Close Games / Meaningful Games

[pnkranger]

Some of you guys need to get off the fence. I would think one would either be a Ken Pom disciple or not.

Mr. Pomeroy was interviewed on a local radio station day before yesterday. He came across as bright, knowledgeable, and articulate. He was asked the specific question about OU's string of close losses in reguard to how his luck factor was computed.

He explained that what happens in the last 2/3 minutes of close games isn't much more than a random outcome. He went on to add that even the complete 18 game Big 12 schedule is so impacted by randomness that it is unlikely that the final standings are an accurate indicator of which teams are actually better than others.

He is right about that. Trying to draw conclusions from outcomes that are highly impacted by randomness is wrong.

I don't think i have to get off the fence because there is no fence here. Let's look at the game of poker for a minute. No one would argue that poker is a game of statistics and that the randomness of the draw can often mean you play much better than your opponent and walk away broke.

But, as a player, you have control over that randomness by getting your money in good. If a player holds 77 and calls a bluff against an opponent holding 72, that means the 77 player has limited the potential randomness of the outcome, though it does still exist.

Basketball is the same way. If given the opportunity, would Foster have missed both of those shots a second time around? Maybe. Likely he would miss one of them at least.

However, in all of our close losses, we're not getting our money in good. In the last four minutes, we're not getting rhythm shots in the flow of our offense. We're not getting defensive rebounds, even with position, we're allowing key opponents (like Foster or Nigel-Goss) to dribble in on their terms and shoot the shot they wanted.

Last night:

• Buddy and Cousins both jacking up ill advised threes around the 3 minute mark with no offense run.
• DJ Bennett getting the rebound then falling down.
• Allowing three offensive rebounds on the final possession.
• Letting Foster set up the final three point shot.
• Isaiah Cousins missing the front end of the free throws.

While a statistical probability is attached to each of those items, I would argue that we negatively impacted the probability of those being a positive outcome through our performance.

Not running any offense and jacking up a contested three while down by 3 points late in the game is akin to getting all your money in the pot with a 72. Sometimes you'll win, but a lot more times you won't.

Last point: Randomness definitely impacts the final outcome, but a trend of negative randomness (1-12 in last 13 games decided by 8 points or less) shows that we are not getting our money in good.

 

[garyeb2]

I love your posts Gary but I disagree.

Playing well late in games is about toughness and discipline, not just random chance.

The reason that Woodard has taken a lot of last second shots over his career is because OU's best options can't get open. That is not random chance, that is teams defending better than OU is executing.

Teams will live with Woodard beating them and not Hield, Spangler, Thomas, etc.

That's what I'm seeing.

That is fine. I know you believe that and you want to assign those fine qualities to those that prevail in close games. But, it wouldn't be accurate to do it all the time. Sometimes the shot maker or winner just got lucky. Or, someone else got unlucky. Maybe it is 40% or 50% or 60% or more or less.

Just one example is the year (I can not remember which one) Kelvin's team had several one point wins. One of them when Eaton was called for a charge coming down the sideline during the closing seconds at the LNC.

All of those great abstract qualities were assigned to that team. They promptly got themselves knocked out of the tourney by a much better team because we had just gotten lucky to get there in the 1st place.

And besides, the only reason I jumped in on this one is because several poster's want to trot out something Ken Pom has come up as evidence their position is correct. Either he is all knowing or he isn't. I'm just shining a light on the flip floppers.

 

[sheepdogs1]

Randomness has an impact on the outcome of close games. I don't know to what degree. Ken Pom doesn't know to what degree. 40 watt bulbs like yourself don't know to what degree. But, it is there.

Because of that, judgments and conclusions based on those outcomes are suspect at best and totally inaccurate at worst.

Thomas Jefferson said "luck occurs to the prepared." If you are prepared you make better decisions. This is not random in nature.

The impact of the game which is close involves two teams. If one team is more prepared then the randomness factor shifts more to the unprepared rather than the prepared.

 

[garyeb2]

I don't think i have to get off the fence because there is no fence here. Let's look at the game of poker for a minute. No one would argue that poker is a game of statistics and that the randomness of the draw can often mean you play much better than your opponent and walk away broke.

But, as a player, you have control over that randomness by getting your money in good. If a player holds 77 and calls a bluff against an opponent holding 72, that means the 77 player has limited the potential randomness of the outcome, though it does still exist.

Basketball is the same way. If given the opportunity, would Foster have missed both of those shots a second time around? Maybe. Likely he would miss one of them at least.

