The Pure Quality Start Thread

[calsnowskier]

– USATODAY.com

This is just one link that I found that is rather team-unspecific. It is from a few years back, though.

Basically, a SP got one point for each of the following...

1) Minimum of six innings.

2) Allow no more than an equal number of hits to the number of innings pitched.

3) Strikeouts must be no fewer than two less than his innings pitched.

4) Strike out at least twice as many batters as he walked.

5) Allow no more than one home run.

A score of 4 or 5 is generals considered "dominant" while a score of 0 or 1 is considered a "disaster". Using this stat as a percentage for a specific time period (Joe Schmoe had a 60% "dominant" and a 15% disaster" PQS for 2012 - 2014) or as a graph to show trending (x=start date; y=PQS score) could be very useful.

Unfortunately, this is a fringe stat (at best) and is not tracked anywhere that I have been able to find. There are a few collections here and there for specific teams for specific years, but I have not been able to locate any sites that track for the entire MLB over any kind of timeframe. I assume that with access to the MLB DB (I have not seen a copy of it in years), the stat could be retroactively added rather simply.

So given that clear disadvantage, I think it has potential as a much better version of the QS, while taking into account more of the modern metrics instead of the "old school" metrics.

It completely ignores ERA, while still focusing on IP. But it also essentially brings in the current big-boy stats like SLG (at least IRT HRs), K/BB and WHIP.

I still like the QS quite a bit. It is just a bit too basic to ever take as anything more than window dressing. The PQS could be a discussion starter when comparing two pitchers instead of a third or fourth point, which I think the QS would be.

 

[MilkSpiller22]

I am not a fan of this stat, of course it is because i have not seen it in action, but i see too many flaws with it... and wonder why it gives somethings such importance...

I also don't se the purpose of it, cause unlike QS i don't see this to be good at measuring consistency due to how the points can fluctuate so much from start to start...

 

[calsnowskier]

I am not a fan of this stat, of course it is because i have not seen it in action, but i see too many flaws with it... and wonder why it gives somethings such importance...

I also don't se the purpose of it, cause unlike QS i don't see this to be good at measuring consistency due to how the points can fluctuate so much from start to start...

How so? Can you give an scenario of how a "real world" disaster start would be rated as a good or dominant start by PQS? Or on the flip, how a "real world" dominant start could be rated as a disaster by PQS?


Let's look at each point individually...

1) 6 IP

I Think it safe to say that this is a standard that we can both agree with. Let's move on.

2) H/IP <= 1

This is a partial WHIP. This begins to look at the QUALITY of the start, not just the inning eating. The hit tool and the walk tool are different for a pitcher, so separating them out is a good idea, I think.

3) K >= IP-2

This begins to get into a bit of a gray area, granted. But it is a fact that if the other team strikes out, than at least your defense is not making errors. Also, you are not getting BABIP'd to death. K's are good, M'Kay...

4) K/BB >= 2

I call this one the "Johnathan Sanchez" point. Great, you are K'ing everyone in the league. But you are walking them twice, first. This ensures that you don't get rewarded for striking out the side, but walking in 4 runs.

5) HR <= 1

This is the most questionable of the points, IMHO. But it is only given if you were not plagued by the long ball. HRs can't be defensed. They are automatic runs scored. Nothing can counter them.

Scores of 2 or 3 are basically discarded when looking at a players season. You want to know how many (%) 4/5s and how many (%) 0/1s he had. That will tell you how effective he was. Each bad start is its own entity, and does not skew his tracking starts (like ERA or WHIP) based on one HORRENDOUS start. It is only one start with a score of 0. If a team loses a game 7-3 or 24-3, it is still just 1 loss. Should the guy who "took one for the team" have his entire season destroyed because if it?

 

[MilkSpiller22]

how about 6 innings, 0 Ks, 7 hits, 1 walk, 2 HRs, 2 ER

thats not a great outing, but is it really a 1?? but thats irrelevant since i just cherrypicked stats to make it look bad, my problem with it is that it doesn't show consistency...

Also it values strike out pitchers too much... how about low whip guys who don't strike out many, and walk a few... how about 3Ks and 2 walks... then they can never get those points...

also what is the difference between allowing a 2 run HR and 2 solo HRs?? here you are penalized for one and rewarded for the other??

 

[calsnowskier]

The game you mentioned would score a 1, and as you mentioned, is a cherry picked stat line (that's fine, that's what I asked for).

They pitcher was flirting with disaster the whole game. He gave up (likely) 2 lead off HRs and was likely aided by either stellar D or just simple good BABIP luck. While his stat line looked OK the next day, no one is leaving the yard saying "Damn, Schmoe was really dealing today!" In short, he wasn't fooling any batters.

