No matter what you think the future holds for JP

[syhuang]

This is about how many points Losman helped the offense score in an average game last year, versus how many he helped it score in an average game this year. In other words, we're dealing with two measured means, and we're testing to see whether the difference in those two measured means is statistically significant. The correct tool to test for statistical significance in differences between measured means is the t-distribution.

817997[/snapback]

Why do you take the subject back to how we got the 7% or 15% improvement number on JP's year-to-year comparison when we're discussing how you set the standard? Shifting focus again?

 

The question is how you set the standard? You wrongfully use one player's individual game distribution to define year-to-year standard.

 

You still don't get it, right? To define statistical significance, you can use t-distribution. But you need to use the t-distribution on yearly data when comparing year-to-year performance. In year-to-year comparison, each data entry is yearly stats. On the t-distribution your mentioned, each data entry is a game. These two t-distributinos have totally different meanings and stop confusing one to another.

[Wraith]

This is about how many points Losman helped the offense score in an average game last year, versus how many he helped it score in an average game this year. In other words, we're dealing with two measured means, and we're testing to see whether the difference in those two measured means is statistically significant. The correct tool to test for statistical significance in differences between measured means is the t-distribution.

817997[/snapback]

 

I have to agree. I don't think using the t distribution here is inappropriate. I'm not sure what having the NFL standard would do for you. The test of mean using the t distribution (I call it the T Test) is dependant only on the sample size n, the variability seen within that sample of size n, and the specified confidence interval.

 

The only problem with the t test, as I mentioned in my previous post is that less sensitive to subtle but definite shifts in the mean over a small sample size.

[Orton's Arm]

In the case of J.P. Losman, it hasn't.

Are the numbers you cited for Losman's performance in 2005 for the whole year, or just for his second stint? I'm perfectly willing to agree that he's playing better than he did in his first stint of 2005. If you've got first stint 2005 data lumped into the totals you cited, he'll obviously look better in 2006. But the real question is: is he playing better now than he was in his second stint of 2005?

[Wraith]

Are the numbers you cited for Losman's performance in 2005 for the whole year, or just for his second stint? I'm perfectly willing to agree that he's playing better than he did in his first stint of 2005. If you've got first stint 2005 data lumped into the totals you cited, he'll obviously look better in 2006. But the real question is: is he playing better now than he was in his second stint of 2005?

818002[/snapback]

 

The breakdown by distance and down are for 2005 as whole versus 2006 as a whole. In general, the distinction between the two stints is important, I agree. However, in the case of completion percentage, it's not. A control chart of his completion percentage from Stint A to Stint B to 2006 shows no statistical differences from Stint A to Stint B and a shift in the mean at the beginning of 2006. The comparisons of the aggregate are appropriate.

[Orton's Arm]

The breakdown by distance and down are for 2005 as whole versus 2006 as a whole. In general, the distinction between the two stints is important, I agree. However, in the case of completion percentage, it's not. A control chart of his completion percentage from Stint A to Stint B to 2006 shows no statistical differences from Stint A to Stint B and a shift in the mean at the beginning of 2006. The comparisons of the aggregate are appropriate.

818004[/snapback]

My next question is this: how current is the 2006 chart? If you were to exclude his last game or two, the 2006 numbers would probably look a lot better.

[Wraith]

My next question is this: how current is the 2006 chart? If you were to exclude his last game or two, the 2006 numbers would probably look a lot better.

818012[/snapback]

 

Which chart? The statistics in the tables by distance and down are current and provided via the splits on ESPN.com.

 

The control chart is also up-to-date (takes about 30 seconds in JMP) and is my own creation. The last two games do not negatively impact his numbers at all (Detroit was right at the mean and NE was above it). CLARIFICATION: This particular control chart was compiled by me, they are a very well established statistical tool in general. :-P

[Dibs]

Which chart? The statistics in the tables by distance and down are current and provided via the splits on ESPN.com.

818016[/snapback]

WOW...I was going to ask where you got the stats from....I never knew they were there.

Thanks for the info.

 

BTW your line of.....

This is the progress that I see. I still see a QB that loses the ball at odd moments and who should learn when to throw the ball away. But anyone who argues against progress is simply ignoring facts.

......is spot on.