However, in all of our close losses, we're not getting our money in good. In the last four minutes, we're not getting rhythm shots in the flow of our offense. We're not getting defensive rebounds, even with position, we're allowing key opponents (like Foster or Nigel-Goss) to dribble in on their terms and shoot the shot they wanted.

Last night:

• Buddy and Cousins both jacking up ill advised threes around the 3 minute mark with no offense run.
• DJ Bennett getting the rebound then falling down.
• Allowing three offensive rebounds on the final possession.
• Letting Foster set up the final three point shot.
• Isaiah Cousins missing the front end of the free throws.

While a statistical probability is attached to each of those items, I would argue that we negatively impacted the probability of those being a positive outcome through our performance.

Not running any offense and jacking up a contested three while down by 3 points late in the game is akin to getting all your money in the pot with a 72. Sometimes you'll win, but a lot more times you won't.

Last point: Randomness definitely impacts the final outcome, but a trend of negative randomness (1-12 in last 13 games decided by 8 points or less) shows that we are not getting our money in good.

Have you ever played poker or do you just watch it on TV? I make a living picking off bluffs and trapping people on draws at the wrong price.

Ken Pom said that even an 18 game schedule is not enough to sort the teams out in the correct order. If 18 isn't enough, neither is 13. Either you are a Ken Pom man or you are not.

You may be 100% correct in your observations. But, the conclusions that you draw from those observations are some number less that 100% accurate due to the impact of randomness on results.

If you don't understand that, my best advice would be not to risk any money betting on ball games or playing poker.

 

[garyeb2]

Thomas Jefferson said "luck occurs to the prepared." If you are prepared you make better decisions. This is not random in nature.

The impact of the game which is close involves two teams. If one team is more prepared then the randomness factor shifts more to the unprepared rather than the prepared.

They weren't basketball games during Jefferson's life.

 

[sdcloud9]

That is fine. I know you believe that and you want to assign those fine qualities to those that prevail in close games. But, it wouldn't be accurate to do it all the time. Sometimes the shot maker or winner just got lucky. Or, someone else got unlucky. Maybe it is 40% or 50% or 60% or more or less.

Just one example is the year (I can not remember which one) Kelvin's team had several one point wins. One of them when Eaton was called for a charge coming down the sideline during the closing seconds at the LNC.

All of those great abstract qualities were assigned to that team. They promptly got themselves knocked out of the tourney by a much better team because we had just gotten lucky to get there in the 1st place.

And besides, the only reason I jumped in on this one is because several poster's want to trot out something Ken Pom has come up as evidence their position is correct. Either he is all knowing or he isn't. I'm just shining a light on the flip floppers.

I would never say that there is no randomness to late game situations, crazy things happen, absolutely they do. I was an assistant high school coach for 20 years and saw many crazy things that no one can prepare.

I am saying, however, that when an individual team struggles often in those situations, there are real events that make a difference.

Maybe OU will win the next four or five close ones and that will even things out but there will be finite reasons why they won those games as well as why they lost the ones currently.

 

[pnkranger]

Have you ever played poker or do you just watch it on TV? I make a living picking off bluffs and trapping people on draws at the wrong price.

Ken Pom said that even an 18 game schedule is not enough to sort the teams out in the correct order. If 18 isn't enough, neither is 13. Either you are a Ken Pom man or you are not.

You may be 100% correct in your observations. But, the conclusions that you draw from those observations are some number less that 100% accurate due to the impact of randomness on results.

If you don't understand that, my best advice would be not to risk any money betting on ball games or playing poker.

This is such a silly way to approach an argument. First, I have watched basketball for 30+ years and also won a large part of my total income through poker for a five or six year period. And I make a decent salary at work...

I understand the principles of both games and have applied them appropriately to this situation.

As for your last point about being 100% accurate ... there is no such thing in poker or basketball. That doesn't mean you should throw away the advantage you might have by chalking it all up to randomness and being happy with 1-12. You HAVE to understand the risk/reward structure of late game play and EXECUTE EFFECTIVELY to improve your chances at managing the randomness that occurs. You HAVE to get your money in good.

If YOU don't understand that, my best advice would be not to risk any money betting on ball games or playing poker.

 

[bocabull]

Randomness has an impact on the outcome of close games. I don't know to what degree. Ken Pom doesn't know to what degree. 40 watt bulbs like yourself don't know to what degree. But, it is there.

Because of that, judgments and conclusions based on those outcomes are suspect at best and totally inaccurate at worst.

Funny because this happens to be right up my alley as I earn a living running a quantitative investment management firm.

Any sophisticated algorithm will backtest with the goal of minimizing random occurrences. In addition the most predictive algorithms include a certain amount of manual discretion to override the formula when certain criteria exists.

