Stupendously Simple Scouting Statistic

[Kris Verdeyen]

Ok, does anyone remember the "Dr Joe Index" from last year's game? It determined what a team's specific contribution was to an alliance by how much it raised or lowered the average score of the alliance's team members. It was pretty cool, in my opinion. But then, I'm a baseball fan, and enjoy games with lots of numbers to keep track of (slugging percentage, anyone?).

So, without further ado, I think I've taken the first step towards a similar goal for this year's game. The essence of it is pretty simple - just take the number of QP's your alliance scored, and divide the number of match points it took you to get those points. Call it "scoring efficiency". If you lose the match, you have a scoring efficiency of 1. If you win the match by one point, you have an SE of almost 3. However, if you win by more than 3 times your opponent's score, your SE will drop below 1. I like it because it penalizes those who run up the score, but I dislike it because, well, because it penalizes those who run up the score too much. I mean, the loser of the match gets a 1, and the winner can get 0.5?

I guess it's a good measure of how well a team knows how to play the game, (by not overscoring), and it still needs to be compensated for the luck of the draw (ally and opponents), but it might have value.

What do you think?



(The attachment is a spreadsheet of most of the possible scores. I used the above formula, but I also took the logarithm and normalized the result, to see if it would help shape the data. The numbers in the sheet are WE's SE.)

[ColleenShaver]

First off, I think it's great, and no where near horrible, when other teams post tools that may be helpful for some of us on the forum.

Personally, I don't totally understand how the sheet works, but maybe the designer can explain exactly how it is used so it can be better understood.

As for scoring high to show off.... you can score a lot of points and 'show off' but not score them all for yourself.. for instance, say you were an awesome ball handler and you wanted to show off your skills.. you can score 20 balls for your team and 19 for your opponent.. this wouldn't risk your QPs or the 'scoring efficiency' that is spoke of here... while allowing you to strut your stuff for all those watching..

and i'd think your ability to do that would show off alot more if you just piled the points onto yourself.. it would show your team having a clear focus on the game and the ability to think on your toes to affect the score..

running up your match score doesn't do much for you or anyone else in this game.... run up the score on the field (the two alliance combined) and win.. the best way to show yourself..

i think so...

[Kris Verdeyen]

Sorry about the lameness of the spreadsheet -

It's not really meant as a tool so much as a display of what the SE would be in the event of specific match scores. The two teams in the match are 'WE' and 'THEY'. The number of match points scored by 'THEY' runs along the top of the sheet, and the number of points scored by 'WE' runs along the side. The cells in the middle contain basically a souped up version of the SE. It takes the Logarithm of the SE in order to make the distribution better, and then scales it so that the highest positive values lie somewhere around 1000 (which was just an arbitrary number).

The interesting thing that I was trying to show with the spreadsheet (and didn't) was that the SE is high when the team's scores are similar, and lower when they differ greatly.

I'm trying to come up with a way to scout teams and find the proverbial 'diamonds in the rough', and i figured that this was a start.

I'm posting an updated version of the spreadsheet that shows good SE's in green, average SE's in yellow, and bad SE's in red. This should make it easier to read

Simpler yet:


How many points did you contributed to your alliance:

-10 You are a turkey and let yourself get captured
0 You didn't help or hurt
10 At least you can stay put in the end zone
20 Wow Pretty darn good.
30 Apply for Woodie's Chair at MIT
40 Get a seat on the Board of DEKA
>40 You mean against good opposition and wih no help from you ally you can move and hold 2 goals in position and fill em with 21 balls? This I would like to see!

[Mike Bulzomi]

Check your excel formulas. Absolute cell references (if that is what you were doing) are in the format "$<Column>$<Row>" like "$B$3". If not, then make sure not to put dollar signs in where they're not needed. Sorry to point out the errors in your spreadsheet, but they run rampant and mess up all the data in column C.

[Kris Verdeyen]

The absolute references are there on purpose - If you use a single dollar sign in front of the column reference: $B2, then the column is absolute, but the row can change. It's very useful to have one reference absolute and one relative when making a chart of this nature. The reason that the column C data is screwed up is because the formula is dividing by zero in column C. The formula itself doesn't work.

Do you have any comment on the actual content - the idea behind the chart?

[kevinw]

I think the concept of scoring efficiency is useful in this year's game. However, I think a little more weight should be placed on actual qualifying score. For instance, a 10-9 victory has a scoring efficiency too close to 20-19 imho. Perhaps the efficiency is close, but then I guess I'd be wanting to measure a combination of efficiency and scoring ability. And consistency (regardless of partners and opponents), but that's a whole other topic.

[Bill Enslen]

I'm having difficulty understanding your logic in computing the scoring efficiency. My difficulty stems from not understanding your purpose in evaluating the SE of other teams. If you intend to use SE to determine which team to add to your alliance when you are one of the top 4 teams in the qualifying round, you're missing the boat. At that point, your team has demonstrated a consistent ability to score points. Your ideal partner is a team which has demonstrated a consistent ability to win matches, regardless of the number of points scored.