Evaluation of the diference between 15-20-30 mn matches

Please discuss issues concerning the Ducati League here. This is the liaison between league players and the league council.
Post Reply
User avatar
tokimi
Private First Class
Private First Class
Posts: 120
Joined: Sat Nov 06, 2004 7:45 am
Location: Storrs CT
Contact:

Evaluation of the diference between 15-20-30 mn matches

Post by tokimi » Fri Jun 02, 2006 6:22 am

Not being a very good coder my self, I never the less finaly got courious
about what are the REAL effects of changing match times. I am sure those
of us who try to be scientists, or more generally those of use who are
courious about our world, would be vaguely interested in this.

Questions:

Do the teams that win at 15 minutes still win at 30 min?
corollary:
Do match results change over (ie if red is winning at the 15 min mark, does
that team still win the match) statisticly often?

What is the change in ratio between the cap scores of the two teams between
the 15, 20 and final time period?

Do the raw magnitudes of each team's scores increase dramatically between
the 15, 20 and 30 minute marks?

What is the rate of change in each team's score in the first 15 minutes of
play, vs the last 15 minutes of play? Also, depending on which team is
winning at the 15 minute mark, how does that effect the activity in
captures in the last 15 minutes?

Also, to be complete, what would a statisticly meaningfull number of match samples be.

Method
Again, I am not a coder, but I would imagine some of these things
following are feasible.

Sampling the cap score every 1 minute, mark only which team has what
score (actual team names are irrevelent). A plugin maybe, added to proxs
dub 59998 and dub 59999 should cover almost all the bases in ducati (as
I am only interested in dealing with ducati at the moment).

After a good statistical number of matches have been recorded, the
analysis would be fairly trivial. I would of course volunteer to do the
analysis, with public posting of data, methods and results as well.

Needs
A plugin would of course need to be built, and implimented in the 3 servers
mentioned above (proxs, dub 59998 and dub 59999). After that, it will
be a waiting game untill enough matches are recorded. Either a nightly
cron job mailing the data to me (or whom ever does the collection) or such
also would be needed.

Any one interested in working with me directly on this (yes I am coping
out and asking for free coders here), please just pm me here (I will get
it in my email) and we can work out a collaboration.

OF COURSE, helpfull suggetions, technical and psychological (yes, I am
also ignoring most psycological effects here), will be needed.
The very young do not always do as they are told, and the very old never listen any way.

wizard
Private First Class
Private First Class
Posts: 63
Joined: Sun Jul 25, 2004 9:02 pm

Post by wizard » Fri Jun 02, 2006 5:08 pm

There are very few matches we lost in the last minutes.

But there is a very large number of matches were we were ahead after 5 min and still lost high.

These are the cases I remember because they were frustrating and I assume the reverse is also true.

So from my experience I would say that there are very few changes to the outcome of a match after 15 min.

User avatar
tokimi
Private First Class
Private First Class
Posts: 120
Joined: Sat Nov 06, 2004 7:45 am
Location: Storrs CT
Contact:

Post by tokimi » Fri Jun 02, 2006 5:10 pm

Well, I have mixed experiences my self, which is why I think a strait up
analysis of the hard facts is important. Because experience can misslead
us if not used properly
The very young do not always do as they are told, and the very old never listen any way.

User avatar
goblin
Private First Class
Private First Class
Posts: 63
Joined: Sat May 06, 2006 10:01 pm
Location: Right here...

Re: Evaluation of the diference between 15-20-30 mn matches

Post by goblin » Fri Jun 02, 2006 6:59 pm

tokimi wrote: Method
Again, I am not a coder, but I would imagine some of these things
following are feasible.

Sampling the cap score every 1 minute, mark only which team has what
score (actual team names are irrevelent). A plugin maybe, added to proxs
dub 59998 and dub 59999 should cover almost all the bases in ducati (as
I am only interested in dealing with ducati at the moment).

After a good statistical number of matches have been recorded, the
analysis would be fairly trivial. I would of course volunteer to do the
analysis, with public posting of data, methods and results as well.
Let me just point out that no matter what the statistics show, we will never know if we got the right answer. The psychological effect of knowing the match will end soon can not be measured in this experiment - would a player in a 30 minute match set in a "final spurt" when the clock approaches 15 minutes?
Tank you!

User avatar
tokimi
Private First Class
Private First Class
Posts: 120
Joined: Sat Nov 06, 2004 7:45 am
Location: Storrs CT
Contact:

Post by tokimi » Fri Jun 02, 2006 9:09 pm

exactly goblin, but it will still give an idea of how matches progress. Atleast
an idea as to what we are dealing with when changing match times. Better
some idea than none.
The very young do not always do as they are told, and the very old never listen any way.

User avatar
goblin
Private First Class
Private First Class
Posts: 63
Joined: Sat May 06, 2006 10:01 pm
Location: Right here...

Post by goblin » Sat Jun 03, 2006 9:45 pm

tokimi wrote:Better some idea than none.
Very true. I just wanted to make sure that we weren't telling ourselves that we could find "the truth" about optimal match times with these statistics. "Some idea", we can agree on :-)
Tank you!

Post Reply