Churchill: Matzek impressing

[allstar1579]

Actually the stats would prove this theory incorrect. HS arms have been more successful than college arms in the top 5 picks of the draft in recent history. HS coaches can abuse guys just like college coaches, but it seems the college coaches really seem to abuse their top arms, especially when they are competing and that leads to arm problems for those college guys later down the road.

Beau Hale is a perfect example of this, ditto for Wade Townsend. I'd much rather take Matzek or Purke than White, Crow or Gibson...

See the big fallacy with your hypothesis is that there are more failures with college arms in the top 5 because there are much more often college arms taken that high than HS arms. There have been 13 college pitchers taken in the top 5 compared to just 2 HS arms since 2002. With the failure rate of first round picks being something like 60% then of course you are going to have more college failures than HS.

Of those 13 college arms, 6 still have a good chance at success, it's too early to pass the "bust" title on them. Neimann is a maybe for me, but he's not a complete bust or anything, and even though Verlander has been all over the place lately, he's still got tons of potential and might just need a change of scenery.

[JTrea81]

See the big fallacy with your hypothesis is that there are more failures with college arms in the top 5 because there are much more often college arms taken that high than HS arms. There have been 13 college pitchers taken in the top 5 compared to just 2 HS arms since 2002. With the failure rate of first round picks being something like 60% then of course you are going to have more college failures than HS.

Of those 13 college arms, 6 still have a good chance at success, it's too early to pass the "bust" title on them. Neimann is a maybe for me, but he's not a complete bust or anything, and even though Verlander has been all over the place lately, he's still got tons of potential and might just need a change of scenery.

Meant to say top 10. That is a larger sample size and you can see the findings of Gregg Pappas:

Boswell's research coincides with my very own, shared here at the OH some time ago.

I used a ten year period from 1993-2002 to determine the success rates of the top ten draft picks from each year.

College Hitter > 69% (11/16)

HS Hitter > 63% (17/27)

HS Pitcher > 35% (8/23)

College Pitcher > 34% (11/32)

Conclusion: Hitters are nearly twice as likely to be better choices.

HS pitchers are slightly less of a risk than college pitchers, but both are riskier picks than hitters, especially college hitters that is for sure.

[Lucky Jim]

Again, this is simplistic.

Nothing has been proven and there is no theory.

This is a hypothesis that is being tested by combining draft set populations where the criteria has definitely changed. To paint all high school pitchers or all college pitchers as being a better or worse bet is nonsense. It is also nonsense to saying high school or college batters are more valuable than pitchers.

Eating plants is more dangerous than eating meat because some plants are naturally poisonous to us. So should we just ignore all plants?

Vegetables are more likely to be poisonous to us than fruit, so should we ignore all vegetables?

Can you see how this is simplistic? Generically characterizing a vast population is not useful. It may provide some context, but it does not help much when trying to make a decision.

And not just the criteria. Also the means of evaluating, info about biomechanics and workload, etc.

The difference between a bust in 1993-2002 and now may be only a marginal increase in our understanding of how to (1) spot flaws in pitching prospects; and (2) take care of that prospect once drafted.

The problem with an opinion derived from a success/failure comparison is that it treats the instances of success and failure in a vacuum. Its only value is in letting us know that it might be riskier now (might be) to draft a pitcher using 1993-2002 standards and then develop him using 1993-2002 standards than it would be to draft a position player using 1993-2002 standards and develop him using 1993-2002 standards.

Growing up in a small town, it was "riskier" to buy a foreign car than a domestic car. The day that the small town got a German auto mechanic, however, that risk changed.

As long as the Orioles have a German auto mechanic around, I'm not so worried about 1993-2002.

[crawdad]

Meant to say top 10. That is a larger sample size and you can see the findings of Gregg Pappas:

HS pitchers are slightly less of a risk than college pitchers, but both are riskier picks than hitters, especially college hitters that is for sure.

You are misusing the statistics. It is not for sure because a college player is not a college player is not a college player. A high school player is not a high school player is not a high school player.

You are using an incredibly crude and flawed metric to determine value. I mean, if you honestly think the draft is a black box . . . then it would make sense to use the stat you are using. If you can discern differences between high risk and low risk talent . . . then it does not make any sense.

[Stotle]

This is no knock on Greg -- I appreciate his passion for the draft and ability to spark and contribute to draft talk these last two years -- but this is why I shuddered when he posted that "study". I believe I've seen that post referenced at least half-a-dozen times as proof that BAL should always draft hitters in the first round, and that just isn't true.

