12-11-2012, 08:45 PM #121
Honestly,.. I look at advanced metrics but I go back to the basics quite a bit because it all gets spun from the basics.
You got 32 more innings out of CC and 5 less earned runs allowed... so I'd say the difference between those two is fairly large. Plus 1.14 to 1.36 WHIP difference.
You got 47 less innings out of Marcum. Similiar WHIP and Similiar QS%. So the difference isn't that large.
As for adjusted ERA... Its just ERA weighted... When you get around the exact center (100)... That is where everyone congregates... It's going to heavily skew the percentage differences because of the volume of players in that average area.
Its a funnel.. Lots of average pitchers and it funnels upward to a much lower number of elite pitchers.
Last edited by Riverbrian; 12-11-2012 at 09:00 PM.
12-11-2012, 10:25 PM #122
Quality Starts are one of the most useless stats in the game. Any stat that can discard a superior start (9 IP, 4 ER) and reward an inferior one (6 IP, ER) should not be used for anything, ever.
12-11-2012, 11:27 PM #123
I will add this angle in support of what you are saying tho... You also have to consider the bullpen itself... A bullpen failure can cause the loss of quality start and it just may not be the fault of the starting pitcher... I'd guess that happens more often than the scenerio that you gave.
Its hard to find a perfect stat... They all need to be taken with a grain of salt and its part of the reason why I think we have an unnecessary lynching of Correia...
Quality Start percentage to me is used for one basic thing... How often did you keep the score reasonable... Its not really for how often you didn't in my opinion.
The didn't part is pretty iffy because... "Didn't" could mean you threw 1 inning and gave up 8 runs and it also could mean that you pitched 5 innings and gave up no runs or as you said 9 innings and 4 runs.
So I tend to look at the "Did" part and thats pretty rock solid... 6 innings and 3 runs or less meant you did well and your defense behind you did well... You gave your team a chance to win. Percentage of times that you did well is therefore reasonable... In my opinion of course.
12-12-2012, 12:45 AM #124
12-12-2012, 06:15 AM #125
9 IP, 4 ER is not a quality start. 6 IP, 3 ER is a quality start.
Which is the better start? QS is way too abstract and scattershot to ever be of much use. The creator of the stat just picked an arbitrary number and used that as the basis of... what, I don't even know. Just the fact that pitchers can go deeper into games with a better ERA and not get awarded the stat shows just how useless it is. Or that the bullpen can blow the "Quality Start" for the pitcher by letting a runner from first score and the stat is taken away from the starter. Or a half dozen other ways that it arbitrarily awards the stat and not others. It's like the save stat, except even dumber.
12-12-2012, 08:17 AM #126
So... In situations... When you are trying to show that Correia isn't as horrible as being portrayed... it shows that he gave you "at least" a decent start 57% of the time and I can point out how it compares to Greinke.
I think I understand your point Brock and yet don't fully understand it at the same time.
Is QS perfect? Not at all and for the reasons that are listed. But it isn't that wrecked either... I just went thru all of Dickeys and Greinkes start in 2012.
I found only one instance of either pitcher losing a QS past the 6 inning due to runs allowed late by himself or a bullpen. That was Greinke against the Twins on June 17. The Twins put 3 runs on Greinke with two outs in 7th after Greinke held the Twins to one earned thru 6. Thats 67 starts between them and only 1 example of a lost QS.
May 4... He gave up a pair of homers to Cozart and Stubbs in the 7th inning after reaching a quality start thru 6...
Another example of corrupted data would be the quick hook.
June 19... Correia against the Twins. Correia has thrown 5 innings of shutout ball. With a 2-0 lead... He retired Span to start the inning... Walked Revere who then stole second followed by a Willingham single. Hurdle then put Watson in to face Morneau who wiggled out of it and no quality start for Correia despite throwing 5.1 and no runs.
So yeah... There is the potential of false data but one thing is true... 6IP 3ER or less... Weather arbitrary or not. Is a traceable benchmark.
Name a stat that doesn't have flaws? They all do. That's why you should use them in combination with your eyes and never use them as some form of proof of anything.
BTW... I encourage everyone to go thru Correia's starts one by one last year. You will feel better.
Last edited by Riverbrian; 12-12-2012 at 08:25 AM.
12-12-2012, 08:45 AM #127
No stat is perfect but I can come up with at least a dozen that show more correlation to a pitcher's true performance than Quality Start.
So, at that point, you have to ask yourself: if there are better stats to show how well a pitcher pitched, why are you intentionally using an inferior statistic unless you want to cherry-pick stats to make one player look good and another look bad? Use the best tools you have to draw a conclusion... No need to intentionally hinder an analysis with flawed methodology.
12-12-2012, 09:25 AM #128
We all cherry pick Brock... I'm being honest... I went out of my way to go thru every start for 3 pitchers and find each flaw and post it. Did I not?
Are you sure I deserve that comment? Intentionally Hinder?
At no point did I ever say that QS% is a case closed stat. I'm saying its flawed and here's why...
However it does show that Correia did reach 6 IP and 3 runs or less at a similiar percentage to Greinke and Marcum and therefore he isn't worth the angst being shown.
I got the day off today... Nothing to do really so I guess I can go thru all MLB pitchers in 2012 and find examples of each such flaw in all pitchers and post them. So we can see the 3 percent variance and its effect to the overall impression.
I think you are expecting too much of stats. I don't know who came up with the arbitrary line of 6IP and 3ER. It is arbitrary but any pitcher who gets there... Had a quality start and gave your team a chance to win... If the arbitrary line was 7IP and 2ER... The percentages would change obviously but would the shuffling of the ranking. I don't know... But with a lot of work . You could find out.
