Starting Pitcher Report

[stembot]

Really appreciate the discussion on this one, too.

How much would it take to include predicted DOM alongside the daily rating figure? While Travis Wood is great at racking up solid PQS scores, his fantasy value is limited due to low strikeouts.

[RDeschain]

How much would it take to include predicted DOM alongside the daily rating figure?

Wouldn't that just be an exercise in futility? I would think it's impossible to accurately predict DOM on a game by game basis and even if you could it would be ridiculously inaccurate.

BUt maybe I'm wrong?

[Gantner]

It would be ridiculously inaccurate to predict Dom. On the other hand, the interplay of a pitcher's home/road Dom with an offensive team's home/road/handedness whiff splits is difficult to visualize without a formula. A "high/medium/low" strikeout favorability rating for a pitching matchup may be helpful.

[stembot]

It would be ridiculously inaccurate to predict Dom. On the other hand, the interplay of a pitcher's home/road Dom with an offensive team's home/road/handedness whiff splits is difficult to visualize without a formula. A "high/medium/low" strikeout favorability rating for a pitching matchup may be helpful.

Can we define the margin of error we're looking for here?

I'd like to think that all baseball player projections are inaccurate to a certain extent. I'm just looking for an edge in knowing when a given pitcher is likely to outperform his projection based on the matchup for that day. Even if it's small, it's something.

[TOLFG]

I wanted to confirm 2 things on the SP report....

1. As of May 1st, it pulls only data from the current year, regardless of sample size, correct?
2. How does it compute rating for an SP who didn't pitch in April due to injury or a rookie coming up? Is the rating for that 1st game only based on the other 3 factors and the SP gets a neutral 0 for his part of the formula?

[RAY@HQ]

1. Correct. Not sure of the exact date, but somewhere around May 1 we cut over to looking at current year data.

2. Also correct... for a pitcher on initial callup, or who does not have a home/road start as appropriate for the matchup, they get a zero for their part of the formula.

[TOLFG]

another question.....I understand from the article that the Last 5 PQS is not a part of the formula and just appears there as another tool to reference. As I play around in past reports, I notice the Last 5 PQS is actually pulling current data into the older reports, is this by design?

for example: 2012-05-01 shows Chris Sales with Last 5 PQS 5-0-5-5-5. More than 2 years later, today's report shows the same last 5. I checked and it's not coincidence, it's this way for other SP's too.

I would think the old reports should remain static, as they were that day in time. As I do research with them, it brings the more important question......is current data being pulled in other fields to the past reports? therefore rendering all past reports useless?

[RAY@HQ]

No, the past reports aren't static.

[TOLFG]

thx Ray. Do you know if it's just that last 5 PQS that is updated retroactively? or is it the formula pieces that make up the rating as well?

I guess I'm trying to understand why it would NOT be static. If we are using history as research to attempt to predict the future, don't we need history to remain exactly as it was at that point in time? or maybe I'm totally missing something here. You are good at explaining so thank you for all you do on these boards.

[RAY@HQ]

Think of it this way: when you put in a specific date in the report URL, all that you're really loading (I believe) that is specific to that day are the scheduled pitchers and game matchups. All of the rest of the material on that page, the data and stats, are read from our database as if the date in question is today.

This is one of the reasons I haven't been able to repeat the initial research from a few years ago about the SP Report's accuracy... because we don't archive the scores on the day of the game they apply to, so I don't have an accurate historical record to use as the research input.

[TOLFG]

darn. Yes, makes sense. I just didn't want to believe it that way, because then the past reports have no value to me. I appreciate you clarifying.