Archive for small sample normalization services

#3: May 3
#2: April 24
#1: April 13

* * *

It’s episode No. 4 of my Small-Sample Normalization Services, which, in Star Wars terms, means this will be, like… the 3rd-best post of the series? Is that how that works? I know to nothing about what’s believed to be the consensus on the merits of each film. I already regret making this stupid comparison.

Allow me, then, to touch upon (and revisit) some players whose performances through six weeks are worth critiquing. Six weeks is still a considerably small sample when it takes hundreds of plate appearances (or, for ball-in-play metrics, batted balls) for certain standard and advanced metrics to become reliable (or, in common but sometimes misused parlance, “to stabilize”). Check previous posts for the rules, but know that a rating of 1 means Hype City and a rating of 5 means, uh, Alarm City. A 3, therefore, would be neutral.

All graphs pulled prior to yesterday’s games.

* * *

Name: Masahiro Tanaka, NYY SP
Read the rest of this entry »

#2: April 24
#1: April 13

If you’ve tuned in before, you know what this is about. If not: the Small-Sample Normalization Service (SSNS) seeks to, ah, normalize a player’s performance in the context of his own previous achievements (or lack thereof). Most of us are human, and our humanity leaves us vulnerable to the biases that cloud rational thought and critical analysis. Such vulnerability is eagerly exploited by the small sample size, never more so than in April. While midseason small samples cower under the cover of hundreds more plate appearances, April performances have no such luxury.

A month’s worth of playing time is certainly more worthwhile to assess than one week’s worth, but 30 innings or 100 plate appearances can still be pretty volatile. Here are a few still-small samples that recently caught my eye.

All graphs pulled prior to yesterday’s games.

Name: Byron Buxton, MIN OF
Read the rest of this entry »

By clicking this link, you (“Reader”) have opted into an agreement (“Contract”) with FanGraphs, Inc. (“Handsome Author”). Handsome Author agrees to provide Small-Sample Normalization Services (SSNS) to Reader in the following post (“Post”). In return, Reader, presumably interested in Handsome Author’s analysis or merely intrigued by Handsome Author’s curious Post title, shall appreciate said services no matter what.

SSNS seeks to normalize good and bad performances witnessed in the first two weeks of the 2017 Major League Baseball (MLB) season. Handsome Author has noted previously, here and elsewhere, that small-sample booms and busts in March and April would go largely unnoticed in other months in which the sport of professional baseball is played, such as May, June, July, August, or even September.

Accordingly, SSNS looks at a player’s past performance as a benchmark for current performance using FanGraphs’ (and not Handsome Author’s) very nifty player graphs. It answers the question, “Has a player done this before?” Perhaps, Reader. Perhaps. But perhaps not. SSNS then assigns an Excitement-to-Panic Level (EPL) on a 5-point scale from 1 to 5 as well as an Adjusted Excitement-to-Panic Level (AEPL) once Handsome Author has properly assessed the historical significance of the player’s performance — within the context of the player’s self.

In this inaugural edition, Handsome Author will use SSNS to evaluate five hitters primarily in terms of their strikeouts (K%) and walks (BB%) through their first X number of games, with some other statistics incorporated as well. SSNS is not the be-all, end-all of player performance, but knowing we’ve seen a player “do this before,” as they say, is enough to calm one’s turbulent heart and mind.

Read the rest of this entry »