Friday, 21 February 2014

Scoring System: Scenario 2 and 3

A second (and third) look as proposals to altering the scoring system revolve around widening the bonus point window.
In season 2013, the rules were set such that;
- award 1 point if you tipped the correct team AND were closest on the margin.
- award an additional 2 points were awarded if you tipped the margin to be within a range of -3 to +3 of the actual margin,
- award no additional points for tipping the exact margin,
- award half the available points if both teams tipped correct and were equally distant from the margin.
The logic in choosing an inclusive -3 to +3 window was that is gave a 7 point target (3 over, 3 under and the exact tip) to land in to get the bonus points. Effectively just over 1 goal in a game.
The bonus points were expected to be hard to achieve, but in reality we had 64 occasions where they were collected - working out to be 3.5 per week.

Tipping under that system brought us the completed home/away season table as below.
Original 2013 table.
Using this benchmark, two other scenarios can now be compared against.

CASE 2: Its a goal!
In this case, lets modify the bonus window to include any tipped result -6 points to +6 points from the actual result (effectively at 13 point target to hit - i.e. 2 goals).

The season 2013 table under this scenario puts the teams as at right.

There is little change in the positioning of teams on the table with the boosted width window.
The top five is retained, with some shuffling of positions, and the same is true of the bottom half.

Also under this 1 goal window;
a) There is a marginal increase in match scores (increased from 2.83/game to 3.26/game).
b) There are slightly less drawn matches (20 draws down from 26).
c) The S.D. (Score difference) numbers show some irregularities, with Wal's remarkably high, and others with negative S.D.'s placed higher on the table than teams with positive S.D.'s. This indicates there were a few lopsided, high scoring matchups, possibly blown out by the bigger bonus window.
Worth noting, boosting the window from 7 to 15 points yields 115 games where the bonus is awarded (almost double the 64 of the 7pt window), which is 6.4 per game.

CASE 3: Two is as good as one.
Lets try widening the window even further. In this case, we now go to a '2 goals away from the actual result gets the bonus' calculation.
As before, this effectively opens the door for a tip to be out by -12 to +12 point spread (a 25 point target window).

Running season 2013 through that window yields the chart below.

Big winners from this are Amul82, who leaps from 8th to 4th, while Wal13Freo and TheHolyBoot suffer sliding from 1st to 5th, and 4th to 7th respectively.

Also under this wide window;
a) This system increases the match scores significantly (increased from 2.83/game to 4.18/game).
b) There are slightly less drawn matches (22 draws down from  26).
c) As a result, the S.D. (Score difference) numbers are about the same.
Worth also noting there was a more even spread of teams winning games (top teams Pts tally 44, down from 56), and the gap from top to bottom changed from 32 pts down to 24.

As before, boosting the window from 7 to 25 points yields 200 games where the bonus is awarded (more than triple the 64 of the 7pt window), which is 11.1 per game.

Opening the tipping bonus window brings higher scores per game, but not a great deal more. Going to a +/- 2 goal window (as per case 3) is clearly too far, while opening the window to +/- 1 goal (case 2) is more practical.
The case 2 scenario in actuality is window of 13 points (+6, exact and -6), whereas the incumbent system (-3, exact and +3) rewards those that are most accurate in their estimations (read as guesses, if you wish).

The FMI board are still favouring the incumbent system, modified to award 1 point for hitting the window and another for being exact.

Discussions and opinions welcome, either at the blog in the comments (for long form commentary) or on twitter.

No comments:

Post a Comment