Saturday Ramblings

[FishDuck Article]

For the first time in three months, outside of a bye week, the Ducks do not have a game this weekend. Of course, they should have had one last night, but they did not. Last Saturday’s soul-crushing meltdown in Corvallis has left us gameless this first weekend in December. Personally, I do not like it. So, I’m going to ramble. ...

 

Saturday Ramblings

FISHDUCK.COM

---

Darren Perkins of FishDuck.com is frustrated that the Ducks did not have a game this weekend.

[Smith72]

Darren, I agree with your statement, "Analytics does not take this into consideration, and it does not take into account momentum, injuries, or emotion." 

 

An experienced play caller knows when to hold 'em and when to fold 'em (punt). An experienced play caller knows to take the 3 and not to chase points for the rest of the game.

 

Thanks Darren for your good "ramblings".  Good points to ponder.

 

 

[McDuck]

Math nerds should not be making football decisions any more than a head football coach should be making atmospheric re-entry calculations for NASA. End of story.

 

Not necessarily true.  Math whiz Nate Silver revolutionized baseball scouting using algorithms.  From an old New York Magazine article.

 

Furthermore, Silver himself invented a system called PECOTA, an algorithm for predicting future performance by baseball players and teams. (It stands for “player empirical comparison and optimization test algorithm,” but is named, with a wink, after the mediocre Kansas City Royals infielder Bill Pecota.) Baseball Prospectus has a reputation in sports-media circles for being unfailingly rigorous, occasionally arrogant, and almost always correct.

 

How Nate Silver Went From Forecasting Baseball Games to Forecasting Elections -- New York Magazine - Nymag

NYMAG.COM

---

Nate Silver is a number-crunching prodigy who went from correctly forecasting baseball games to correctly forecasting presidential primaries—and perhaps the election itself. Here’s how he built a...

 

[DazeNconfused]

Darren, I totally agree with what you said below. This may be the paragraph of the year in my book!

 

"Analytics essentially allows math nerds to make gametime decisions for you, and they shouldn’t. It is completely logic-driven, and football is a game played and coached by completely illogical humans. Analytics does not take this into consideration, and it does not take into account momentum, injuries, or emotion.

Math nerds should not be making football decisions any more than a head football coach should be making atmospheric re-entry calculations for NASA. End of story."

[Mic]

On 12/3/2022 at 7:58 AM, DazeNconfused said:

Math nerds should not be making football decisions any more than a head football coach should be making atmospheric re-entry calculations for NASA. End of story."

When schools and teams decide it's time for "math nerds" with all their algorithms and analytics to start calling the shots (plays)  it'll be time to replace frail human bodies with stainless steel robots and machines.  Imagine the sound of colliding bodies that'll generate! 

[DazeNconfused]

On 12/3/2022 at 7:43 AM, McDuck said:

Math nerds should not be making football decisions any more than a head football coach should be making atmospheric re-entry calculations for NASA. End of story.

 

Not necessarily true.  Math whiz Nate Silver revolutionized baseball scouting using algorithms.  From an old New York Magazine article.

 

Furthermore, Silver himself invented a system called PECOTA, an algorithm for predicting future performance by baseball players and teams. (It stands for “player empirical comparison and optimization test algorithm,” but is named, with a wink, after the mediocre Kansas City Royals infielder Bill Pecota.) Baseball Prospectus has a reputation in sports-media circles for being unfailingly rigorous, occasionally arrogant, and almost always correct.

 

How Nate Silver Went From Forecasting Baseball Games to Forecasting Elections -- New York Magazine - Nymag

NYMAG.COM

---

Nate Silver is a number-crunching prodigy who went from correctly forecasting baseball games to correctly forecasting presidential primaries—and perhaps the election itself. Here’s how he built a...

 

Like Allen Iverson sort of said.. 'what are we talking about, baseball? We talking about baseball"

 

Baseball is not the same game. Properly striking a pitch involves one player in the batters box, fielding a ball cleanly does not rely on teammates. Football has 11 players on each team involved in every play with.

 

Lastly the math nerds can't account on the fly for things like:

if you punt the ball is the wind at your back and in the other teams face on offense

if your punter strong and they have a weak punter

does your punter get you lots of fair catches 

is it starting to rain

is the other teams QB, RB, OL or WR now banged up in this game

if you don't get the 4th down will your teams emotions let down on both sides of the ball

if you punt and get a 3 and out will your team get fired up

if you punt with the wind behind your back and and get a 3 and out and they punt into the wind will you pick up field position 

if you go for it and fail what's the effect of getting the home teams crowd fired up more

is your QB. RB, OL, or WR banged up in this game making it harder to convert this 4th down

if you go for it at your own 29 yard line on 4th down and fail - what's the emotional effect on you defense being given a short field to defend

 

 

[Darren Perkins]

On 12/3/2022 at 7:43 AM, McDuck said:

Math nerds should not be making football decisions any more than a head football coach should be making atmospheric re-entry calculations for NASA. End of story.