Jim and Craw have already debunked it, but I just wanted to reiterate what a bad idea it is to throw out individual analysis of players in favor of Greg's study. It is a good jumping off point for someone who wants to dig into a deeper look (the types of players that have succeeded vs. those that have not), but you can't stop with Greg's post. Not only is it incorrect, but it's being used by others (not Greg) to incorrectly frame draft issues to the point that people unfamiliar with the draft would actually come away from some of these threads with a much worse understanding of how things work than if they knew nothing at all about the Rule 4.

[JTrea81]

And not just the criteria. Also the means of evaluating, info about biomechanics and workload, etc.

The difference between a bust in 1993-2002 and now may be only a marginal increase in our understanding of how to (1) spot flaws in pitching prospects; and (2) take care of that prospect once drafted.

The problem with an opinion derived from a success/failure comparison is that it treats the instances of success and failure in a vacuum. Its only value is in letting us know that it might be riskier now (might be) to draft a pitcher using 1993-2002 standards and then develop him using 1993-2002 standards than it would be to draft a position player using 1993-2002 standards and develop him using 1993-2002 standards.

Growing up in a small town, it was "riskier" to buy a foreign car than a domestic car. The day that the small town got a German auto mechanic, however, that risk changed.

As long as the Orioles have a German auto mechanic around, I'm not so worried about 1993-2002.

Boswell also evaluated this in his article and came up with the same conclusion and he expanded his selection set to include all pitchers drafted in the top 5 since 1965.

If you take a larger sample size, the evidence is even more conclusive. Since '65, 102 pitchers have been taken within the first five picks. Not one is going to the Hall of Fame. None is close. Only one won more than 200 games (Kevin Brown). Rounding out the top five -- Dwight Gooden (194 wins), Bill Gullickson, Moore and Benes. The only reliever of note: ex-Oriole Gregg Olson. Josh Beckett (89-62) may end up high on the list eventually.

So top draft prospect pitchers in general are far less likely to pan out than hitters, no matter if they are HS or college. And this is going back all the way to 1965. The Orioles have experienced this first hand by not having any success in drafting a pitcher in the first round since Mike Mussina in 1990, let alone a pitcher in the top 10 or 5.

[crawdad]

Boswell also evaluated this in his article and came up with the same conclusion and he expanded his selection set to include all pitchers drafted in the top 5 since 1965.

So top draft prospect pitchers in general are far less likely to pan out than hitters, no matter if they are HS or college. And this is going back all the way to 1965. The Orioles have experienced this first hand by not having any success in drafting a pitcher in the first round since Mike Mussina in 1990, let alone a pitcher in the top 10 or 5.

This is descriptive . . . completely descriptive. It does not go around trying to explain anything. I don't know why you keep insisting that it advances the conversation.

Underage drinking does not cause teenage pregnancy.

[Three Run Homer]

You are misusing the statistics. It is not for sure because a college player is not a college player is not a college player. A high school player is not a high school player is not a high school player.

You are using an incredibly crude and flawed metric to determine value. I mean, if you honestly think the draft is a black box . . . then it would make sense to use the stat you are using. If you can discern differences between high risk and low risk talent . . . then it does not make any sense.

He's not misusing the statistics. You and Stolte are willfully ignoring the statistics. 67 percent versus 33 percent is an enormous difference, and even with the small sample size I'd guess it's statistically significant.

Statheads sometimes drive scoutheads crazy because they think that stats are everything, which leads statheads to believe that guys like Jackie Rexrode or Jeremy Brown can be major leaguers even though the scouts correctly recognize that their physical limitations will eat them up at higher levels of competition.

But scoutheads drive statheads crazy when they just dismiss statistics the way that you are doing. Of course the numbers don't mean that no team should ever pick a pitcher--every case is different, and no responsible statistician would say otherwise. But you can't just blow off the numbers. They are what they are. Facts are facts.

[Lucky Jim]

Boswell also evaluated this in his article and came up with the same conclusion and he expanded his selection set to include all pitchers drafted in the top 5 since 1965.

So top draft prospect pitchers in general are far less likely to pan out than hitters, no matter if they are HS or college. And this is going back all the way to 1965. The Orioles have experienced this first hand by not having any success in drafting a pitcher in the first round since Mike Mussina in 1990, let alone a pitcher in the top 10 or 5.

That you don't understand that Boswell's "evaluation" suffers from the same limitation as Greg's (and "limitation" is not a criticism - all analysis has limitations) is just evidence that you don't understand how to read and interpret the data.

The fact that you try to rebut me by "going all the way back to 1965" when I clearly state that marginal or incremental changes in evaluation and development could undo all prior mistakes (theoretically) only reinforces the fact.