Best I can tell Pom's "luck" variable is calculated by the outcome going against his base algorithm. What are the odds of me landing a major pension fund as a client pitching that formula? lol ZERO.

 

[stormspencer]

Funny because this happens to be right up my alley as I earn a living running a quantitative investment management firm.

Any sophisticated algorithm will backtest with the goal of minimizing random occurrences. In addition the most predictive algorithms include a certain amount of manual discretion to override the formula when certain criteria exists.

Best I can tell Pom's "luck" variable is calculated by the outcome going against his base algorithm. What are the odds of me landing a major pension fund as a client pitching that formula? lol ZERO.

You are correct. The luck factor is a variable that is quantified after an outcome has already taken place in his formular. It isn't heavily weighed into his final metric, but it is accounted for none the less.

Kenpom also doesn't factor in "wins" and "losses" into his formula....there are some folks out there who don't realize this. The most important variables in his formula are "point per 100 possessions" on offense and "points given up per 100 possessions"....both adjusted for opponents SOS and adjusted pace/tempo. He does include luck as a part of his metric, but it is a tertiary measure of your performance in "close' games.

One example where he busts up conventional wisdom is Pitt. For the last decade or so, people have thought of Pitt as a tough physical team that played great defense. That actually isn't the case at all. The reason they play in low scoring games is because their "adjusted pace/tempo" is always in the 300s. Thus, there are a limited amount of possessions in their games, but their actual adjusted defense ranking has rarely been in the top 50 in his rankings.....because the pace dictates the low scoring games rather than Pitt's supposed "great" defense.

The reason I site the above example is because people write off Kenpom because they don't understand it and/or don't want to put the effort into it to see how his system operates. I realize it is "rankings", but it's more than just a ranking....it's a probability measurement to perform in the future. A more accurate description would be a list of teams and their probability to perform in future games based upon points per possession. I think it is the best site out there because you can go back 2, 5, or 10 years and you will find that his top 10 or 15 teams perform very well in the NCAA tournament...with a few outliers of course. That is why I'm not too worried about OU moving forward this year. Last year, we were 33rd in his final list, which means we were "likely" to win no more than 1 game in the NCAA tournament. This year we are 10th....and if we stay around that, we have just under a 50/50 shot at the Elite 8....according to the probability. Sorry for the length of this.

 

[garyeb2]

This is such a silly way to approach an argument. First, I have watched basketball for 30+ years and also won a large part of my total income through poker for a five or six year period. And I make a decent salary at work...

I understand the principles of both games and have applied them appropriately to this situation.

As for your last point about being 100% accurate ... there is no such thing in poker or basketball. That doesn't mean you should throw away the advantage you might have by chalking it all up to randomness and being happy with 1-12. You HAVE to understand the risk/reward structure of late game play and EXECUTE EFFECTIVELY to improve your chances at managing the randomness that occurs. You HAVE to get your money in good.

If YOU don't understand that, my best advice would be not to risk any money betting on ball games or playing poker.

I could suggest a couple of books for you to read. But, past that. I can not help you. Please, don't try to BS me with your poker results. I know better. When someone tells me that they are a winning poker player my first thought is that either their recollection or record keeping is faulty.

Here are the facts. 90% of the people that play regularly in a raked game lose. 5% of the players are break even to the most marginal of winners. Only 5% have the ability to generate any measurable long term profit.

 

[pnkranger]

I could suggest a couple of books for you to read. But, past that. I can not help you. Please, don't try to BS me with your poker results. I know better. When someone tells me that they are a winning poker player my first thought is that either their recollection or record keeping is faulty.

Here are the facts. 90% of the people that play regularly in a raked game lose. 5% of the players are break even to the most marginal of winners. Only 5% have the ability to generate any measurable long term profit.

I could suggest a couple of books for you to read, including peer reviewed research and respected texts focusing on game theory and end game strategy. But, past that, I can not help you.

You can believe what you want to believe, but the bottom line is you haven't addressed the point being made. Instead, you have acted like an arrogant know-it-all without actually saying anything relevant to the discussion.

You are a great contributor when it comes to practice reports, but when anyone challenges anything about the program, you immediately take on a condescending tone and focus on the individual poster rather than the discussion being held. I have been a member of this group of posters since the early 2000s, long before we left the old board to come here, and that tact is not in line with the tone of this board.

Let's get back to the discussion. I will say for you very clearly what I have articulated multiple times in this thread and would like to hear your rebuttal.

According to game theory and, specifically, end game strategy, in any odds based game, the player must take certain specific actions to reduce the impact of chance if they hope to win over the long term.