[Orton's Arm]

I just went back through the play-by-plays from 2005, to look at the drives some people want to eliminate from Losman's second stint by using a 15 or 20 yard threshold. Below are the descriptions of the drives. I'll leave it to the reader to decide whether Losman did enough on those drives to deserve some credit for the points they produced. I personally feel he did.

 

Drive 1 (against San Diego): This drive started on the San Diego 47, and moved to the San Diego 35 for a FG. Losman passed for 11 yards, McGahee rushed for 6, and the Bills lost 5 due to a penalty.

 

Drive 2 (against San Diego): Losman completed a 9 yard pass to Reed on 2nd and 10. He completed a 3 yard TD pass to Shelton on 1st and goal from the 3. He also had a 30 yard run.

 

Drive 3 (against Miami): This drive started on the Miami 26, so I deducted 3 points because the Bills were already in FG range. McGahee had 9 rushing yards to Losman's 17 passing yards, including Losman's 4 yard TD pass to Evans. Using a 20 yard threshold, Losman gets no credit for this particular drive.

 

People are accusing me of dishonesty and manipulating statistics because I gave Losman credit for those three drives.

[syhuang]

I'll leave it to the reader to decide whether Losman did enough on those drives to deserve some credit for the points they produced. I personally feel he did.

820344[/snapback]

Wow, when people follow your designed rules and find your method produces inconsistent results, you want to change your rules again. Instead of using your hard threshold, you want to use judgement calls now!?!?

 

If you want to use judgement calls, stop creating rules to pretend your results are "fair". It's ok to analyze the drive details, but you need to apply it on all drives, not just the ones don't fit into your "rules". For example, on field goals, why do you give same credits when the drives started at opponent's 47 yards and at own 20 yards? Everyone watches football knows these have very different difficulties. Or why don't you give partial credits to JP when his pass brings a field goal attempt from 52 yards to 44 yards?

 

You can keep modifying your rules and bringing in personal judgements to adjust the results. But at least you can avoid manipulating stats by applying your rules fairly. Stop contradicting yourself.

 

You said you deducted 3 points from Drive 3 in your last post and this is not true. In your original post, you have JP contributed 12.2 points per game in his 2nd stint last season, it's 61 points totally.

 

The scores of the 5 games in JP's second stint last season:

 

KC: Bills scored 14 points. You give all credits to JP

SD: Bills scored 10 points. You give all credits to JP

CAR: Bills scored 9 points. You give all credits to JP

MIA: Bills scored 23 points, including one safety. You give 21 points to JP (never deduct 3 points as mentioned in your rules and your last post)

NE: Bills scored 7 points. You give all credits to JP

 

Total: Bills scored 63 points

 

How come you give JP a total credit of 61 ponits if you deducted three points from Drive 3 and two points from a safety?

 

I guess you'll say that the results don't affect your conclusion since You already design the rules to keep JP's credit last season as much as possible. You design the rules to omit some of JP's stats this season, but when the same should be applied to his 2005 stats by your rules, you choose to ignore it.

 

Everyone can create numbers like you did by creating his own rules (and even bringing his personal judgements into the rules).

 

You make up your own rules to omit the stats which don't support your arguments. You can not even apply your rules fairly. When your rules don't produce the numbers you like by using your hard threshold, you start to adjust them and bring in personal judgements. Sure, it's ok to bring in personal judgements, but stop acting like your opinions have fair stats to back up. It has the same effort by simply saying "I think JP hasn't progressed".

[Orton's Arm]

Wow, when people follow your designed rules and find your method produces inconsistent results, you want to change your rules again. Instead of using your hard threshold, you want to use judgement calls now!?!?

Both your numbers and your conclusions are incorrect. First, the numbers. I gave Losman (2005, v2) credit for the following:

- KC game: 11 points (2 TD drives, minus three points for a drive that started in FG range)

- SD game: 10 points

- Carolina game: 9 points

- Miami game: 18 points (3 TDs, minus three points for a drive that started in FG range)

- NE game: 7 points

 

Total: 55 points.

Then I divided by 4.5, because Losman had played in four and a half football games. Had Losman been the Bills' QB for both halves of the KC game, I would have divided by five. 55 points/4.5 games = 12.2 points per game.

