Oracle Sporting Performance Index: Difference between revisions

From RabastorWiki
Jump to navigation Jump to search
 
(One intermediate revision by the same user not shown)
Line 6: Line 6:


==History==
==History==
OSPI was developed by Oracle as a means to allow comparative ranking of teams in different conferences, for their NSCF coverage. The idea of non-conference games had been floated during [[NationStates College Football season 4|season 4]], and [[Marcus Resheph]], Oracle's head of data, tasked his analysts to devise a method to allow these games to contribute toward a fair ranking system.
OSPI was developed by Oracle as a means to allow comparative ranking of teams in different conferences, for their NSCF coverage. The idea of non-conference games had been floated during [[NationStates College Football season 4|season 4]], and [[Marcus Resheph]], the new Head of Data instructed the team to adapt a previous system devised by James Greenwall – the original Oraacle – to create a method to allow these games to contribute toward a fair ranking system.


Drawing from the system of [[wikipedia:RPI|RPI]] in popular [[RLStates]] games and media, the team deconstructed the formula and applied it to NSCF competition for testing.
Drawing from Greenwall's formula and the system of [[wikipedia:RPI|RPI]] in popular [[RLStates]] games and media, the team deconstructed the logic and applied it to NSCF competition for testing. After some trials and tweaking, a completed version was put into use on the OSN website, and later presented to the NSCF Committee. The Committee approved, and the NSCF Commissioner implemented the system for official NSCF rankings as of season 5, when non-conference games were also introduced for the first time.
 
After some trials and tweaking, a completed version was put into use on the OSN website, and later presented to the NSCF Committee. The Committee approved, and the NSCF Commissioner implemented the system for official NSCF rankings as of season 5, when non-conference games were also introduced for the first time.


==Formula==
==Formula==
Line 61: Line 59:
| '''Utica'''
| '''Utica'''
|-
|-
| Turic
| '''Turic'''
| 21
| 21
| 16
| 16
Line 71: Line 69:
'''WP'''
'''WP'''


:Utica: ((2*0.4)+(1*0.6))/3 = (0.8+0.6)/3 = 1.4/3 = 0.4667
:Utica: ((2*0.6)+(1*1.4))/3 = (1.2+1.4)/3 = 2.6/3 = 0.8667
:Burningham: ((1*0.4)+(0*0.6))/3 = (0.4+0.0)/3 = 0.4/3 = 0.1333
:Burningham: ((1*0.6)+(0*1.4))/3 = (0.6+0.0)/3 = 0.4/3 = 0.2000
:Mora: ((0*0.4)+(1*0.6))/3 = (0.0+0.6)/3 = 0.6/3 = 0.2000
:Mora: ((0*0.6)+(1*1.4))/3 = (0.0+1.4)/3 = 1.4/3 = 0.4667
:Turic: ((1*0.4)+(0*0.6))/3 = (0.4+0.0)/3 = 0.4/3 = 0.1333
:Turic: ((1*0.6)+(0*1.4))/3 = (0.6+0.0)/3 = 0.6/3 = 0.2000


'''OWP'''
'''OWP'''


:Utica: (0.1333 + 0.2000 + 0.1333)/3 = 0.4666/3 = 0.1555
:Utica: (0.2000 + 0.4667 + 0.2000)/3 = 0.8667/3 = 0.2889
:Burningham: (0.4667 + 0.2000 + 0.1333)/3 = 0.8000/3 = 0.2667
:Burningham: (0.8667 + 0.4667 + 0.2000)/3 = 1.5334/3 = 0.5111
:Mora: (0.4667 + 0.1333 + 0.1333)/3 = 0.7333/3 = 0.2444
:Mora: (0.8667 + 0.2000 + 0.2000)/3 = 1.2667/3 = 0.4222
:Turic: (0.4667 + 0.1333 + 0.2000)/3 = 0.8000/3 = 0.2667
:Turic: (0.8667 + 0.2000 + 0.4667)/3 = 1.5334/3 = 0.5111


'''OSPI'''
'''OSPI'''


:Utica: ((2*0.4667)+0.1555)/3 = (0.9334+0.1555)/3 = 0.3630
:Utica: ((2*0.8667)+0.2889)/3 = (0.9334+0.1555)/3 = 0.6741
:Burningham: ((2*0.1333)+0.2667)/3 = (0.2666+0.2667)/3 = 0.1778
:Burningham: ((2*0.2000)+0.5111)/3 = (0.2666+0.2667)/3 = 0.3037
:Mora: ((2*0.2000)+0.2444)/3 = (0.4000+0.2444)/3 = 0.2148
:Mora: ((2*0.4667)+0.4222)/3 = (0.4000+0.2444)/3 = 0.4519
:Turic: ((2*0.1333)+0.2667)/3 = (0.2666+0.2667)/3 = 0.1778
:Turic: ((2*0.2000)+0.5111)/3 = (0.2666+0.2667)/3 = 0.3037


Final rankings would therefore be:
Final rankings would therefore be:

Latest revision as of 08:33, 15 November 2016

The Oracle Sporting Performance Index, commonly known as OSPI, is a sports rating system developed by the Oracle data analysis branch of Osarius Sports Network. It is used to rank sports teams based upon a team's wins and losses and its strength of schedule. OSPI is the primary method by which NSCF teams have been ranked and seeded since season 5. The system has also been in use in various Osarian collegiate sports since 2204 to aid in the selecting and seeding of teams appearing in the playoffs. In its current formulation, the index comprises a team's winning percentage, its opponents' winning percentage, with adjustments made for home and away games.

