Hypothetical pitcher question

[Hambowen]

I am wondering which player everyone thinks would have the best ERA only based on the numbers I am giving you. Assume everything about them is the same inlcuding "luck" except for the numbers provided. These are not real player stats just numbers I made up.

Who would have a better ERA:
k/9--bb/9---hr/9
10----3------1.3----Player A
8-----4------0.8----Player B
7-----3------0.9----Player C
5-----2------1.0----Player D

Rank them, group them, or say they are all equal. Anyway you want to give your opinion is fine.

[AllstonRockCity]

Stealing Gary's thunder??

I'll go w/ Player A as the lowest and then I'd expect Players B,C, and D to all have similar numbers.

PlayerA has the highest K:BB and the highest K/9, though not sure how much of that would be mitigated by the fact that he also has the highest HR/9.

B-D all have a K:BB between 2 and 2.5 and a HR/9 right around 1. To me that makes them all the same for this exercise regardless of their actual k/9 and bb/9 rates.

[msugray]

Assume everything about them is the same

Who would have a better ERA:
k/9--bb/9---hr/9
10----3------1.3----Player A
8-----4------0.8----Player B
7-----3------0.9----Player C
5-----2------1.0----Player D

I'll take them in the order listed:

Code: Select all

k/9  bb/9   hr/9
10    3     1.3  Player A
8     4     0.8  Player B
7     3     0.9  Player C
5     2     1.0  Player D

This assumes, of course, your premise that all else is equal.

[Guest]

Player A- 4.02
Player B - 3.92
Player C - 4.02
Player D - 4.26

Thats what the ERAs would be with neutral luck.

A piece on this subject probably within the next couple days.

[cwk1963]

I was going to say:

A
B
C
D

Even after seeing the results from Gary, I'll still take Player A first. He has a better K:BB than player B and strikes out guys at a better rate; thus has more control over his own destiny. The difference in HR rate is roughly 1 HR every other game or so, inconsequential IMO.

[msugray]

Even after seeing the results from Gary, I'll still take Player A first. He has a better K:BB than player B and strikes out guys at a better rate; thus has more control over his own destiny. The difference in HR rate is roughly 1 HR every other game or so, inconsequential IMO.

I agree with cwk, but...........I am willing to be swayed be GaryJ if he can show me the data.
I believe those ERA estimates are accurate given the k, bb, and hr rates stated. I wonder, however, about using the data as a predictive measure.

I went and looked at 40-50 random pitchers of all types. Their hr/9 rates vary significantly from season to season. Is hr/9 a stable enough statistic to use for projection? I think the only pitchers that would be very predictable with respect the hr/9 are pitchers that are extreme with regard to GB/FB. That may also be a false assumption. I guess, I view hr/9 as an outcome of pitcher skill; not a skill of it's own. If a pitcher is giving up fewer HR as a result of giving up fewer fly balls, I view the ability of inducing ground balls the skill part and hr/9 a measured outcome of that skill. Chicken or the egg, I suppose?

Does GB/FB ratio when matched with k/9 and bb/9 have more or less correlation to ERA than using hr/9?

GaryJ, I await your findings.

[rotodog]

Now which one has most Potential upside if those are numbers form a second year pitcher going into his 3rd year?

I say A and hope the HR/9 gets right....

[Shyguy30]

The question was which pitcher would have the lower ERA, not which would you draft first. High strikeout totals don't necessarily correlate to low ERAs (see Perez, Oliver). Without reading any of the other responses, I was going to say a close group that ranked CABD.

[Todd Zola]

Hopefully I am not spoiling Gary's article, but FWIW, there are xERA (expected ERA) calculations that use these skills to come up with the xERA, which IF YOU BELIEVE IN THE FORMULAS, make the determination objective. When Gary says "normal luck", the assumption is hit rate and LOB% are normal. Hit rate is the BABIP. DIPS theory suggests MOST pitchers will settle around .290-.310. Extreme groundballers and knucklers are notable exceptions. LOB% measures the percentage of baserunners that are left on base, Baseball HQ has their own metric they call strand rate, LOB% is more in the public domain. The average is about 70%. Lucky pitchers have a LOB% above 75%. It is these factors (BABIP and LOB%) that add the subjective nature to the analysis.