 

Not necessarily true.  Math whiz Nate Silver revolutionized baseball scouting using algorithms.  From an old New York Magazine article.

 

Furthermore, Silver himself invented a system called PECOTA, an algorithm for predicting future performance by baseball players and teams. (It stands for “player empirical comparison and optimization test algorithm,” but is named, with a wink, after the mediocre Kansas City Royals infielder Bill Pecota.) Baseball Prospectus has a reputation in sports-media circles for being unfailingly rigorous, occasionally arrogant, and almost always correct.

 

How Nate Silver Went From Forecasting Baseball Games to Forecasting Elections -- New York Magazine - Nymag

NYMAG.COM

---

Nate Silver is a number-crunching prodigy who went from correctly forecasting baseball games to correctly forecasting presidential primaries—and perhaps the election itself. Here’s how he built a...

 

I knew somebody would call me out on that. I meant "in general." I figured there was probably a coach who was a rocket scientist. Thanks for pointing that out. 

[Mic]

On 12/3/2022 at 8:16 AM, DazeNconfused said:

Like Allen Iverson sort of said.. 'what are we talking about, baseball? We talking about baseball"

 

Baseball is not the same game. Properly striking a pitch involves one player in the batters box, fielding a ball cleanly does not rely on teammates. Football has 11 players on each team involved in every play with.

 

Lastly the math nerds can't account on the fly for things like:

if you punt the ball is the wind at your back and in the other teams face on offense

if your punter strong and they have a weak punter

does your punter get you lots of fair catches 

is it starting to rain

is the other teams QB, RB, OL or WR now banged up in this game

if you don't get the 4th down will your teams emotions let down on both sides of the ball

if you punt and get a 3 and out will your team get fired up

if you punt with the wind behind your back and and get a 3 and out and they punt into the wind will you pick up field position 

if you go for it and fail what's the effect of getting the home teams crowd fired up more

is your QB. RB, OL, or WR banged up in this game making it harder to convert this 4th down

if you go for it at your own 29 yard line on 4th down and fail - what's the emotional effect on you defense being given a short field to defend

 

 

Oh, I'm pretty sure some math nerd could probably develop some sort of statistical analysis's that could generate some algorithms that could take all those eventualities in stride and come up with something.  But we may have to extend the play clock to 1 minute, 30 seconds.

[lownslowav8r]

On 12/3/2022 at 7:43 AM, McDuck said:

Not necessarily true.  Math whiz Nate Silver revolutionized baseball scouting using algorithms.

These findings are always based upon initial conditions and are by nature probabilistic. When the initial conditions aren’t matched, then you are guessing. Game time decisions are especially fraught given the ever changing circumstances. Lot harder than the relatively static art of scouting. 

[Mic]

On 12/3/2022 at 8:57 AM, lownslowav8r said:

These findings are always based upon initial conditions and are by nature probabilistic. When the initial conditions aren’t matched, then you are guessing. Game time decisions are especially fraught given the ever changing circumstances. Lot harder than the relatively static art of scouting. 

Bottom line:  Let's just hope the game stays essentially as it is and the math nerds stay in the classroom or in the stands.

[HappyToBeADuck]

Darren thanks for the article. Been playing the "what if" game all week. Disappointment is not the emotion that I was hoping would rule my feelings this weekend.

 

10 days ago, i was certain we would beat the Rodents then have a shoot out against usc. Sure didnt land that way. However the Utes humbling the traitors was special and moved my disappointment meter.

 

As for analytics?

 

Did LR use analytics to go for it on 4th and 8? They failed, Utah got a short field and momentum. Little did we know that the game for the traitors was over at that point. The Utes seized the moment, took control and started their journey to the Rose Bowl. Analytics failed.

 

Did KW use analytics when he went for it on 4th and 2 in trojan territory? Or did he use common sense or a gut feeling? They made the first down and later scored a TD. If it was analytics then it succeeded.

 

Later in the game, KW ignored the analytics and punted around mid-field on 4th down. His in game experience told him that his D had taken control. A measurement that Nerd filled analytics may not be able to compute. KD ignored analytics. He did not give the trojans a chance to get momentum back. Thats smart coaching.

 

Analytics had very little to do with the butt kicking the Utes put on the traitors after LR's analytical mistake.

 

You dont hold a team to 1 yard of total offense in the 3rd quarter with analytics. Half time adjustments and smash mouth football by the Utes crushed the traitors false hope of a coronation.

 

Analytics does not create turnovers nor take into account a QB injury. The Ute DC turned his beasts loose on a wounded prey. Caleb Williams could not overcome his injury. He gave it is all but fell short.