K's are the stat du' jur (sp?). Sabers love them their BABIP. K pitchers avoid the BABIP swings, and avoid defensive variations. They also have the disadvantage of tending to not go deep in games due to pitch counts. That tends to be countered by the uber-specialized bullies in today's game. This is one of the reasons I think that the 6 IP standard used for both the QS and the PQS is appropriate for today's game.

Allowing a 2-run HR vs allowing two solo-shots is looked upon as more-favorable because of cluster-luck. The fact that there was a runner on base has already been accounted for in other points. The ball traveled the same distance. It doesn't matter if there was anyone on base. Is this over-simplified? Yes! But that is sabermetrics.

 

[MilkSpiller22]

another cherry picked

6 innings, 6 hits, 2 walks, 1 HR,6 Ks, and 6 ER(one bad inning) to be give a 5 for that?? how is that a dominant start?? that might be cherrypicked but it happens all the time that a pitcher has one terrible inning...

Again, i don't understand the purpose of this stat, i am sure it is because we never saw it in action... But it seems very different than QS... and does not measure consistency as well, once you play with points good games will beef up the point total or %, so it is not about how often it is about how good...

 

[calsnowskier]

This is a better line to "break" the stat.

Let's look at the game...

It looks like he went 6 SUPER strong innings (no hitter, 1-hitter?) then went out for the 7th and could not figure anything out.

Well, if Mattingly had any kind if bully, Kershaw would have been taken out of the game before having to eat the shit sandwich.

This line would likely not exist in the real world outside of playoff baseball. The SP DID have a good game. He just had to eat it due to the lack of a bully.

 

[MilkSpiller22]

This is a better line to "break" the stat.

Let's look at the game...

It looks like he went 6 SUPER strong innings (no hitter, 1-hitter?) then went out for the 7th and could not figure anything out.

Well, if Mattingly had any kind if bully, Kershaw would have been taken out of the game before having to eat the shit sandwich.

This line would likely not exist in the real world outside of playoff baseball. The SP DID have a good game. He just had to eat it due to the lack of a bully.

and at the end of the day he got the loss, while losing the QS...

 

[calsnowskier]

I am not understanding how you don't think this shows consistency. Can you expand?

Remember, you do not "count" this stat. You do not, at the end of the season, add up all his points and have a number. You would really only look at his rates, or MAYBE look at his average score. I think a very good use would be to apply his scores to a graph, then you can even turn it into a predictive stat.

I wish I had some totals to show you that are not Giants-centric (a few Giants bloggers have really championed this stat over the last few years, so there are some Giants seasonal studies on the inner-webz). The link I provided in the OP was from '11 and it pointed out that it was passing the eyeball test (the clear leaders were Verlander, Sebathia, Greinke, Halliday, Kershaw).

 

[calsnowskier]

and at the end of the day he got the loss, while losing the QS...

You are looking more results oriented than process oriented.

I take it you are not a saber-guy? (Neither am I, really, I just happen to like this stat)

 

[calsnowskier]

and at the end of the day he got the loss, while losing the QS...

Also, "the loss"?

Come on man, you are better than that.

 

[MilkSpiller22]

You are looking more results oriented than process oriented.

I take it you are not a saber-guy? (Neither am I, really, I just happen to like this stat)

no i am not a saber guy per se... I am a stat freak, but i create my own statistics... and often they look like sabremetric stats... the weights and measures are subjective, the raw data is not...

 

[MilkSpiller22]

Also, "the loss"?

Come on man, you are better than that.

again, i just don't see the purpose of this stat... it is not a good measure of consistency, and it does not talk about theoretical wins, so it is really nothing like QS... it basically is just a combination of specific statistics that someone valued highly...

 

[calsnowskier]

Again, I just don't get your "no measure of consistency" argument. I think it is a better judge than QS (and I like QS). If a guy goes out every game and does a 6/3, he will have 32 QSs at the end of the year. That's awesome, but he is not an Ace. He is not the guy you are going to start game 1 of the World Series.

At the same time, a guy goes out and gives a 9/4 every time out. He has helped his team MUCH more over the season (approx 100 extra IP over the season with a much lower ERA), but does not have a single QS to his name. A cherry picked scenario, but that is what we are doing...

 

[MilkSpiller22]

the reason QS is better at showing consistency than PQS is because QS is black and white, you are either quality or not... while PQS you are getting points, every DOMINANT start with skew the points, so it is not about how often you are good, it is more about how good you are in your games...

and like i brought up earlier, that one bad inning game, you could still have a Great PQS, but not have a good game... so basically PQS does not care about results...