[Lucky Jim]

He's not misusing the statistics. You and Stolte are willfully ignoring the statistics. 67 percent versus 33 percent is an enormous difference, and even with the small sample size I'd guess it's statistically significant.

Statheads sometimes drive scoutheads crazy because they think that stats are everything, which leads statheads to believe that guys like Jackie Rexrode or Jeremy Brown can be major leaguers even though the scouts correctly recognize that their physical limitations will eat them up at higher levels of competition.

But scoutheads drive statheads crazy when they just dismiss statistics the way that you are doing. Of course the numbers don't mean that no team should ever pick a pitcher--every case is different, and no responsible statistician would say otherwise. But you can't just blow off the numbers. They are what they are. Facts are facts.

He IS misusing the statistics. No one is "blowing off" the numbers. All anyone is saying is that you can't draw the conclusions from the numbers that Trea is attempting to draw.

The things that made pitchers risky in 1965 are different than the things that make them risky now. And the means of mitigating that risk have changed, too.

A guy failing in 1965 or 1993 has no bearing - none - on the risk of a guy evaluated and drafted in 2009, unless you allege that evaluating and developing is static.

Plenty of things that were risky in the past are no longer risky.

Think of improvements in safety for autmobiles. Cars are lighter and faster - and safer - than they were. And judging one's risk in driving based on fatality information from 1965 or even 1992 would be misguided.

[allstar1579]

He's not misusing the statistics. You and Stolte are willfully ignoring the statistics. 67 percent versus 33 percent is an enormous difference, and even with the small sample size I'd guess it's statistically significant.

Statheads sometimes drive scoutheads crazy because they think that stats are everything, which leads statheads to believe that guys like Jackie Rexrode or Jeremy Brown can be major leaguers even though the scouts correctly recognize that their physical limitations will eat them up at higher levels of competition.

But scoutheads drive statheads crazy when they just dismiss statistics the way that you are doing. Of course the numbers don't mean that no team should ever pick a pitcher--every case is different, and no responsible statistician would say otherwise. But you can't just blow off the numbers. They are what they are. Facts are facts.

That is entirely the problem, because he IS saying you should never take a pitcher from that sample size, so in fact by your own admission, he IS misusing statistics.

[crawdad]

He's not misusing the statistics. You and Stolte are willfully ignoring the statistics. 67 percent versus 33 percent is an enormous difference, and even with the small sample size I'd guess it's statistically significant.

Statheads sometimes drive scoutheads crazy because they think that stats are everything, which leads statheads to believe that guys like Jackie Rexrode or Jeremy Brown can be major leaguers even though the scouts correctly recognize that their physical limitations will eat them up at higher levels of competition.

But scoutheads drive statheads crazy when they just dismiss statistics the way that you are doing. Of course the numbers don't mean that no team should ever pick a pitcher--every case is different, and no responsible statistician would say otherwise. But you can't just blow off the numbers. They are what they are. Facts are facts.

This is funny. In the past three weeks I have been called a scout geek on four separate occasions. This amuses me because my background is statistical analysis. I understand statistics. I understand population experiments. I used to do a lot of work with ecological modeling. With this background behind me . . . I am saying that the simple descriptive analysis of 30 or 40 years of the draft does not tell you why it happened like that. It does not tell you how drafting philosophies have changed. It does not tell you about differences in team philosophies. It does not tell you about drafting for slot consideration.

It is grouping together a bunch of data that should not be grouped together.

This is not a scout-focused guy saying this . . . this is an environmental toxicologist with 10 years in the field dealing with population models telling you that your data set is pretty much useless trying to answer the question you are trying to answer. Get a better data set.

Also . . . 67 vs 33 is not an enormous difference. It can be, but that is not always the case. Again, this is a misunderstanding of statistical analysis.

[allstar1579]

This is funny. In the past three weeks I have been called a scout geek on four separate occasions. This amuses me because my background is statistical analysis. I understand statistics. I understand population experiments. I used to do a lot of work with ecological modeling. With this background behind me . . . I am saying that the simple descriptive analysis of 30 or 40 years of the draft does not tell you why it happened like that. It does not tell you how drafting philosophies have changed. It does not tell you about differences in team philosophies. It does not tell you about drafting for slot consideration.

It is grouping together a bunch of data that should not be grouped together.

This is not a scout-focused guy saying this . . . this is an environmental toxicologist with 10 years in the field dealing with population models telling you that your data set is pretty much useless trying to answer the question you are trying to answer. Get a better data set.

Also . . . 67 vs 33 is not an enormous difference. It can be, but that is not always the case. Again, this is a misunderstanding of statistical analysis.

Yeah, I'm the scout geek. Get it right.