 

Onto the logic portion of your post. Earlier, you made various accusations against me, because I chose a 10 yard threshold instead of a 15 or 20 yard threshold. I responded by asking people to look at the drives from 2005 that your 15 yard and 20 yard thresholds would require be thrown out. Especially on the two TD drives, it was quite clear that Losman made a meaningful contribution. Adopting the 20 yard thresholds you mentioned would result in Losman (2005 v2) losing all credit for two TD passes and a 30 yard run. I do not feel my system should be changed with the intention of producing this result. Others may disagree, and in the end it is a judgement call. Under the rules of the 10 yard threshold, Losman v2 retains credit for both those TD passes, as well as the FG drive I mentioned.

 

You bring up a valid point in saying that it's easier to march 20 yards into FG range, than it is to march 60 - 70 yards to get that FG. I agree the system I've designed makes only a crude attempt to capture this distinction, by deducting 3 points from any scoring drive that begins inside FG range. However, the solutions to this problem you pose would result in a system that would be far more complex than my present system. Given that one of the loudest complaints leveled against my present system was its level of complexity (and implied statistical distortion), I was less than eager to double or triple the system's complexity.

 

I don't feel this one system is an end-all, be-all. It's just one way of looking at the numbers. You also have to look at other data, such as yards per attempt, TD/INT ratio, things like that. If all those numbers seem to tell the same story, then maybe that story is true. On the other hand, it's always possible that changes in the supporting cast or quality of defenses or something caused a distortion in the numbers. Is Losman playing better this year than in his second stint from last year? In the end, the question is a judgement call.

[syhuang]

Both your numbers and your conclusions are incorrect. First, the numbers. I gave Losman (2005, v2) credit for the following:

- KC game: 11 points (2 TD drives, minus three points for a drive that started in FG range)

- SD game: 10 points

- Carolina game: 9 points

- Miami game: 18 points (3 TDs, minus three points for a drive that started in FG range)

- NE game: 7 points

 

Total: 55 points.

Then I divided by 4.5, because Losman had played in four and a half football games. Had Losman been the Bills' QB for both halves of the KC game, I would have divided by five. 55 points/4.5 games = 12.2 points per game.

This is another example that your system has major flaws. Why do you only count JP played 0.5 game against KC last season?

 

If you consider number of pass attempts (JP: 16, Holcomb: 6), when Holcomb was injured (5 minutes into the 2nd quarter) and when was JP's first series (5 minutes left in 2nd quarter), JP clear played more than 0.5 game.

 

The results can be varied between 11.6 to 12.2 depending on which number you pick, and 0.5 is clearly too low.

 

 

Onto the logic portion of your post. Earlier, you made various accusations against me, because I chose a 10 yard threshold instead of a 15 or 20 yard threshold. I responded by asking people to look at the drives from 2005 that your 15 yard and 20 yard thresholds would require be thrown out. Especially on the two TD drives, it was quite clear that Losman made a meaningful contribution. Adopting the 20 yard thresholds you mentioned would result in Losman (2005 v2) losing all credit for two TD passes and a 30 yard run. I do not feel my system should be changed with the intention of producing this result. Others may disagree, and in the end it is a judgement call. Under the rules of the 10 yard threshold, Losman v2 retains credit for both those TD passes, as well as the FG drive I mentioned.

No, I just follow your original rules and state that with different thresholds, the results are inconsistent. You just pick a number which favors your argument and never mention the impact of different thresholds. You try to bring in judgement calls to retain the drvies which would be ommitted with different thresholds (by your rule). Anytime you bring in judgement calls, the results already lose any remaining meaning.

 

 

You bring up a valid point in saying that it's easier to march 20 yards into FG range, than it is to march 60 - 70 yards to get that FG. I agree the system I've designed makes only a crude attempt to capture this distinction, by deducting 3 points from any scoring drive that begins inside FG range. However, the solutions to this problem you pose would result in a system that would be far more complex than my present system. Given that one of the loudest complaints leveled against my present system was its level of complexity (and implied statistical distortion), I was less than eager to double or triple the system's complexity.

Good, you admit your system has problems. Complexity is not an excuse for your flawed system and flawed results. You can not ignore many factors and simplify the whole system, the results will not be meaningful. Using simplified and incomplete system with questionable thresholds can only get invalid results, it's not better than your personal opinions.

