Real Week 2 update

[Cory]

Originally Posted by Boltman

As a follow up to my scheduling observations from CVR here are the key matches where "top teams" met on the same field of play as partners or as opponents. Of course there were many good teams but these tend to be the best by historic achievements. 179 peaked my interest to look a bit deeper. My definition of top teams are those I noted going in based on achievement. Forgive me if I missed someone.

as of now, every team # in picture identified is less than number 1679

What I did was only extracted every game in which top teams crossed paths or partnered up and the results are seemingly not random at first glance of the same field games.

Three matches the top teams faced against each other BUT there were SIX matches where they were also alliance partners

I did our matches and got TWO such pairings whereas 5136 had more (they also made WC) and we are within one Team# of them.

So I think the algorithm may weight previous WC pairings more favorably than head to heads. Again I only looked at games where giants crossed paths in qualification.

It can seem random (about two three games per, probably that way for most teams) but one would think that Head to head versus Side by Side alliance members would be more even. Not off by a about a factor of two.

I get what 179 is saying also about game scheduling and time between matches also one regional is not enough sample size but it is curious to look at. Like I said I noticed this last year, I see patterns quite well my job as scout/strategy mentor and it carries on into this year. Perhaps with more years I will see it start to even out. For now it doesn't really matter other than to satisfy my curiosity on how match scheduling really occurs.

The colors have nothing to do with my personal ranking of individual teams. Rather its just so I can identify the different top teams with different shades of green to better visualize it.

I'll run similar test next year to see if it is similar or balances it out assuming we go to one the powerhouse regional's again..they are fun. Hard to miss them around CA.

In the end I would not doubt this is done for entertainment value, as a bunch of 50 point games is not real exciting (140 is) however its not the most efficient way of getting as 179 calls it the "next generation" to show up regularly at worlds . They seem to have an uphill swim and scheduling in regional's could explain some of it. Perhaps it can be a reason why some don't make it for quite a while. But occasionally they do punch through. In no way did this affect us we got what we earned in both regional's and could not ask for more than that this season.

Whatever you think you're seeing here, you're wrong. There is nothing about the algorithm that weights whether a team has made it to Champs. You're grasping at straws.

[wireties]

Originally Posted by Cory

Whatever you think you're seeing here, you're wrong. There is nothing about the algorithm that weights whether a team has made it to Champs. You're grasping at straws.

QFT - The algorithm (at least a few years ago) does not include team numbers or any kind of history. It starts with team #1 (attending that regional) and goes through team X (attending that regional). It is randomized a bit but it has nothing do with the FRC team number until it fills in that the nth team to register was team ABCD for the report they feed to pit admin.

With respect you are way out in left field on this topic.

[Boltman]

Originally Posted by wireties

QFT - The algorithm (at least a few years ago) does not include team numbers or any kind of history. It starts with team #1 (attending that regional) and goes through team X (attending that regional). It is randomized a bit but it has nothing do with the FRC team number until it fills in that the nth team to register was team ABCD for the report they feed to pit admin.

With respect you are way out in left field on this topic.

I may be "in left field" a simple check will either prove a trend or not.

[matthewdenny]

I would like to note that with small sample sizes it is statistically likely to see patterns that aren't actually there.

[Boltman]

Originally Posted by matthewdenny

I would like to note that with small sample sizes it is statistically likely to see patterns that aren't actually there.

I know that is just one regional, I am not drawing a conclusive statement just what I noticed in two years of actually analyzing stuff. Just presenting limited data in one regional here.

It did not affect us one bit. Neither did losing 2 RP in San Diego. We were good enough to get what we deserved either way.

[Boltman]

Originally Posted by Cory

Whatever you think you're seeing here, you're wrong. There is nothing about the algorithm that weights whether a team has made it to Champs. You're grasping at straws.

Perhaps, anyhow we had a ton of fun playing with all the great teams up in
NoCal. Good luck 254 your bot rocks.

[Lil' Lavery]

There is no "perhaps" here. The algorithm does not factor in team numbers or team history. Rather than pontificate here, consider talking with an FTA or reading up on the matchmaking algorithm to understand how it works.

http://www.idleloop.com/matchmaker/

[Boltman]

Originally Posted by Lil' Lavery

There is no "perhaps" here. The algorithm does not factor in team numbers or team history. Rather than pontificate here, consider talking with an FTA or reading up on the matchmaking algorithm to understand how it works.

http://www.idleloop.com/matchmaker/

Thanks for the link! I like having more info.

"it's desirable to minimize duplicates so that teams see as many different teams as possible. It's especially desirable for a given team not to have the same team as a partner more than once, nor as an opponent more than once."

I think they could have split 1678's second match against us to another match just sayin'. Its not really ideal for any team to face consensus #1 twice in qualifications or 20% of their matches.

I'm sure the algorithm tries to be unbiased however as we know it can be imperfect (as the above link states) a human FTA should look for duplicates and adjust as necessary. Just flip teams above or below any "peculiarity" a one game switch in the schedule is doable.

I am not pontificating (you chose to read a post in my week 2 update thread), just pointing out the realities of a single regional's scheduling that happened to have 4 World Champions everyone in FRC knows.

Also I disagree with this statement "This is not only a scheduling algorithm issue, but a simple matter of not knowing in advance which teams will turn out to be more effective than others."

