FWIW, Fangraphs Projects O's in Last Place in 2015

[DrungoHazewood]

He could be a cheaper version of Delmon, with a lot more upside, on this year's roster.

Probably about the same outfield defense, too.

[Phantom]

I realize that, but at some point you have to confront the reality that these two pitchers outperform their FIP every single year, and factor that in to your projections if you want the projections to be as good as they can be. I don't think the creators of Steamer are inclined to make individual exceptions to their methodology. That's fine, and I don't necessarily blame them, but in these two cases the chances are very good that their projections will be far off base.

The problem is, there have been a lot of cases where pitchers have "outperformed their FiP" over what you would assume to be a decent sample size, and then regressed to their Fip (Cain and Hellickson both did it in the same year, IIRC). So the question as always is, when is this performance "real" (or better yet, sustainable) and when is it luck (or better yet, unsustainable)?

If you are going to flip a coin 10 times, and the first four are heads, is it smarter to predict 10 heads, or that it will go 50/50 from then on out? Maybe the coin is weighted, but usually the smarter bet is on random variance. Now, I don't think all FiP outperforming is random variance, but I do think predicting it from season to season is an art, not a science.

[Number5]

The problem is, there have been a lot of cases where pitchers have "outperformed their FiP" over what you would assume to be a decent sample size, and then regressed to their Fip (Cain and Hellickson both did it in the same year, IIRC). So the question as always is, when is this performance "real" (or better yet, sustainable) and when is it luck (or better yet, unsustainable)?

If you are going to flip a coin 10 times, and the first four are heads, is it smarter to predict 10 heads, or that it will go 50/50 from then on out? Maybe the coin is weighted, but usually the smarter bet is on random variance. Now, I don't think all FiP outperforming is random variance, but I do think predicting it from season to season is an art, not a science.

You do know that FIP, by definition, is fielding independent, right?

Pitchers on good fielding teams outperforming their FIP is not random. They are not outliers. They are definitely projectable, and definitely to be expected.

Your coin flip scenario is pointless. If the Orioles' fielding is expected to once again be well above average, the pitchers will once again most probably outperform their FIPs by a good margin. Using FIP alone to project ERA is simply lazy on Fangraph's part and will almost certainly be very wrong in the Orioles' case yet again. Again, this is certainly not random variance.

[Frobby]

The problem is, there have been a lot of cases where pitchers have "outperformed their FiP" over what you would assume to be a decent sample size, and then regressed to their Fip (Cain and Hellickson both did it in the same year, IIRC). So the question as always is, when is this performance "real" (or better yet, sustainable) and when is it luck (or better yet, unsustainable)?

If you are going to flip a coin 10 times, and the first four are heads, is it smarter to predict 10 heads, or that it will go 50/50 from then on out? Maybe the coin is weighted, but usually the smarter bet is on random variance. Now, I don't think all FiP outperforming is random variance, but I do think predicting it from season to season is an art, not a science.

I can't disagree with you here. But I'd stress that baseball players aren't coins. There are a million reasons why someone's performance might not be captured in a model.

[Phantom]

I can't disagree with you here. But I'd stress that baseball players aren't coins. There are a million reasons why someone's performance might not be captured in a model.

Similarly, there are a million reasons why a coin might not flip perfectly. You might see that a side is dented for the tails side is a bit worn, but if I asked you to predict the next ten flips, you would still be smart to guess 5 and 5. And that's the point here. No one is saying all the things Fip or any other projection system don't factor in don't matter. All they are saying is that as of now, there is no good way to make better projections with them/

[Number5]

Similarly, there are a million reasons why a coin might not flip perfectly. You might see that a side is dented for the tails side is a bit worn, but if I asked you to predict the next ten flips, you would still be smart to guess 5 and 5. And that's the point here. No one is saying all the things Fip or any other projection system don't factor in don't matter. All they are saying is that as of now, there is no good way to make better projections with them/

FIP is not a projection system. FIP is useful to have an idea of the pitcher's part of the equation independent of the defensive team behind him. This helps in assessing how a pitcher on another team might perform if he were on your team, when factoring in your team's defense. When making projections for a team's ERA, the team's defense is a known factor. To intentionally ignore a known factor in projecting results is an obvious flaw.

[Phantom]

FIP is not a projection system. FIP is useful to have an idea of the pitcher's part of the equation independent of the defensive team behind him. This helps in assessing how a pitcher on another team might perform if he were on your team, when factoring in your team's defense.