In poker, for instance, a player in late position might raise against four callers to isolate against one player in early position. This limits chance by reducing the # of opponents/cards to draw against AND puts the raiser in a position of power betting AFTER the initial move of the earlier player.

I have already covered what this looks like in basketball, but players can reduce randomness through execution by running an offensive set (not shooting early shot clock threes without rebounders in place), positioning yourself well for a defensive rebound off of a free throw, or double teaming the opponent's best player, who has the ball in his hands, even though he beat you with a last second shot only weeks before.

I will admit that one season is a small sample size. However, these issues stretch back to last season and, against quality opponents tend to happen in every close game we've had under Kruger. Ignoring that ... even if both teams played equally in the last two minutes of OU's last 13 close games, you would expect that OU would have won more than 1 of those 13 (dating back to last season). It very well could be that the odds are simply against OU at this point; however ...

My argument is that OU is not doing enough to reduce the impact of chance in late game situations and, in many cases, is allowing the other team to reduce the impact of chance through our lack of execution.


Please, let's focus on the discussion and leave the personal barbs out of it. I'd like to hear your thoughts on this perspective.

 

[pnkranger]

Also, I want to be clear: I hope like hell that we have just been incredibly unlucky, and that the odds will even out when we get to post-season play. I would love to win 9 straight post-season games by 1 point...

 

[Big Old Booger]

Under Lon Kruger:

Games decided by 5 points or less: 14-15 (0-4 this season and 0-fer our last 6)
Overtime Games: 0-7
Post-season games: 0-5

Dating back to last season's loss @Iowa State, we are 1-12 in games decided by 8 points or less.

Do we lack a player with a killer instinct in our program?

Key depth, mainly at the bigs. We had Osby and nothing, then Osby and M'Baye, then Spangler and nothing now Thomas and Spangler. Young or inexperienced depth behind them or guys that really aren't good enough to play many minutes in this conference.

It's getting better though but it takes time to build a program the right way, which is what LK is doing.

I would like to see the breakdown of these games by season dating back to LK's first season. I know we have lost a lot of close games this year though and it is frustrating but we just don't have the player to make the play.

Look back at some of Kelvin's teams. He always had 1 guy whether it was Minor, Erdmann, Brewer, Najera, Price, etc., heck even Jameel Haywood made the last second shot at Mizzou in the lane to win. We just don't have those guys yet but we will.

 

[KBaden145]

Key depth, mainly at the bigs. We had Osby and nothing, then Osby and M'Baye, then Spangler and nothing now Thomas and Spangler. Young or inexperienced depth behind them or guys that really aren't good enough to play many minutes in this conference.



It's getting better though but it takes time to build a program the right way, which is what LK is doing.



I would like to see the breakdown of these games by season dating back to LK's first season. I know we have lost a lot of close games this year though and it is frustrating but we just don't have the player to make the play.



Look back at some of Kelvin's teams. He always had 1 guy whether it was Minor, Erdmann, Brewer, Najera, Price, etc., heck even Jameel Haywood made the last second shot at Mizzou in the lane to win. We just don't have those guys yet but we will.

This and nothing more. Completely agree.


Sent from my iPhone using Tapatalk

 

[stormspencer]

Also, I want to be clear: I hope like hell that we have just been incredibly unlucky, and that the odds will even out when we get to post-season play. I would love to win 9 straight post-season games by 1 point...

You may want to read this guy's blog and scroll down to the game summary. He maintains that "luck" is not necessarily indicative of execution down the stretch. He postulates that you can execute with precision in the waning minutes of the game (making the right pass, getting the shot you want etc.) and you might still miss the game winning-shot. Likewise, teams can be lackluster down the stretch and put up a bad shot....and it goes in for a game-winner. His point is that luck doesn't necessarily correlate to execution, but in a one-possession game, crazy stuff can happen. And one game doesn't necessarily translate to future games. Here is a quick summary of his OU/KSU recap:

"In addition, the end of this game was a demonstration of what I often talk about: late game execution doesn't really correlate with late game winning percentages. Kansas State had the ball with 1:01 to go in a tie game, which meant that they should go for a quick 2-for-1. Instead they intentionally burned clock, ruining the 2-for-1, but got their own offensive rebound off the miss. And then when their next possession went nowhere, Marcus Foster was forced to launch a deep three-pointer... which he hit. That's how it goes sometimes."

http://basketballpredictions.blogspot.com/

He also maintains that luck usually evens out to a degree over the course of the season....but it hasn't yet for us.

 

[pnkranger]

I don't disagree with him in a one game scenario. Anything can happen when you execute perfectly because you don't control all of the variables.