I believe "Class" has a lot to do with year to year success in a 20 year organization of 4000 or so teams... just like in horse racing. So to say that 254 or 1678 or 973 or 1323 are all the sudden not going to be good is a false statement. they will likely be very good because of their programs and their history. They would not accept anything less. Same here we are not going to fall off the face of FRC either every year we will iterate better. You can assume certain things from how teams have done in the past. Just look at their year over year results its all there.

I think to really do scheduling "correctly" some unbiased FTA needs to "adjust" when the computer prints out what it thinks is optimal based on some algorithm. The FTA knows or should know the perennial teams and could easily see the duplicates or inordinate amount of powerhouse pairings with a single highlighter . They could adjust anytime up to when its posted. It can be done and I think it improves the game play as well if instituted correctly.

The main problem is.. teams only get maybe 20 matches in two regionals so scheduling draw is huge just like "missed RP's" as a high order ranking metric.

I'm not afraid to point things out to improve the game hopefully for every team. FRC is great yet can still improve and tweak. Its good to be questioned and not accept always the status quo. I may just have a fresh view on an older institution. Things can always improve.

This is not why we are not in St Louis. All I'm sayin' it could be done better...and perhaps it makes a smidgen of difference if they actually tweak the way schedules are made. Might take a half hour or less at each regional to flip a few teams a few spots with the analytic help.

-----------------------------------------------

Attached is the way I do schedules lets pretend its a different season and in this scenario 3495 is a super powerhouse and consensus #1 or won 4 WC divisions and its our second match with them on our side. Great for us but seemingly "unfair" for having #1 twice on our side (or against us)

I mark partners in green and opponents in red for my scouts. Whites we do not play.

So if an FTA wanted to by switching 5137 for 2135 a one game switch a second powerhouse pairing match is avoided then all they have to do is make sure the other 5 teams did not play that team on same side or not. There are plenty of white spaces to switch teams around IF the goal is to be more random and possibly more fair to all teams.

If that does not work then there are two other "blue" teams that could make the switch.

We could have probably used the extra game in between as well there...in that scenario.

Just like with instant replay, changes can be made that are fairly simple and don't open a can of worms.... what might actually happen is all the sudden more teams get super good alliance partners and avoid facing powerhouses too much until eliminations. Schedules are always pretty late anyhow.

I think a human eye on it could easily sort out inconsistencies, much easier than trying to program some algorithm..its not rocket science. Let the algorithm do the sort.. then highlight several inconsistencies and move things around slightly.

[Chris is me]

Originally Posted by Boltman

I think to really do scheduling "correctly" some unbiased FTA needs to "adjust" when the computer prints out what it thinks is optimal based on some algorithm. The FTA knows or should know the perennial teams and could easily see the duplicates or inordinate amount of powerhouse pairings with a single highlighter . They could adjust anytime up to when its posted. It can be done and I think it improves the game play as well if instituted correctly.

This proposition is inherently contradictory. You are asking for an "unbiased" volunteer to introduce bias into the schedule based on their perception of historical robot / team quality. No matter how you slice it, that is biasing the results even farther from random than what they are, and there isn't even a chance of a volunteer doing this "fairly".

Changing the algorithm to try and pair teams of different ages with each other was tried once, in 2007, with generally disastrous results. The match making generator is already far from random when you consider all of the constraints on it - adding more constraints to make schedules "fairer" all but requires loosening one of the other constraints. In 2007, this meant matches had a consistent mix of young teams, old teams, and middle aged teams... at the cost of having to play with and against the same robots over and over again. It was almost universally disliked.

My advice? Get over the fact that schedules won't ever be "fair" however you see it, and work on becoming a team that isn't dependent on a perfect schedule to play well into Saturday / Sunday afternoon.

[1493kd]

Originally Posted by Chris is me

Changing the algorithm to try and pair teams of different ages with each other was tried once, in 2007, with generally disastrous results.

I will agree 100% with the 2007 system of scheduling being a mess. It was my second year on the team and we ended up being ranked 2nd at UTC with a robot that never hung a tube once (maybe 2) and was complete garbage. We were lucky and got paired up with 25 twice and many other great teams to push up to being ranked 2nd.

I felt guilty ranking so high with such a poor performing robot. I refuse to even count it as our highest ranking ever at an event.

[Boltman]

Originally Posted by Chris is me

This proposition is inherently contradictory. You are asking for an "unbiased" volunteer to introduce bias into the schedule based on their perception of historical robot / team quality. No matter how you slice it, that is biasing the results even farther from random than what they are, and there isn't even a chance of a volunteer doing this "fairly".

Changing the algorithm to try and pair teams of different ages with each other was tried once, in 2007, with generally disastrous results. The match making generator is already far from random when you consider all of the constraints on it - adding more constraints to make schedules "fairer" all but requires loosening one of the other constraints. In 2007, this meant matches had a consistent mix of young teams, old teams, and middle aged teams... at the cost of having to play with and against the same robots over and over again. It was almost universally disliked.

My advice? Get over the fact that schedules won't ever be "fair" however you see it, and work on becoming a team that isn't dependent on a perfect schedule to play well into Saturday / Sunday afternoon.

Been there done that already as I repeatedly said this did not affect us (but may affect others).... we held our own in CV and Ventura last year both days . I think that small changes could be made though to benefit all teams and saying that's "just the way it is" is not really acceptable.

I foresee maybe about three glaring inconsistencies that may need to be adjusted per "powerhouse" event... an hour of work maybe just making sure it does not mess another team up. The system is decent I believe it can be improved. Just like missed defense crossings that also can improve. Pretty sad when Portcullis crossings are missed especially the way we did them