[Guest]

Nothing being spoiled.

We have an xERA formula that will get written up this weekend that uses nothing but the indicators provided and produces an xERA that correlates at a level of over 0.5 R-squared which is pretty damn good considering the luck factors involved.

The problem with the question is that a 0.8 HR/9 pitcher and a 1.3 HR/9 pitcher won't have the same luck, the extra HR are killers.

An extra half a HR a game is worth 0.50 to ERA by itself. It's not inconsequential.

If the question asked about potential, obviously the thought would be different, as the high K rate might suggest a hope of lower future HR totals but that wasn't really the idea here.

MSU - other than a piece which teases out our xERA formula, what other evidence would you like? It's hard to exactly tease out the data because a high HR pitcher will have lower LOB% rates than a low HR pitcher on avg and therefore they benefit doubly by the low HR rate.

[babl]

now, what if we knew their names, would that persue you any different?...before i did some digging i wouldve ranked them BCAD (dont know if id have room on a roster for a 5 k/9 with a 2 hr/bb)

k/9---hr/bb--hr/9
10----3------1.3----Player A (one REALLY doesnt exist)
8-----4------0.8----Player B (cc sabathia, 8.3--4.2--.8)
7-----3------0.9----Player C (cliff lee, 6.7--2.8--.9)
5-----2------1.0----Player D (brandon looper, 5.0--2.2--1)

these were the closest guys i came up with to the numbers supplied.

[msugray]

"10----3------1.3----Player A (one REALLY doesnt exist)"

Fictitional pitcher resembles 2008 Scott Kazmir with with an extra walk (9.81 k/9, 4.14 bb/9, 1.36 hr/9).

[Guest]

Of course, that extra walk by itself is worth roughly .30 in ERA....

[Hambowen]

I basically asked the question to see where my thoughts were compared to others. I had them all ranked basically the same.

Gary's era projections did show me that I weight low walk guys too highly. The first 3 guys all have very similar era's with Player B having the lowest era even with the 4 walks per 9 and Player D with the lowest walk rate having the highest era because of his low K totals.

I need to rethink a few things based on people's thoughts here and Gary's projections. K's need to be weighted higher when I do things now.

[Guest]

Ham - have you read the pitching theory piece I did? I think it probably does as good a job as I can in expressing the value of each peripheral.

(and the issue with the players is more HR related than K related)

[Hambowen]

Yes I did read the article. I might need to re-read it however. I was trying to find a simple formula to weight those 3 stats into 1 number. When i did this however I did not weight things accurately and need to re-do it.

[Guest]

I promise that by mid-week, you'll have that formula, available for all Platinum Subscribers (gratuitous plug).

[Hambowen]

That would be amazing. The way my keeper list looks in one league i might be going hardocre Lima so the info would be real helpful.

[rotodog]

I promise that by mid-week, you'll have that formula, available for all Platinum Subscribers (gratuitous plug).

Gary,

Any thoughts about incorporating any batted ball data such as gb/fb % in the future? Or Swing and Miss percentages? The amount of information baseball folks are now recording is incredible. I just cant comprehend it all...

[Guest]

These are sort of different questions:

1. We have our own Independent ERA number - like DIPS/FIP etc which does what the original request was. So we have that in place now.

2. As to the second - as soon as there's enough data to test, as in several years we can back-test, and we get some comfort that there's some consistency year over year in the extra data, sure we'd use it. I agree that the data is voluminous, and very cool, but so much of it gets trapped in the other stats that I feel pretty comfortable with it. Again, we're getting to an R^2 over .5 using nothing but 3 statistics, if the others help us towards BABIP, we'll get there. The research is very promising at this point but I'm comfortable someone using our projections isn't at a disadvantage to that type of research, as we are incorprating batted ball results in our projections.

[rotodog]

Thanks for clarifying.. I didnt realize batted ball data was already in the formula...

[Guest]

Batted ball data isn't in the formula.

Batted ball data is in our projection set. The formula is a quick and dirty way to evaluate the three true outcomes peripherals.

However, many pitchers have varying BABIP's and we incorporate that in our projections. To treat a SP with a consistent .310 BABIP the same way we treat Jon Papelbon would be foolish.