Car Nack Predicts 16-1

[Car Nack]

Car Nack predicts that in the regional and district events that the number one seed will have a very difficult time winning the event. In fact the number one seed will win in less than 25% of the events.

Car Nack has spoken.

[Christopher149]

Originally Posted by Car Nack

Car Nack predicts that in the regional and district events that the number one seed will have a very difficult time winning the event. In fact the number one seed will win in less than 25% of the events.

Car Nack has spoken.

The almighty Car Nack has returned!

Hmm, seems like a 2012 sort of prediction. Considering we were picked by the #2 seed at an event in 2012 and lost in QF, I do not disagree.

[Steven Donow]

Originally Posted by Christopher149

The almighty Car Nack has returned!

Hmm, seems like a 2012 sort of prediction. Considering we were picked by the #2 seed at an event in 2012 and lost in QF, I do not disagree.

Bouncing off of this, does anyone have the data on the % of #1 seeds that won their events over the past few years?

[JesseK]

This prediction boils down to minimizing the advantages of first seed (first pick, playing vs 8th seed in QF's which allows the alliance to become cohesive before facing tougher opponents) while maximizing the disadvantage of first seed (16th pick).

• No 2nd picks which fill in the gaps for the 1st seed are available as the 16th pick
• Theory that the first seed will likely be a team who did the 'defense agreement' too much, therefore no other captains want to be their partner since no one knows what the 1st seed is REALLY capable of

Sure I'm missing others...

Curiously, I wonder if what led to this prediction is the thought that 8th seed (or 4th/5th seed) may have the upper hand in QF's simply due to variety in selection.

[T3_1565]

Maybe this is based off of a 3rd robot being more important in this game then others?

I don't think there is much room for "hold my ramp and don't move" robots this year. Plus even for the argument on defensive bots as a 3rd bot, the bottom list of teams will be teams that attempted a drivetrain for cross defenses and failed, There drivetrain may be worse off then a kitbot that far down the list.

Just my thoughts on the reasoning.

[IronicDeadBird]

I'm personally just blown away by such a strong read by someone day 4 of the competition. Everything in it lines up so well. I just feel dumb right now.

[Kevin Sevcik]

I suspect part of this is the assumption that those bonus ranking points are going to shake up the standing more than usual. 1 RP in quals probably is worth more than 25 match points in elims, so Car Nack has a bit of a point. The attributes that make a #1 seed aren't as well aligned with winning Elims as they usually are.

[plnyyanks]

Originally Posted by Steven Donow

Bouncing off of this, does anyone have the data on the % of #1 seeds that won their events over the past few years?

I was curious too, so I whipped up a quick script to find out. Script and full output are available on my GitHub: https://gist.github.com/phil-lopreia...276a608edc7de4

Overall 266 of 435 events were won by top seeds (61.1494252874 percent)
  In 2011, 41 of 62 events were won by top seeds (66.1290322581 percent)
  In 2012, 46 of 73 events were won by top seeds (63.0136986301 percent)
  In 2013, 46 of 81 events were won by top seeds (56.7901234568 percent)
  In 2014, 51 of 102 events were won by top seeds (50.0 percent)
  In 2015, 82 of 117 events were won by top seeds (70.0854700855 percent)

So Car Nack's prediction of <25% would make this year a pretty significant outlier when compared to the past 5.

[The Lucas]

Originally Posted by plnyyanks

I was curious too, so I whipped up a quick script to find out. Script and full output are available on my GitHub: https://gist.github.com/phil-lopreia...276a608edc7de4

Overall 266 of 435 events were won by top seeds (61.1494252874 percent)
  In 2011, 41 of 62 events were won by top seeds (66.1290322581 percent)
  In 2012, 46 of 73 events were won by top seeds (63.0136986301 percent)
  In 2013, 46 of 81 events were won by top seeds (56.7901234568 percent)
  In 2014, 51 of 102 events were won by top seeds (50.0 percent)
  In 2015, 82 of 117 events were won by top seeds (70.0854700855 percent)

So Car Nack's prediction of <25% would make this year a pretty significant outlier when compared to the past 5.

I made my on bold prediction of <40% #1 seed winners back in '09 and I (along with most of Michigan) pretty much got dominated by HOT (leading to a total of 50% #1 seed winners). I hit my makeup prediction of exactly one #1 seed division winner (again HOT with capt Wildstang and Spartan Robotics, the eventual champions). This is a much bolder prediction, perhaps Car Nack's boldest ever. Major props to him if he is right. I wonder what the biggest contributing factor (extra RPs, lack of depth + serpentine + random match sched, indiv ability ceiling) was to prompt this prediction.