Fip was not CREATED as a tool for projection, but it turns out it is far more accurate for predicting performance than other tools such as ERA. In fact, it's better at predicting ERA rest of season than ERA is, so I completely reject that its only or even best use is for predicting performance on another team.

When making projections for a team's ERA, the team's defense is a known factor. To intentionally ignore a known factor in projecting results is an obvious flaw.

If this were true, there would be another prediction system out there that incorporates defensive performance that is far more accurate.

[Phantom]

Sure there is a better way and the coin analogy isn't analogous at all. Do you believe the Orioles will continue to be a significantly better than average defensive team in 2015 as they were the last 3 seasons or not? If you do, then FIP should be adjusted. If you don't, say so and we will discuss that.

I believe that the Orioles team ERA will be much, much closer to their FiP than the last two seasons.

[MurphDogg]

I believe that the Orioles team ERA will be much, much closer to their FiP than the last two seasons.

Why, because they lost defensive studs Markakis and Cruz?

[waroriole]

Probably about the same outfield defense, too.

Exactly my thoughts. I was waiting for someone to say he doesn't play OF.

[CA-ORIOLE]

In the end the OP references a chart based on mathematical calculations and provides a projected performance based on that data. Part of that chart utilizes FIP. I see no reference to anyone indicating that anybody must follow this without considering other factors. EVERYBODY here seems to agree that defense (and other factors like park adjustments) are not accounted for in FIP. Hate it or not, I think everybody agrees that FIP is a stat that tries to isolate the value of a pitcher in a NEUTRAL environment and does not relate all the external factors related to actual run prevention.

I think Phantom is probably right about why the data for fielding (and other adjustments i.e for for park) are omitted. In the end, it's not clean data and is going to be more difficult to capture in a statistical model. Can you make such adjustments for these things? Yes. Bref does it, but even they do it as ongoing and after the fact, not as a projection.

It does seem as though the projection systems in general do not account for defense, but has Cameron (as much as I dislike him) or anybody else on the site stated that we should take the posted data of runs allowed numbers from FIP literally?

[Phantom]

Why, because they lost defensive studs Markakis and Cruz?

For a bunch of reasons, but here are two:

- The difference between our Fip and ERA was (largely) not due to our defensive prowess, so even if our defense stays exactly the same, we regress. There are a bunch of reasons to believe this including it sustaining even as our best defenders went down. Additionally, our pitching staff was composed of arms that were poorly constructed to take advantage of our stellar infield. Hell do people remember how awesome the Royals defense was last season? They outperformed their FiP by .15. I do not think our defense this season will be near as good as theirs last.

- Predicting defensive performance from year to year is very difficult, similar to relief pitcher performance.

[MurphDogg]

- Predicting defensive performance from year to year is very difficult, similar to relief pitcher performance.

I don't think this is true. Manny Machado isn't going to be Brooks Robinson one season and Mark Reynolds the next. I think that you are equating the relatively high variance in advanced defensive statistics with a high variance in defensive skill. The skill itself doesn't change.

[Phantom]

I don't think this is true. Manny Machado isn't going to be Brooks Robinson one season and Mark Reynolds the next. I think that you are equating the relatively high variance in advanced defensive statistics with a high variance in defensive skill. The skill itself doesn't change.

Fair enough. But since we are talking about assigning value to defense, and then using that value to make predictions, variance in skill and variance in the best numbers we have to measure that skill are basically the same.

In other words, my bad. Should have been more clear.

[Frobby]

For a bunch of reasons, but here are two:

- The difference between our Fip and ERA was (largely) not due to our defensive prowess, so even if our defense stays exactly the same, we regress. There are a bunch of reasons to believe this including it sustaining even as our best defenders went down. Additionally, our pitching staff was composed of arms that were poorly constructed to take advantage of our stellar infield. Hell do people remember how awesome the Royals defense was last season? They outperformed their FiP by .15. I do not think our defense this season will be near as good as theirs last.

- Predicting defensive performance from year to year is very difficult, similar to relief pitcher performance.

I guess I just feel FIP is very overrated. If I can go back to basics, pitching results have three components: skill of the pitcher, skill of the defense, and luck. I don't think FIP fully captures skill of the pitcher. Not all fly balls are alike. Not all ground balls are alike. I believe some pitchers are better than others at eliciting the kind of contact that leads to outs, and I don't think FIP captures it well. Maybe in another five years, with the advent of hit f/x and StatCast, someone will come up with a more sophisticated system than FIP. For now, I just use it as a tool to see if I should have any concerns about whether ERA results are sustainable, but I don't see it as a great predicter of future results.