 

I don't feel this one system is an end-all, be-all. It's just one way of looking at the numbers. You also have to look at other data, such as yards per attempt, TD/INT ratio, things like that. If all those numbers seem to tell the same story, then maybe that story is true. On the other hand, it's always possible that changes in the supporting cast or quality of defenses or something caused a distortion in the numbers. Is Losman playing better this year than in his second stint from last year? In the end, the question is a judgement call.

820778[/snapback]

Don't shift the focus. I'm only talking about your questionable attempt to omit some stats which don't help your argument.

 

Everyone can design his own rules to omit the stats he doesn't want and gets the results he likes. Take your system as example, some people prefer 10 yards, some prefer 15 yards, some prefer 20 yards, some prefer using pass/rush yard ratio, and so on. People may think some drives should count, some drive shouldn't count. There is no standard on all of these and you basically just pick what you like to get the results benefit you. In the end, it's your own personal judgement, it shows nothing more than "I think ......".

[PacificCoastBillsFan]

But the Bills haven't been sustaining drives. I saw a statistic about the percentage of third downs that actually get converted when a team chooses to pass. Guess which team was at the very bottom of the list?

 

The problem I have with completion percentage is that you can game the system. If it's 3rd and 10, throw a 5 yard pass. That'll help your completion percentage and passer rating, but it won't do much to help the team.

 

817793[/snapback]

 

But if your OL isnt giving you time to put the ball down field, you take what you can get an hope your reciever gets some YAC, or JP could hold onto the ball and take a sack........Sorry I dont have any stats on how many 3rd & long presures there were in relationship to short yds passes. I will leave it up to all you statitions argue JP's stats, rankings, and how come not to mention what they really mean in relationship to this team.

 

All I know is I hope the OL shake up helps this TEAM, and they do some serious recruiting this off season!

[Orton's Arm]

If you consider number of pass attempts (JP: 16, Holcomb: 6), when Holcomb was injured (5 minutes into the 2nd quarter) and when was JP's first series (5 minutes left in 2nd quarter), JP clear played more than 0.5 game.

Since Losman came into the game with 5 minutes left in the second quarter, I redid the numbers with him getting charged for 35/60 of a game. As a result, his points per game fell from 12.2 down to 12. I hope this pleases you.

 

You just pick a number which favors your argument and never mention the impact of different thresholds.

Didn't I address this issue before? I feel like we're going in circles here.

You try to bring in judgement calls to retain the drvies which would be ommitted with different thresholds (by your rule). Anytime you bring in judgement calls, the results already lose any remaining meaning.

Wrong. There are always judgement calls. Throwing out outliers is to some extent a judgement call. Deciding on an acceptable alpha level is a judgement call. Which independent variables to test for is a judgement call.

 

Yes, it always pays to ask whether the wrong judgement call was made. I give you credit for doing that here. And yes, choosing 10 yards as a threshold is a judgement call. But after looking at the drives that would have been thrown out with a 15 or 20 yard threshold, I feel the 10 yard threshold was the right judgement call.

 

Do you feel differently? Do you feel that ignoring those two TD drives (including the 2 TD passes and the 30 yard run) would produce a more accurate depiction of Losman's contribution during his second stint from last year? If you do, feel free to argue your case. If not, kindly stop arguing with mine.

[syhuang]

Since Losman came into the game with 5 minutes left in the second quarter, I redid the numbers with him getting charged for 35/60 of a game. As a result, his points per game fell from 12.2 down to 12. I hope this pleases you.

No, it doesn't. Why do you only use the time JP entered the game after I listed other ways to calculate it? Why don't you use pass attempt ratio? why don't you use the combination of the time Holcomb injured and the time JP entered the game?

 

This is another judgement call. It's not related to the outliers you mentioned. This is a judgement call of setting up rule/system, not a judgement call of regarding a data entry as a outlier.

 

Again, you still pick the number which favors you most. This is not about what pleases who, it's about the whole system which benefits one side heavily. What the final numbers are is not the main problem, the main problems are your designed rules which omit the stats you don't like, the questionable thresholds, and your personal judgements throughout the whole process.

 

You just pick a number which favors your argument and never mention the impact of different thresholds.