OSPI could be said to lack theoretical justification from a statistical standpoint as it does not take into consideration the margin of victory. However, because the margin of victory has been manipulated in the past by teams or individuals in the context of gambling, OSPI can be used to mitigate motivation for such manipulation.

As the winning percentage of the opposition counts for one third of the OSPI value, it is not always prudent to schedule weaker opposition. To mitigate the potential for tactical scheduling in NSCF, for example, the league requires teams to schedule non-conference games at the start of the season.

History

OSPI was developed by Oracle as a means to allow comparative ranking of teams in different conferences, for their NSCF coverage. The idea of non-conference games had been floated during season 4, and Marcus Resheph, the new Head of Data instructed the team to adapt a previous system devised by James Greenwall – the original Oraacle – to create a method to allow these games to contribute toward a fair ranking system.

Drawing from Greenwall's formula and the system of RPI in popular RLStates games and media, the team deconstructed the logic and applied it to NSCF competition for testing. After some trials and tweaking, a completed version was put into use on the OSN website, and later presented to the NSCF Committee. The Committee approved, and the NSCF Commissioner implemented the system for official NSCF rankings as of season 5, when non-conference games were also introduced for the first time.

Formula

The current formula is as follows:

OSPI = ((WP * 2) + OWP) / 3

where WP is Winning Percentage and OWP is Opponents' Winning Percentage.

The WP is calculated by taking a team's home wins multiplied by 0.6, plus away wins multiplied by 1.4, and dividing this total by the number of games it has played (i.e. wins plus losses), like so:

WP = ((HomeWins * 0.6) + (AwayWins * 1.4)) / Games Played

Games won at a neutral site are counted as is (effectively with a multiplier of 1.0).

The OWP is calculated by taking the average of the winning percentages for the opponent in each game the team has played. This means an opponent a team faces twice would have its WP counted twice in the OWP calculation.

Extended example

Assume the following game results:

Home Score Away
Utica 27 10 Burningham
Turic 17 42 Mora
Utica 31 13 Turic
Burningham 17 10 Mora
Mora 33 35 Utica
Turic 21 16 Burningham

Here is the calculation of the WP, OWP, and OSPI for each team:

WP

Utica: ((2*0.6)+(1*1.4))/3 = (1.2+1.4)/3 = 2.6/3 = 0.8667
Burningham: ((1*0.6)+(0*1.4))/3 = (0.6+0.0)/3 = 0.4/3 = 0.2000
Mora: ((0*0.6)+(1*1.4))/3 = (0.0+1.4)/3 = 1.4/3 = 0.4667
Turic: ((1*0.6)+(0*1.4))/3 = (0.6+0.0)/3 = 0.6/3 = 0.2000

OWP

Utica: (0.2000 + 0.4667 + 0.2000)/3 = 0.8667/3 = 0.2889
Burningham: (0.8667 + 0.4667 + 0.2000)/3 = 1.5334/3 = 0.5111
Mora: (0.8667 + 0.2000 + 0.2000)/3 = 1.2667/3 = 0.4222
Turic: (0.8667 + 0.2000 + 0.4667)/3 = 1.5334/3 = 0.5111

OSPI

Utica: ((2*0.8667)+0.2889)/3 = (0.9334+0.1555)/3 = 0.6741
Burningham: ((2*0.2000)+0.5111)/3 = (0.2666+0.2667)/3 = 0.3037
Mora: ((2*0.4667)+0.4222)/3 = (0.4000+0.2444)/3 = 0.4519
Turic: ((2*0.2000)+0.5111)/3 = (0.2666+0.2667)/3 = 0.3037

Final rankings would therefore be:

    • Utica
    • Mora
    • Turic (via H2H results)
    • Burningham

Other uses

When used in sports which allow draws, OSPI can be modified to give similar functionality by counting draws as half a win.

This adjusts the winning percentage formula to the following:

WP = ((HomeWins * 0.6) + (AwayWins * 1.4) + (HomeDraws * 0.3) + (AwayDraws * 0.7)) / Games Played

As before, neutral games are included in the calculation with a modifier of 1.0 (or 0.5 for draws).

References

Lua error: Cannot create process: proc_open(/home4/iltornan/lua/error.log): Failed to open stream: No such file or directory