8-day pitcher matchup tools

[Hoser$]

On subject table showing individual SP ratings for upcoming week -- I see a column headed "Last 5 PQS" which in past years I studied carefully. It is now BLANK. I know nobody has had 5 starts, but can't you show the PQS for the starts that HAVE been made? Thanks!

[RAY@HQ]

That will be added this week. We're almost done building all of the PQS logs.

[Hoser$]

Thank you Ray!

[gravy028]

Is there an update to PQS logs? I'm on the Daily Matchups and still see nothing as of today. Want to ensure I am not a total moron and looking in the wrong place or missing it somehow? I too rely on the PQS logs heavily! Thanks BHQ!

[RAY@HQ]

PQS logs are ready, I just need to post them on the site. Should be up this afternoon. And now that they're ready, we can get the Last 5 PQS popped into the SP Matchups tool too.

[gravy028]

Awesome, thanks Ray!!

[Lion]

Speaking of Matchups, at what point in the season are the pitcher ratings based primarily on outcomes and skills from the current season? Is it 5 PQS starts? A better reader than I probably knows the answer to this, but I was unable to discern it from this very erudite article: https://www.baseballhq.com/articles/...-matchups-tool.

[RAY@HQ]

It's basically this table:

Eventually we came to create a tiered projection, where we used a different formula based on the TBF in the last 30 days.

Tier 0: TBF30 < 10 Tier 1: 10 ≤ TBF30 < 50 Tier 2: 50 ≤ TBF30 < 100 Tier 3: TBF30 ≥ 100

Tier 0 used only the projected K% and park factor as input. The other tiers used all of our input parameters, weighted differently by tier, to produce an output formula.

So once a pitcher is at >100 BF over 30 days, you're mostly weighted on the in-season data as opposed to the projected data.

[AutomatedTeller]

I love this tool. Sometimes, it gives results that I cannot believe - like Kyle Hendricks being a green matchup this week vs the Braves. I'm not criticizing the tool or anything - I think there's a reasonable shot that the tool is seeing something I do not, which doesn't mean I'm streaming Kyle Hendricks vs the Braves

What I'm wondering is if BHQ has a way of tracking predictions vs actuals, particularly for outliers like this? (I don't even know how to define "outliers like this", so that may answer the question right there)

Is there value to that?

[RAY@HQ]

This question has been asked a few times over the years, and I'd love to answer it. It's one thing when the tool gives "expected" numbers and they work out. When it gives unexpected numbers, what's the track record? I'd love to answer that question, but haven't really come up with a good methodology to do so. Open to suggestions...

[WillinTO]

I can see that the tracking versus outliers is a methodological challenge. But Ray, how about an update on how good the tool is doing at predicting outcomes in general?

[RAY@HQ]

I can see that the tracking versus outliers is a methodological challenge. But Ray, how about an update on how good the tool is doing at predicting outcomes in general?

Yeah, it's been a while since I've done that, and definitely want to do it for this year. I think I'll wait until around the ASB for the data set to build up, and run some analysis then.