Since this blog is the continuation of 10+ years of emails, we have naturally developed our own language, as well as adopting measures and stats from other places. In order to allow new readers to have some idea of what is going on, I plan to keep a glossary of terms in this section. I don’t promise to keep it up-to-date, but if you email me that you want a definition of some term I use but don’t define, here is where I will post the answer.
I use pretty understandable abbreviations for typical stats: BA = batting average, OBP = on-base percentage, SLG = slugging average, OPS = on-base plus slugging, H = Hits, R = runs scored, RBI = runs batted in, HR = Home Runs, ERA = Earned Run Average, etc.
WAR – war is bad, as we all know. WAR, however, means Wins Above Replacement and is an overall measure of a player’s value, in a season or a career. The theory is that you can measure a player’s contributions, whether by hitting, fielding or pitching, and figure out how much he contributed to his team, above the amount that a replacement-level player would have contributed with the same amount of playing time. There are two flavors of WAR: one from Baseball-Reference.com (which I sometimes call bWAR) and one from FanGraphs.com (which I sometimes call fWAR). Usually when I just use WAR I am referring to fWAR.
Projected WAR – I sometimes try to estimate the value of a player in future years with something I call projected WAR. I am not always careful about this, if you see WAR for the future, this is what it is. I calculate it as follows: 3 times the most recent season plus 2 times the prior season plus the season before that, of whatever WAR I am using (bWAR or fWAR). That gives me a weighted WAR on a 3-2-1 scale (some others use 5-3-2). I then adjust it by the age of the target season, as follows: <26 multiply by 1.15, 26-28 multiply by 1.1, 29-32 multiply by 1 (no change), 33-35 multiply by .9, 36+ multiply by .85. If I am projecting multiple seasons going forward (say, for contract analysis), I just take projected WAR and multiply by the age factor above.
Win Shares (and Loss Shares) – these are the invention of Bill James, the father of modern baseball analysis (you can find them on bjol.com) and they are sort of like WAR. Each win share or loss share represents 1/3 of a win or loss, and they are theoretically adjusted to be above replacement, that is, a replacement-level player would theoretically earn no win shares, and would earn loss shares depending on playing time. The nice thing about win shares is that they give a W/L record, which looks kind of like a starting pitcher’s, giving some additional perspective. Thus one player might have a 2.0 WAR season, with a 10-10 win shares record, while another might have a 2.0 WAR season with a 7-1 win shares record.
Game Score (GS) – also a Bill James invention, Game Score (often GS on this site) is a calculation of how well a starting pitcher performed in a given game. 50 is the nominal average, though the real average varies a bit from season to season. It is NOT park-adjusted, so a GS from Petco is worth less than a GS from Wrigley, but I like it because it is easy. I use verbal ratings derived from game scores as follows: <20 Horrible, 20-29 Awful, 30-39 Bad, 40-49 Mediocre, 50-59 Decent, 60-69 Good, 70-79 Excellent, >79 Outstanding. I also use GS to determine the W/L record that a starter “deserves” by considering GS <40 to be a loss, 40-49 0.25 of a win, 50-59 0.75 of a win, and >59 to be a win. (How to calculate GS: start with 50. Add 1 for each out recorded. Add 2 for each inning FINISHED starting with the 5th. Add 1 for each strikeout. Subtract 1 for each walk. Subtract 2 for each hit allowed. Subtract 2 for each unearned run. Subtract 4 for each earned run. Easy!) (Shortcut: Step 1 is 50 + innings times 3 + innings after 4 times 2. Step 2 is (H+R+ER) times 2. Step 3 is strikeouts – walks. GS is step 1 minus step 2 plus step 3. Easier!)
