teams into my spreadsheet. Then added up the Fan Mood column and calculated an average. At first I thought it was a copying mistake or just the result of rounding by BrokenBat. Well, I think I eliminated copying mistake by running a line by line cumulative total...I never got a line total that was less than or equal to the line above—thus every cell contained a valid fan mood number. And none of the Fan Mood numbers was in the “impossible range”. As for rounding, the report by BrokenBat gives Fan Mood with 2 decimal point precision. If Fan Mood is to have a 100 average, the total of Fan Mood scores for all teams should sum to 75,600. But per my spreadsheet, all Fan Moods summed to 75,234.04. That 365.96 points short. 75,234.04 Fan points when divided by 756 teams yields an average of 99.5159...not 100. That is too big a shortfall from 100 to be attributed to rounding when Fan Mood is reported to two decimal places—the most any rounding error could be would by .01 per team to 7.56 total Fan Mood points, not the 365.96 point difference noted above. So I can't say rounding causes the difference. Is the difference due to corrupted data? It seems plausible that sometime in the last 10 years the data may have gotten corrupted, and efforts to correct may not have brought the average Fan Mood back to 100 per team. So was that it, or are there other reasons the average is not 100? Updated Wednesday, April 22 2020 @ 4:02:01 pm PDT

#73059	04/22/2020 7:15:25 pm	Apr 29th, 2045
hurstdm Joined: 01/18/2017 Posts: 579 Murfreesboro Moo Cows IV.2	Fascinating number crunching. Why would the average have to be exactly 100.00? Why couldn't it just "approach" 100?

#73063	04/23/2020 6:18:06 am	Apr 29th, 2045
Ken_Kennilworth Joined: 11/26/2019 Posts: 417 Charleston Hawks Legends	approach 100 -- it would be good if it was. That would mean Fan Mood is not zero sum and implies regarding fan mood : being competitive in games while losing is less costly than being blown out; winning awards and championships earn a bonus; press releases and forum posts could help; there are penalties for ownership change, bankruptcy or star player cuts. As you might guess, I am being a BrokenBat detective here by trying to solve the mystery of Fan Mood -- the ultimate goal is to perform better in the game

#73080	04/23/2020 5:29:20 pm	May 3rd, 2045
Ken_Kennilworth Joined: 11/26/2019 Posts: 417 Charleston Hawks Legends	Here some more information about Ratings and how they work based on some crunching I did tonight. Ratings are zero-sum (Fan Mood is not, but that is for another thread). When two teams play, after the game is over, the losing team has its rating reduced by 1%, while an equal number of rating points is added to the winning teams rating. For example. Team A with a rating of 120 plays Team B that has a rating of 110. Team A loses. Team A has its rating reduced by 1.2 points (1% of 120) down to 118.8 Team B has its rating increase by 1.2 points up to 111.2 from 110. The net change for the two teams is zero-sum! So it's always the losing team giving points to the winning team. I discovered this my recording the Rating and Fan Mood points before and after each game played tonight in my division. I then calculated the change in rating points, and after a minute or two of puzzling, I saw the change described happened after all 10 games. I would urge others to try repeating my work to see if they get the same results and then add their findings to this thread--no need to test as many games as i, even just one game would help validate or refute this finding. Updated Friday, April 24 2020 @ 1:13:05 am PDT

#73139	04/25/2020 10:01:38 pm	May 10th, 2045
Ken_Kennilworth Joined: 11/26/2019 Posts: 417 Charleston Hawks Legends	Fan Mood, for league games, changes for wins and losses only it seems. It doesn't matter if you win by 1 run or 20, you get the same increase. And it doesn't matter if you beat the first place team in you league or the bottom dweller. (Thus, there is no accounting for a good loss, a close loss to a good team). The increase/decrease for league games depends only on your Fan Mood at the start of a game. If the mood is 100, you will gain 2 if you win, or lose 2 points if you lose...its the inflection point. At Fan Mood of 125 (25 points above 100) you will gain 1.5 points if you win...at Mood of 75 (25 points below 100) you will be deducted 1.5 points if you lose. If you lose at 125 points, you will have 2.5 points deducted, while if you win at 75 points, you will gain 2.5 points. Symmetry around 100. The farther above 100, the less gain a win gives, but the more deduction for a loss...the reverse is true for below 100. I learned this by downloading the Ratings page into a spreadsheet for three consecutive league dates and then analyzing the game to game changes in Fan Moods for every one of the 756 teams. This gave me over 1500 data pairs sets. From there it becomes very easy to see there is a linear relationship between Fan Mood increase and Fan Mood prior to increase, and there was no bonus points for beating a good team or penalty for losing to a poor one... So now I have a chart that can give me a good idea the increase/decrease to expect if you win/lose when you Fan Mood before a league game is anywhere from 20 up to 150. My chart is only for league games. I suspect CUP games result at least twice as much as league games...I will be looking at that next CUP play. Don't know yet if there is a difference for Group Cup play vs Elimination play. And of course, league championship games are probably even more impactful to fan mood (4 times as much?) For Fan Mood here is no ELO based system as I suspected. Updated Saturday, April 25 2020 @ 10:10:28 pm PDT