 

My hat is off to an injured Bo Nix and his offensive team mates for beating Utah. They overcame......

 

Still injured Bo and the Duck O beat the Beavers. Poor coaching, poor analytics and a invisible Defense lost that game.

 

I really hope DL and his staff recorded the PAC title game. KW, his staff and players laid out the blue print for back to back trips to the Rose Bowl.

 

That's a jouney we all would like to see our Ducks take........

Edited December 3, 2022 by Smith72
Typos KW instead of KD

[Duck Fan 76]

On 12/3/2022 at 8:28 AM, Darren Perkins said:

I knew somebody would call me out on that. I meant "in general."

Daren, first let me say that I truly appreciate the article you wrote for us and I think you've tapped into a really important theme not just in football but in 21st century society.  Books galore are literally (pun intended) being written on this subject although too few in my opinion focus on football specifically.

 

Let me say right out in the open that I'm a football loving math nerd.  In fact my background is as a data scientist and I routinely work with AI/ML algorithms.  I would like to politely enter the discussion on the topic at hand and wish to do so without pissing anybody off.  ("hope in one hand...)

 

"Did analytics cause the decision for Dan Lanning to go for it late in the game with a 4th and 1 from his own 29-yard line? He doubled down on his 4th down from his own territory call against the Huskies. He wanted to prove that going for it in your own territory is the right thing, and he failed."   More on this statement later.

 

Let me open with agreeing with the theme of your statement that going for it on 4th and 1 from our own 29 was the wrong call in a tough rivalry game.

 

My position on data analytics in sports is a bit more nuanced and to explain it I need to get some history out on the table.

 

Mathematicians have been studying games for a very long time.  The famous mathematician Blaise Pascal helped to discover the concept of "expected value" around 1650 and that discovery has underpinned optimal strategies for games that revolved around chance ever since (dice, cards, roulette, etc.).  There would be no professional gamblers without this area of math as games of chance that pay out are purpose built to favor the house.  There are a lot of mathematicians employed in Vegas working tirelessly to stay ahead of these developing counter strategies.

 

The science of games really took a major leap forward however in the 20th century.  John von Neuman published a revolutionary paper on the theory of games of strategy.  Along with Alan Turing, von Neuman is the Einstein level genius that invented the modern science of computers.  His theory on games of strategy was established as the new field of study named "game theory" and was a heavily funded area of research by the 1950's due to the cold war. 

 

The US and Soviet governments spent BIG money trying to figure out what was the optimal strategy for a nuclear armed power and game theory and computer simulation was the primary methods employed.  The obvious concern was that "going with the gut" would get everyone killed.  Eventually it was game theory that proved that nuclear conflict was in reality a no-win game and both countries separately arrived at the conclusion to agree to arms control discussions.  That's an oversimplification of course but the point is game theory occupied some of the greatest minds of the 20th century.

 

That's all fine and dandy but what does it have to do with football?  My point with the nuclear discussion was to say that game theory is pretty well developed and it's not a hair brained flat earther scheme sitting out on the dark web.  In fact it quickly (1950's) moved into the area of strategic games of pleasure and almost nobody was happy with that.  Well maybe some math nerds.

 

Everyone over age 10 mostly understands that Tic-Tac-Toe isn't a very fun game to play, the reason?  If both players are paying attention nobody wins.  In fact game theory proved that Tic-tac-toe was unwinnable if both players play an optimal strategy.  When a proven optimal strategy is found for a game we call it solved.  Connect four, battleship, checkers are all solved games.  All of these games however have something in common, they are deterministic zero sum games with perfect information.  Meaning if I take a move in the game I always get exactly that move (deterministic).  Zero sum means there are no cooperative solutions where both players win, tie counts as a mutual loss.  Perfect information means you can see every piece on the board exactly at any point in time.

 

There are deterministic zero sum games with perfect information that we haven't solved such as chess and go.  The reason they aren't solved is the number of possible moves is to big to account for and we haven't found patterns where we can reduce those possible moves to a manageable number for an algorithm to work on.  That doesn't mean however that game theory isn't used to win these games.  

 

Before checkers was a solved game we had developed computer algorithms that could beat any human player in the world.  In 1996 Gary Kasparov lost the first round to IBM's deep blue chess computer which was the first time a human chess champion ever lost to a computer.  Kasparov battled back to win the match 4-2 and said "... I sensed a new kind of intelligence".  In 1997 he lost the match to deep blue and today the worlds best mostly refuse to play computers as they find it demoralizing.  