But again, maybe i need to see this stat in action for a while, but i just don't see the story it is trying to tell...

 

[calsnowskier]

As you said in the other thread, no stat is perfect..

You have come up with a stat line that, while not "real", mimics a very recent, high-profile game. It is a scenario that needs a LOT of outside factors to happen (playoffs, no bullpen, bad manager). If ANY of those outside factors were not present, Kershaw would have pulled earlier, and his "story" for that game would have been completely different.

Sabermetrics ATTEMPTS to mute all the noise that is out there. Kershaws game WAS a good game. Not his best, but it was solid for 6 innings. The 7th came along and because it was a playoff game and Mattingly had zero faith in his bully, he left Kershaw out there to eat it. If even ONE of those noise factors were not there, he would have had a QS and there would not have been much argument about it being a strong to dominant start.

1) Regular season - The win in the game is less important than making sure Kershaw is right. Also, the bully needs to do its job, and the result is less important than seeing if the guy is actually able to do his job.

2) Bad bully - If the dodgers did not have a complete shit-show of a bully, than a floundering Kershaw would not have STILL been their best option in that situation. Mattingly would have been able to go to SOMEONE and preserve Kershaw.

3) Bad Manager - Granted points 1 and 2. But you don't let the best pitcher in the game wear that game like he did. 9 out of 10 managers would have gotten him out of there before what happened happened. Donnie is just THAT bad.

 

[MilkSpiller22]

As you said in the other thread, no stat is perfect..

You have come up with a stat line that, while not "real", mimics a very recent, high-profile game. It is a scenario that needs a LOT of outside factors to happen (playoffs, no bullpen, bad manager). If ANY of those outside factors were not present, Kershaw would have pulled earlier, and his "story" for that game would have been completely different.

Sabermetrics ATTEMPTS to mute all the noise that is out there. Kershaws game WAS a good game. Not his best, but it was solid for 6 innings. The 7th came along and because it was a playoff game and Mattingly had zero faith in his bully, he left Kershaw out there to eat it. If even ONE of those noise factors were not there, he would have had a QS and there would not have been much argument about it being a strong to dominant start.

1) Regular season - The win in the game is less important than making sure Kershaw is right. Also, the bully needs to do its job, and the result is less important than seeing if the guy is actually able to do his job.

2) Bad bully - If the dodgers did not have a complete shit-show of a bully, than a floundering Kershaw would not have STILL been their best option in that situation. Mattingly would have been able to go to SOMEONE and preserve Kershaw.

3) Bad Manager - Granted points 1 and 2. But you don't let the best pitcher in the game wear that game like he did. 9 out of 10 managers would have gotten him out of there before what happened happened. Donnie is just THAT bad.

then how about pitchers who have first inning jitters, lets up the 4+ runs and then shut out the rest of game... That bad inning is not always because of the bullpen and manager... point is , that this stat does not care about results it cares about someones subjective weights and measures...

 

[calsnowskier]

the reason QS is better at showing consistency than PQS is because QS is black and white, you are either quality or not... while PQS you are getting points, every DOMINANT start with skew the points, so it is not about how often you are good, it is more about how good you are in your games...

It would not skew the points. It would become a dominant that is added to the big picture.

 

[calsnowskier]

then how about pitchers who have first inning jitters, lets up the 4+ runs and then shut out the rest of game... That bad inning is not always because of the bullpen and manager... point is , that this stat does not care about results it cares about someones subjective weights and measures...

That game is not a QS, but potentially a 4 or 5 PQS. If it is a 4 or 5, though, he would have to be pretty damned awesome the rest of the way, and while it happens, it is VERY rare.

As for it not caring about results, agreed. But to the same extent as QS. The 6/3 is arbitrary (I think we are both OK with that, though), but a subjective limit.

Weighing Ks heavily and avoiding HRs is actually NOT subjective. Avoiding HRs favors GB pitchers (something most scouts look for) and K pitchers avoid BABIP and defensive variances.

 

[MilkSpiller22]

what i noticed about most sabremetrics is that it is not result based, it is talent based... And due to that it is actually best used as a predictor... But i give sabremetrics very little merit to answer any result question like who had the better season, who had the better game, or who had the better career...

For example BABIP, at the end of the day, do you really care about BABIP over BA or OBP, but if a player cuts down on strikeouts, then BABIP is a great indicator of BA...

Also due to the predicting aspect, i don't find them helpful with veterans, we know what they are...