Didn't I address this issue before? I feel like we're going in circles here.

I mean before I point the problem out. You never address this issue before being pointed out that different threshoulds generate inconsistent results. After I point out this issue, instead of following your original rules, you try to use personal judgements to retain the drives which would be omitted by a different threshold. These drives may or may not need to be ommited, however, they already reveal your rule/system is incomplete and have problems that you keep bringing in something new to "fix" it. (here "new" means something never be mentioned before in your rules)

 

Wrong. There are always judgement calls. Throwing out outliers is to some extent a judgement call. Deciding on an acceptable alpha level is a judgement call. Which independent variables to test for is a judgement call.

Don't try to use outliers to confuse the whole issue. An Outlier is a data entry which is quite different to rest of the data. It's irrelevant to our discussion of not applying your rules in some cases. If you insist, please explain why (only) these drives are so different to others to be considered outliers. If you want to use the passing yard and total drive yard ratio, why don't you use it on all drives?

 

Yes, it always pays to ask whether the wrong judgement call was made. I give you credit for doing that here. And yes, choosing 10 yards as a threshold is a judgement call. But after looking at the drives that would have been thrown out with a 15 or 20 yard threshold, I feel the 10 yard threshold was the right judgement call.

It's your own personal judgement and doesn't represent most people's opinions. What you feel is right doesn't mean it's right.

 

As I said, you can simply say "I think..." instead of going through your designed rules to omit the stats you don't want and going through judgement calls of whether to apply your own rules to come up with manufactured stats.

 

Do you feel differently? Do you feel that ignoring those two TD drives (including the 2 TD passes and the 30 yard run) would produce a more accurate depiction of Losman's contribution during his second stint from last year? If you do, feel free to argue your case. If not, kindly stop arguing with mine.

820832[/snapback]

It's not about what I think, it's about your system which heavily benefits your argument and omits stats you don't like. As I mentioned, if you want to look at the drive details, you need to apply it on all drives instead of merely the ones don't benefit you. We already discuss this part which brings up the drive length and other issues. It also brings up the problem of over-simplified system you designed and the invalid results from your incomplete system.

[BillnutinHouston]

Bottem line is he doesnt win ball games weather hes "progressed"or Regressed.

817629[/snapback]

 

Doesn't the defense have a role in winning ball games?

[Orton's Arm]

No, it doesn't. Why do you only use the time JP entered the game after I listed other ways to calculate it? Why don't you use pass attempt ratio? why don't you use the combination of the time Holcomb injured and the time JP entered the game?

To some extent a QB creates his own pass attempt ratio. If you complete a 12 yard pass on 3rd and 10, you'll get more pass attempts than the guy who didn't make that completion. So in general, pass attempt ratio isn't a good way of dividing up games.

 

The other method you suggest--using a time halfway between the Holcomb injury and JP entering the game--seems reasonable. Doing that gets Losman v2 charged an extra five minutes, and lowers his adjusted points per game from 12 down to 11.8. I hope this pleases you.

 

What the final numbers are is not the main problem, the main problems are your designed rules which omit the stats you don't like, the questionable thresholds, and your personal judgements throughout the whole process.

The fact that you endlessly repeat this doesn't make it true.

 

After I point out this issue, instead of following your original rules, you try to use personal judgements to retain the drives which would be omitted by a different threshold. These drives may or may not need to be ommited, however, they already reveal your rule/system is incomplete and have problems that you keep bringing in something new to "fix" it.

Wrong! The ten yard threshold system I originally used retains the three drives in question. It doesn't require any modification or judgement calls or anything to keep those three drives in. Most people looking at those drives would feel the 10 yard threshold system produced the right result.

Don't try to use outliers to confuse the whole issue. An Outlier is a data entry which is quite different to rest of the data. It's irrelevant to our discussion of not applying your rules in some cases.

There is no discussion about whether I applied those rules in some cases or not. There are only two relevant facts: that I applied the 10 yard threshold consistently, and the fact that you don't understand this.

 

The sole point I was making in mentioning outliers, independent variables to test for, and alpha levels was to show that judgement calls will always be a part of statistics.