 

Sports in general are a LOT harder for computers to form winning strategies with.  They are stochastic games instead of deterministic meaning when I call a jet sweep sometimes I get a jet sweep and sometimes I get a fumble.  Perfect information is debatable but what I can say is that there are a lot of hidden variables for a football player.  Is the RB going to accelerate into the gap at 7 m/s^2 or did he go out on the town last night and today it's 6 m/s^2?  Sports in general might have something that "looks like" a player turn (aka a down) but the play itself is real time so observing variables in real time is in itself a very hard problem.

 

My point is we aren't there yet where computers are spitting out the "perfect play call" given all observable data.  Will we ever be?  Almost certainly barring human extinction.  When?  That's really hard to say any we have a word for people that predict when algorithms will accomplish some goal, "marketing".  The truth is innovation comes in bursts and they are mostly unpredictable so we could be 5 days or 5 centuries away from that capability.  It's why full self driving cars is a bad investment, the technology has promise but it's not quite there yet so keep your hands on the wheel.

 

Baseball saw the rise of data scientists with Sabermetrics and it's proven to be a competitive advantage for teams but not as big a one as huge salary pools.  Football absolutely isn't baseball.  Baseball is checkers and Football is Starcraft.  Statistics in baseball are more relevant due to the highly repeatable situational effort (pitches, hitting, fielding, etc.).  You have a pitcher literally on an island performing a task.  A batter trapped in a box countering the pitchers move.  It really does lend itself quite well to stochastic methods in game theory.  Even so sometimes players get the yips and the data models fail.

 

Does all this mean that Football is immune to data analytics?  Unfortunately no, it doesn't.  Data analytics is a very useful tool for making competitive decisions of all sorts.  Business uses it, the DoD uses it, the Yankees use it, Manchester United uses it and so do the Oregon Ducks.  The real key here is understanding that data analytics is just one tool in the box and while it has a purpose it's not a swiss army knife... at least not yet.

 

Now I return to Daren's quote "Did analytics cause the decision for Dan Lanning to go for it late in the game with a 4th and 1 from his own 29-yard line?"  Daren answered this question himself "He wanted to prove that going for it in your own territory is the right thing, and he failed."  He in this instance is Dan Lanning and the fact that he wanted to see a particular outcome other than just winning the game means he DIDN'T use data analytics to make the decision.  He used his emotions and possibly justified it with ad hoc data analytics.

 

How the human mind generates a decision is extremely complicated and we absolutely don't understand it.  We do however think we might understand some of it and if you're interested in getting a non-technical overview it can be found in the book "Thinking Fast and Slow".  The main idea is that "slow" rigorous logical reasoning is extremely calorically expensive (mental exhaustion anyone?) and while it's a competitive advantage in nature in some situations over-thinking will actually get you killed.  Enter the "fast" thinking of the human "gut" that we came pre-equipped with.  Fast thinking takes a lot of logical short cuts and arrives at a cheap but usually right answer for areas where the individual has a LOT of experience.

 

The idea then is if you are doing something you are an expert in, don't overthink it just make the call and move on.  If you are in new territory you should slow down and proceed cautiously.  This rule of thumb, follow your gut, is generally speaking the optimal solution for humans.  Where the problem can arise is that thinking "fast" will sometimes make errors.  In the case of the optimal nuclear strategy those errors have devastating costs.  How do we avoid them?  We force ourselves to slow down and think it through when the risks are too high.

 

Game theory, data analytics, computer simulations, computer decision algorithms are all just tools that are meant to help us slow down and think it through.  In the end they are just tools (for now anyway) and it's how the people in charge use them that matters in getting to the right situational decisions.  This concept is called human-machine teaming and we've been using it since we learned how to make pointy sticks.  The stick isn't the thing, it's the hand holding it, the eye aiming it and the mind deciding on what to do with it that has mattered.  The machines have certainly gotten fancier but we really haven't.  Same hand, same eye, same mind.

 

In football when it's 4th and 1 on your own 29-yard line and you are a new head coach late in a competitive rivalry game you are in an area of devastating consequences.  Those are exactly the times you need to force yourself to take 15 seconds to slow down and use some tools to help you make the right decision.  Data analytics is one of those tools and the danger is that the "thinking fast" system never stops so it's already made a call on what to do.  It says "go for it" because we didn't get here without being aggressive, aggressive has paid off for us. 

 

The hardest thing to do in that moment is to force yourself to be aware that part of you has already decided and the calorically efficient thing to do with an exhausted mind is to use a ready tool to tell yourself a story on why that decision is the right decision.  This is hard stuff.  This is hard stuff for football coaches, this is hard stuff for squad leaders in the Rangers, this is hard stuff for me when I'm debating on writing 2k words on the forum or getting started on that project I'm supposed to be working on.

 

It helps to work with people you respect that can challenge your instincts and give you pause when you have a tough decision.  Mistakes will happen, mistakes were made. 

 

 

 

 

 

 

Edited December 3, 2022 by Duck Fan 76