#73140	04/25/2020 11:42:47 pm	May 10th, 2045
Ken_Kennilworth Joined: 11/26/2019 Posts: 417 Charleston Hawks Legends	Here is a forumla that will calculate the change in your Fan Mood if you win a league game Change in Fan Mood = 4(200-Fan Mood Before Game)/200 If you lose a league game Change in Fan Mood = (-4 Fan Mood Before Game )/200 After calculating the change, update per following Fan Mood After = Fan Mood Before + Change in Fan Mood ; duh! Fan Mood ranges from a possible 0 to 200, that's why 200 in the formula...the absolute value of the sum of change when a win plus change of a loss is always 4...that why multiply by 4. This was per my review and analysis of the my downloaded data. The Rating page rounds to 2 decimals, the formula will go several decimals beyond but should give the same result as the page when the formula's figure is also rounded. So I believe I solved the mystery of Fan Mood changes for league games. Please try repeating my analysis and/or applying the formula to your games. And let us know if you find the analysis / formula works or not. Updated Sunday, April 26 2020 @ 6:00:07 am PDT

#73141	04/26/2020 1:15:22 am	May 10th, 2045
MukilteoMike Joined: 08/09/2014 Posts: 3294 Inactive	Nice work. I wonder why in the world Steve didn't use the same method for both rating and mood. If it's good for one, isn't it fine for the other? While this isn't a zero sum, it works out to very nearly the same.

#73143	04/26/2020 5:50:33 am	May 10th, 2045
Ken_Kennilworth Joined: 11/26/2019 Posts: 417 Charleston Hawks Legends	Thank you. Why Rankings in the game? I don't know either but it must have some purpose; why spend time and computer resource to develop something merely ornamental? My goal in all this work is to have a formula/method to predict attendance. All I know about that right now is fan mood (home and visitor?) and weather influence attendance, perhaps ratings are involved as well. But now with the formula I can work backwards from the present to calculate fan mood for past games for both my team and opponents. And the attendance figures already reported on the schedule page --so some correlation analysis/testing can begin. Additionally, we will learn the relative importance of CUP games...I will confess that in my first CUP season I played my subs for most games because I wanted to avoid injury to my regulars--looking back it probably was harmful to my attendance/finances.

#73144	04/26/2020 10:19:37 am	May 10th, 2045
Ken_Kennilworth Joined: 11/26/2019 Posts: 417 Charleston Hawks Legends	Sorry to take over this thread, but here is another piece of the Fan Mood puzzle. This time for CUP game pool play. Cup Pool Play games make 1.5 times the change that League play games do. Otherwise, all is the same. Thus, take the same league play formula and multiply the result by 1.5 Here is a forumla that will calculate the change in your Fan Mood if you win a CUP game Change in Fan Mood = 1.5 * (4(200-Fan Mood Before Game)/200) If you lose a Cup Pool Play* game Change in Fan Mood = 1.5 * ((-4 * Fan Mood Before Game )/200) again, if you try this calculation please share your results *updatefirst round of Cup Knockout Game uses the same forumla as Cup Pool Games...just confirmed by capturing before and after Fan Mood numbers for Allen v Huntington, recalculated change per formula and got a result that agreed with Fan Mood reported by Broken Bat Updated Friday, May 1 2020 @ 2:30:05 pm PDT*

#73147	04/26/2020 2:27:55 pm	May 10th, 2045
MukilteoMike Joined: 08/09/2014 Posts: 3294 Inactive	I predict that playoff games ramp up the multiplier to 2.5 or maybe even more. And that is terrible! It penalizes the playoff loser ridiculously. Winning the division should never hurt your mood, but it can destroy it. I stand by my method of correcting this travesty and am interested to see what the actual multiplier is. Again, great work, Ken.