[syhuang]

To some extent a QB creates his own pass attempt ratio. If you complete a 12 yard pass on 3rd and 10, you'll get more pass attempts than the guy who didn't make that completion. So in general, pass attempt ratio isn't a good way of dividing up games.

 

The other method you suggest--using a time halfway between the Holcomb injury and JP entering the game--seems reasonable. Doing that gets Losman v2 charged an extra five minutes, and lowers his adjusted points per game from 12 down to 11.8. I hope this pleases you.

Didn't I tell you that this's not about who pleases who? This is about your incomplete system and your invalid results

Wrong! The ten yard threshold system I originally used retains the three drives in question. It doesn't require any modification or judgement calls or anything to keep those three drives in. Most people looking at those drives would feel the 10 yard threshold system produced the right result.

You're wrong again. You did try to retain these three drves after being pointed out that using different thresholds will have inconsistent results by your own rules.

 

Can you read "different thresholds"?

 

Don't try to use outliers to confuse the whole issue. An Outlier is a data entry which is quite different to rest of the data. It's irrelevant to our discussion of not applying your rules in some cases.

There is no discussion about whether I applied those rules in some cases or not. There are only two relevant facts: that I applied the 10 yard threshold consistently, and the fact that you don't understand this.

You not only omit stats but also omit facts.

 

Why did you only mention your 10 yard threshold and forget to mention other thresholds which give you inconsistent results? You indeed tried to retain the data entries which should be ommited by different thresholds based on your original rules. You keep avoiding your questionable thresholds and using your personal judgements on several key problems.

 

The sole point I was making in mentioning outliers, independent variables to test for, and alpha levels was to show that judgement calls will always be a part of statistics.

821165[/snapback]

It seems like you alway fail to respond to your incomplete system and your invalid results from yoru over-simplified rules.

 

Everyone can create any numbers he likes by

(1) Create his own rules to omit the stats he doesn't want

(2) Pick the threshold to favor him most

(3) Use personal judgement to retain the stats which should be ommitted by different thresholds based on his rules

(4) Simplify the whole system to favor his argement

(5) Get the manufactored numbers benefit his opinion

 

These manufactored numbers are useless and show nothing more than "I think ......".

[Orton's Arm]

Everyone can create any numbers he likes by

(1) Create his own rules to omit the numbers he doesn't want

(2) Pick the threshold to favor him most

(3) Use personal judgement to retain the stats which should be ommitted by his rules

(4) Simplify the whole system to favor his argement

(5) Get the manufactored numbers benefit his opinion

 

These manufactored numbers are useless and show nothing more than "I think ......".

None of this is true, and the bolded part is egregiously false. I'm done arguing with you.

[syhuang]

None of this is true, and the bolded part is egregiously false. I'm done arguing with you.

821184[/snapback]

All of above are facts. You not only omit stats but also omit facts. You did try to use your personal judgements to retain the stats which should be ommited by different thresholds based on your original rules.

 

Sure, you're running away after you cannot rationalize your questionable threshold, over-simplified method, incomplete system, and most importantly, your manufactored results.

 

Let me repeat:

 

Everyone can create any numbers he likes by

(1) Create his own rules to omit the stats he doesn't want

(2) Pick the threshold to favor him most

(3) Use personal judgement to retain the stats which should be ommitted by different thresholds based on his rules

(4) Simplify the whole system to favor his argement

(5) Get the manufactured numbers benefit his opinion

 

These manufactured numbers are useless and show nothing more than "I think ......".

[Ramius]

All of above are facts. You not only omit stats but also omit facts. You did try to use your personal judgements to retain the stats which should be ommited by different thresholds based on your original rules.

 

Sure, you're running away after you cannot rationalize your questionable threshold, over-simplified method, incomplete system, and most importantly, your manufactored results.

 

Let me repeat:

 

Everyone can create any numbers he likes by

(1) Create his own rules to omit the stats he doesn't want

(2) Pick the threshold to favor him most

(3) Use personal judgement to retain the stats which should be ommitted by different thresholds based on his rules

(4) Simplify the whole system to favor his argement

(5) Get the manufactored numbers benefit his opinion

 

These manufactored numbers are useless and show nothing more than "I think ......".

821188[/snapback]

 

apparently, HA likes getting mopped up and down the floor in multiple threads...thats the only reason for his assinine factless intelligence-less rants