You can find more data for '09 and '08 in my prediction thread. Check out this great Jim Zondag chart for '07-'10.

[Donut]

I think Car Nack has this one wrong, 25% is too low, but I can see where this prediction is coming from.

The ranking points from breaching and capturing will skew rankings, perhaps making the 1-4 seeds more equal.

The bonus points from a capture and breach will be critical to winning an event. At events lacking depth, the last pick may not have a drivetrain that can consistently cross a defense. This could put the 1 seed in a predicament of asking the last pick to play defense with a large risk of them not capturing at the end, or camp in the opposing courtyard for the capture but play the match 2v3. This could be an issue against a more balanced 2 - 4 seed in the semis or finals.

The third alliance partner is worth 30 points just by Challenging for a capture. How many other years was the third partner that critical?

[Edxu]

Originally Posted by Donut

I think Car Nack has this one wrong, 25% is too low, but I can see where this prediction is coming from.

The ranking points from breaching and capturing will skew rankings, perhaps making the 1-4 seeds more equal.

The bonus points from a capture and breach will be critical to winning an event. At events lacking depth, the last pick may not have a drivetrain that can consistently cross a defense. This could put the 1 seed in a predicament of asking the last pick to play defense with a large risk of them not capturing at the end, or camp in the opposing courtyard for the capture but play the match 2v3. This could be an issue against a more balanced 2 - 4 seed in the semis or finals.

The third alliance partner is worth 30 points just by Challenging for a capture. How many other years was the third partner that critical?

Third partner was worth 20 points per cycle in Aerial Assist for purely spitting the ball out of their intake, and most teams got around to 4-5 cycles per match at Champs, so that's a value of 80-100 points in a game where average score at champs was probably around 150-180.

But yeah, I would agree that the third alliance partner is crucial in this game, especially if they can be optimized to play anti-shooter defense and Secret Passage defense at the same time.

[plnyyanks]

Originally Posted by The Lucas

I made my on bold prediction of <40% #1 seed winners back in '09 and I (along with most of Michigan) pretty much got dominated by HOT (leading to a total of 50% #1 seed winners). I hit my makeup prediction of exactly one #1 seed division winner (again HOT with capt Wildstang and Spartan Robotics, the eventual champions). This is a much bolder prediction, perhaps Car Nack's boldest ever. Major props to him if he is right. I wonder what the biggest contributing factor (extra RPs, lack of depth + serpentine + random match sched, indiv ability ceiling) was to prompt this prediction.

You can find more data for '09 and '08 in my prediction thread. Check out this great Jim Zondag chart for '07-'10.

For completeness sake, I ran my script going to back to 2007 (as far back as TBA has ranking data)

Overall 363 of 628 events were won by top seeds (57.8025477707 percent)
  In 2007, 17 of 41 events were won by top seeds (41.4634146341 percent). 0 events were skipped
  In 2008, 24 of 45 events were won by top seeds (53.3333333333 percent). 0 events were skipped
  In 2009, 21 of 44 events were won by top seeds (47.7272727273 percent). 8 events were skipped
  In 2010, 35 of 50 events were won by top seeds (70.0 percent). 5 events were skipped
  In 2011, 41 of 62 events were won by top seeds (66.1290322581 percent). 0 events were skipped
  In 2012, 46 of 73 events were won by top seeds (63.0136986301 percent). 0 events were skipped
  In 2013, 46 of 81 events were won by top seeds (56.7901234568 percent). 0 events were skipped
  In 2014, 51 of 102 events were won by top seeds (50.0 percent). 0 events were skipped
  In 2015, 82 of 117 events were won by top seeds (70.0854700855 percent). 0 events were skipped

[lynca]

Originally Posted by plnyyanks

For completeness sake, I ran my script going to back to 2007 (as far back as TBA has ranking data)

So awesome Phil ! Thanks for the statistics

[Richard Wallace]

Through Week 1, Car Nack's 25% prediction is not looking good. 9/17 (52.9%) of #1 seeds have captained winning alliances.

[Kevin Leonard]

Originally Posted by Richard Wallace

Through Week 1, Car Nack's 25% prediction is not looking good. 9/17 (52.9%) of #1 seeds have captained winning alliances.

That number is, however, lower than in previous years, especially 2015. I was very surprised by the number of upsets this weekend, and I wonder if Car Nack was onto something, or if Week 1 was a fluke.