Calculating my own BPIs

[brian]

Before becoming an online member today I had put together my own spreadsheet which calculates different BPIs using the formulas from the Baseball Forcaster. I noticed that my numbers were different from Baseball HQ and I did a little searching to find out why. It seems the differing numbers I get are due to the FB/LB/GB data that I am getting from Baseball Prospectus' site (this is the only sight I know of that I can get the data I need without having to pay again somewhere else. I know you can get data here but that doesn't help me when I don't have a subscription to the site). It does not match up with HQ data.

For example, here is Chien-Ming Wang's data:

GB% - LD% - FB%
HQ 58 - 15 - 27
BP 57 - 21 - 20

BP doesn't add up to 100% because they also have a Popup% listed for starters and that is 2% for Wang. So, while my H%, S%, Ctl, Dom, Cmd, hr/9 data all match up with HQ my GB/LD/FB data is causing me to have a BPV that is off and an xERA that is off (my formula says Wang's xERA should be 3.34 and HQ says 3.95, my formula says Wangs BPV should be 55 and HQ says 52 as of 5/29 at 5 PM ET).

I guess my question is - is this a big deal/difference? For starters my numbers don't seem to far off from HQ but relievers it is a bit more dramatic. Dan Wheeler for example is a 145 BPV for me and a 2.31 xERA, HQ says 111 BPV and 3.20 xERA. A noticable difference.

I guess I can add my Popup% into the FB% but that still doesn't account for the difference in LD%. Any thoughts?

[RobR]

See this post for xERA discussions. I believe we're using a slightly different formula than the forecaster.

Off the cuff, I would guess that BP and HQ use different data providers for G/L/F data who use different criteria to distinguish GB from LD from FB. I think HardballTimes had an article a few years back about how subjective LD vs. FB can be. ESPN shows only G and F and has Wang with 93 GB and 39 FB, which are pretty close to our numbers of 91 and 43. Who knows what ESPN does with LD though? Are the differences as pronounced for many players or is Wang one of the largest outliers?

Out of curiosity, where does BP publish their G/L/F data?

[brian]

See this post for xERA discussions. I believe we're using a slightly different formula than the forecaster.

I'll check that out, thx.

Out of curiosity, where does BP publish their G/L/F data?

http://www.baseballprospectus.com/statistics/sortable/

Custom report drop-down -> League Batting/Pitching, select the cats you want, there you go.

[RAY@HQ]

RobR's got it right, I believe... different stat providers will reflect slightly different G/L/F numbers, for whatever reason. The discrepancies should be small enough that they don't really impact the BPI calculations to any significant degree.

[Michael@HQ]

You don't have to go through the long Boof Bonser thread if you don't care for the blow-by-blow account. The formula used by the website for xERA has been corrected, the on-line glossary defines xERA correctly, and you can look at the projections spreadsheets or the player link page to find the inputs to the formula.

I agree that you've got to use the same FB, GB, LD data to match BBHQ's results.