Relief Ratings – Brian and I developed a language and a calculation for evaluating a relief outing. There are four values: Effective, Ineffective, yikes and YIKES! You don’t need to know how to calculate these values to understand what they mean, I suspect. If you want the details, here they are: Effective means more outs than baserunners, AND twice as many innings pitched as runs allowed. To calculate whether an outing qualifies as a yikes or YIKES! you total up the following good things (outs recorded and innings pitched in) and the following bad things (hits, walks, runs, earned runs [yes, earned runs count double], wins, losses, saves and blown saves). If the bad things are at least 5, and outnumber the good things, then the difference (bad – good) is the value of the yikes. If it is lower than 5, it is a yikes, and 5 or more is a YIKES! An ineffective outing is one which is not Effective, but doesn’t rise to be a yikes. Like 1/3 IP with 1 hit and no runs.
Relief Seasons – to get the value of a relief season by this method, you score each Effective as +1, each Ineffective as -1, each yikes as -2 and each YIKES as -3. A relief pitcher’s contribution is simply the sum of these scores. A relief pitcher’s rating is the ratio of his contribution to the number of outings in which he pitched, converted into a grade like in school: F to A+, on a scale I’ll put here someday, if I am not too lazy.
Report Card – I calculate report cards for teams, from time to time. I haven’t found an automated way to do this, so I don’t do it too often, as it is a bunch of work. There is a tab on the site for the latest report card, which should also tell you when it was calculated. No promises on frequency. The report card gives a letter grade for hitters, starters and relievers, and is meant to suggest the following: a .600 team is an A team. A .550 team is a B team. A .500 team is a C team. A .450 team is a D team. a .400 team is an F team. If your hitters are rated B, then you should be a .550 team if your pitching is average. Thus if you have .550 hitters, .550 starters and .550 relievers, you will be MUCH better than a .550 team. These report cards are EVALUATIVE not PREDICTIVE. They are intended to answer the question “What is working?”, NOT the question “Where am I likely to finish?”
FIP stands for Fielding Independent Pitching, and it is a way to take Strikeouts, Walks and Home Runs allowed and combine them into a number that looks like an ERA and is on the same scale. The theory is that pitchers have little control of balls in play, but lots of control of these three elements, and so this rates pitching better than the traditional stats. xFIP is adjusted FIP, adjusting for HR/fly ball rate, so that pitchers that are [theoretically] “lucky” in that balls stay in the yard, have that adjusted away. FIP and xFIP are available on FanGraphs.
ISO stands for isolated power, and is just SLG – BA. It is a better measure of true power than slugging average – Did Ty Cobb REALLY have more power than Mike Schmidt? No, but he had a better slugging average, though a much lower ISO.
Runs Created are an invention of Bill James (abbreviated RC) and are an attempt to improve upon Runs and RBI by removing the team context, and asking the question: how many runs SHOULD a player score/drive in if he has THESE hit elements (singles, homers, outs, etc.) and played on a neutral team? I am lazy, and so if I calculate the RC myself I use the most basic formula, which is (Times on base X Extra bases on hits) / Plate Appearances. Actually, I don’t use this at all, instead using OBP * SLG * AB which mathematically is the same thing. (Trust me – or don’t: write out how OBP and SLG are calculated, and compare above). RC are useful for making sense of otherwise odd ratios (Ichiro in 2010 had 214 hits and 74 runs scored, because the Mariner offense was so wretched). For real RC analysis I look up the RC on Bill James’ website, which is WAY more complicated but presumably more accurate, as well.
WPA – this stands for Win Probability Added, and is complicated. It is a product of FanGraphs, and attempts to measure how much each player contributed to an individual win or loss, play by play. That is, a player who comes up with his team down 3, bottom of the ninth, two outs, bases loaded and hits a HR geta about 0.9 of a win for that hit – 90% of the time the team loses that game, and his hit (all by itself) changed that to 100% win. Thus he added .9 to the win probatility for his team. It is possible to have a WPA for a given game of more than 1.0, but ultimately the WPA for the winning team sums to .5 and for the losing team to -.5. It is fun to play around with.
OPS+ and ERA+ – Baseball-Reference.com, perhaps my favorite website (besides this one, of course) takes a player’s OPS or ERA, adjusts it for the ballpark he plays in, and compares it to the league average, posting the result as OPS+ or ERA+. Thus 100 means exactly average, 120 means 20% over average, and 50 means 50% worse than average.