12-12-2012, 09:29 AM #129
- Liked 535 Times in 354 Posts
I love the quality start stat. What matters for a start is did you give your team a good chance to win or not. Much better than ERA which blends every start together. Now, whether or not the precise definition is correct can he debated, but I find the stat highly useful for assessing performance.
Last edited by mike wants wins; 12-12-2012 at 09:35 AM.
12-12-2012, 09:50 AM #130
- Liked 340 Times in 216 Posts
Quality starts only consider earned runs. With a guy like Correia, he's going to be giving up a ton of unearned runs since he's a groundball pitcher with an awful defense. Who cares if the runs are earned or not, they still lose you the game. I'm still laying the loss on the pitcher since those runs could be avoided by keeping the ball off of the other teams bats.
12-12-2012, 10:02 AM #131
What if the pitcher plays in Yankee Stadium half his games and is right-handed? The place plays like a bandbox for lefty hitters and he gets to face the AL East in half his games and maybe the Yankees defense is awful. What if the guy pitches half his games in Dodger Stadium and then gets to face the often weak NL West hitters and what if the Dodgers are the best defensive team in baseball? What does QS really tell you about those pitchers? What about a guy who pitched in the late 90s versus a guy who pitches today when run scoring has dropped around a run per game? Continue ad nauseum. By using a small sample size (~30 starts) and by using a stat that is nearly 100% arbitrary (again, what is a "Quality Start"?), you are hindering your analysis for no good reason. You can take offense to that statement but you shouldn't... it's just the truth. Yes, guys will fluctuate from start to start but if you draw from a large enough sample size (say, ERA+, K/BB per 9, and IP), you'll get a much better idea of what kind of pitcher you're looking at. There's just no reason to look at the Quality Start stat. There are much better options out there.
12-12-2012, 10:03 AM #132
12-12-2012, 10:06 AM #133
Some great posts here. I tend think some of the disagreement is based upon different views of what the Twins should be trying to accomplish. I for one am disappointed in the signing because it shows the Twins don't seem to care about actually winning and that they really don't understand or care about acquiring the types of pitchers that can consistently help them win games. I think others are pretty satisfied that at least we got someone that has a major league resume and won't drool all over himself when on the mound.
I will say that based upon Correia's track record, there is nothing to suggest he's going to be better than what he was last year, and many more possibilities that he will be worse, the most important being the move to the AL and having to face 9 real hitters many nights.
12-12-2012, 10:22 AM #134
- Liked 1 Time in 1 Post
12-12-2012, 10:34 AM #135
ERA is flawed to a certain degree because of team defense, luck and ball park factors.. ERA plus is flawed because of defense factors and in my opinion it's dangerous in the hands of the majority of stat heads who don't understand the formula and the shades of percentage difference that appear larger when the results congregate around the middle. Any defense independent stat can be called into question for the simple reason that defense is a part of baseball and a large part of baseball.
They are all flawed and I love all of them... Personally... I think WHIP is as good as you get but the ability to get out of jams is something that can't be discounted...
There are no perfect stats... And that's the flat out beauty of baseball.
Last edited by Riverbrian; 12-12-2012 at 10:38 AM.
12-12-2012, 10:35 AM #136
- Liked 1 Time in 1 Post
Third and first struggled according to fangraphs. Justin Struggled last year with a UZR/150 of -3.9 after 6 straight years of positive values there. Odds are if he's right it will go back up. Trevor had the most innings at 3b, and showed a UZR/150 of -9.2
Florimon will likely get the most play at Short, where he was +5. I'm betting on Carroll everyday over at 2B were he was 9.4. Eduardo Escobar will be the utility guy and has shown great range at all three positions in his career.
This looks to be at least a +2 net infield, so basically neutral to above average for ground ball pitchers.
12-12-2012, 10:42 AM #137
12-12-2012, 10:54 AM #138
I just saw on another blog that Baseball Reference has a cool little tool that allows you to adjusts a players stats for a year to other teams/league. Here are the results when I put Correia's 2012 line, his BEST season in 3 years, into the 2012 Twins:
Looks pretty ugly and not worth investing 2 years in the guy. I know these just calculations/estimates, but they are based upon sound mathematics anyway.
2012 31 7 11 .389 4.91 163 186 98 89 21 49 85 3 1.442 32
12-12-2012, 11:08 AM #139
Slippery slope when you talk about more flawed than others. 6IP 3ER or less in each start... Did he get there or didn't he.
If he did... Mark 1... If he didn't... You can go thru each start and throw red flags on every instance that he or the bullpen gave up runs after reaching the required 6 innings or anytime he was pulled before reaching 6 innings. With the red flags... I believe that all pitchers will then have a margin of error of 3 to 6%. Although... I admit I would have to do a **** load of stat work to back that up.
You get to that 3 percent margin of error and it will be fairly consistent across the board. It won't be enough to discount it because we won't be talking about any subtracting of quality starts. Only addition in the event of giving up runs after reaching the goal or being pulled early. There may be some examples of pitchers that were screwed to the tune of 5 QS... I don't know... But I'm guessing the 5 QS screwed guys will be few. But... Most importantly... The Red Flags can be easily identified.
If its Important to you to differentiate between the quality start 6P 3ER or the Uber quality start. QS isn't going to satisfy you but you can then go to the other indicators.
Correia was alright last year and if you want to prove it statistically that he wasn't... You will have to cherry pick the stats that say He wasn't... Just like I can Cherry pick the stats that say he was.
12-12-2012, 11:13 AM #140
I don't know how that works...
Its quite possible that Correia will have a 4.91 in 2013. I won't argue that.
What happens when you run Edwin Jackson thru it?