What I'm saying is that we haven't done ourselves any favors late in tight games to minimize the impact of luck/chance,

 

[garyeb2]

I could suggest a couple of books for you to read, including peer reviewed research and respected texts focusing on game theory and end game strategy. But, past that, I can not help you.

You can believe what you want to believe, but the bottom line is you haven't addressed the point being made. Instead, you have acted like an arrogant know-it-all without actually saying anything relevant to the discussion.

You are a great contributor when it comes to practice reports, but when anyone challenges anything about the program, you immediately take on a condescending tone and focus on the individual poster rather than the discussion being held. I have been a member of this group of posters since the early 2000s, long before we left the old board to come here, and that tact is not in line with the tone of this board.

Let's get back to the discussion. I will say for you very clearly what I have articulated multiple times in this thread and would like to hear your rebuttal.

According to game theory and, specifically, end game strategy, in any odds based game, the player must take certain specific actions to reduce the impact of chance if they hope to win over the long term.


In poker, for instance, a player in late position might raise against four callers to isolate against one player in early position. This limits chance by reducing the # of opponents/cards to draw against AND puts the raiser in a position of power betting AFTER the initial move of the earlier player.

I have already covered what this looks like in basketball, but players can reduce randomness through execution by running an offensive set (not shooting early shot clock threes without rebounders in place), positioning yourself well for a defensive rebound off of a free throw, or double teaming the opponent's best player, who has the ball in his hands, even though he beat you with a last second shot only weeks before.

I will admit that one season is a small sample size. However, these issues stretch back to last season and, against quality opponents tend to happen in every close game we've had under Kruger. Ignoring that ... even if both teams played equally in the last two minutes of OU's last 13 close games, you would expect that OU would have won more than 1 of those 13 (dating back to last season). It very well could be that the odds are simply against OU at this point; however ...

My argument is that OU is not doing enough to reduce the impact of chance in late game situations and, in many cases, is allowing the other team to reduce the impact of chance through our lack of execution.


Please, let's focus on the discussion and leave the personal barbs out of it. I'd like to hear your thoughts on this perspective.

I'm well versed in game theory and end game strategies as it relates to poker. You need to stop with the poker analogies if you can not get past poker pro 101 thought processes. I took that course back in 1989 and my grand son knows to raise the limpers and bet the flop and why.

If you want to argue, try confining it to the position that I have actually taken. Ken Pom said that the result of close games is highly impacted by randomness. I agree with that. I said that the poor execution of the losing team has to be discounted as the reason for the loss by an amount that is equal to the random impact. I don't know how much that is. Neither does ken Pom.

I am going on to say that even the poor execution could be a random event. How often does Jordan make that layup and how often does Foster hit a 3 with a hand in his face. When Jorden misses a layup he is favored to make and Foster hits a 3 he is favored to miss, those are random outcome and we got unlucky. The outcome we got was only one of many possible outcomes.

The result of any basketball game is only one of many possible outcomes. If we played West Virginia 10 times we would get 10 different results. You seem to be too result focused. It can not be about the result in poker or anything else. It is about the process. Lurching and jerking around over a bad result and advocating process changes is a mistake. It is a small sample size mistake.

Like I said earlier. Even the poor execution could be a random event. It is like when Buddy bounces the ball off his foot every 2-300 dribbles or so. No one works harder than Buddy trying to improve. But, it looks like he is going to continue bouncing the ball off his foot every once in awhile. If he did it in a close late game situation and we lost, here would come the critics calling Buddy a choker and offering that as proof Kruger can not coach. When in reality all that happened was that Buddy randomly bounced the ball off his foot.

Sure, our team needs to continue trying to improve their processes. We all do. I'm sure that they will. I'm also sure that many of our posters have some pretty flimsy small sample size evidence to be so critical.

The Drunkards Walk: How Randomness Rules Our Lives. Leonard Miodinow
Fooled By Randomness: The Hidden Role of Chance Nassim Nicholas Taleb

These are just a couple of the books that have helped shape my opinion on this topic.

 

[Kewllane]

I've been saying this for a while now. If it's close at all, OU usually doesn't pick up the win. That's why if you look at an OU spread in Vegas and you see it less than 6 or 7, then its basically picking if you think OU will win or not. If they win, they'll cover that easily. If they don't cover that, they're probably going to lose. Might as well take the moneyline.

But still, it is frustrating that they can't seem to win a close one. Unfortunately, I don't see them making the Sweet 16 without winning at least one close game. Can't beat everyone by 10 or more.

 

[Escape Goat]

Bump. Hope this turns out better.


Sent from my iPhone using Tapatalk

 

[Escape Goat]

Finally!!


Sent from my iPhone using Tapatalk