Most fantasy managers evaluate talent on numbers like R, HR, RBI, SB and Ave, which they use to rank for their team in leagues. They consider comparison to former stats a good judge and good performance should be repeated.
The problem is the amount of luck in those stats (even OBP is luck since it takes Ave into account). I think analysis should be run on several factors that I discuss below.
Hitter
1. BB% and K%
These are 2 of the most important numbers in fantasy. Most leagues K% is most important as it directly influences Ave. BB% is a direct factor for OBP, which is less important, but OBP is important for R and SB.
2. HR/FB
Although for pitchers this number should revert to league average (10-12%). Hitters have a more variable number based on their power. Hitters like Ryan Howard will hit around 30% while Juan Pierre is destined to hold onto that 1%.
Judging how many HR a player is likely to hit is fairly straightforward based on HR/FB. This number can vary though and power growth can occur in the late 20’s, so it must be analyzed thoroughly.
3. Speed Score
This one is a Bill James invention and can get quite involved. There are several ways to use and calculate this. I include the double and triple calculation because my league uses total bases. In most leagues though you want the guy who will steal a base. Here is the calculation based on steal attempts: SQRT ((SB+CS)/(H-2B-3B-HR+BB+HBP))/0.07. This will give you a number between 1 and 10 with 10 being fastest. This number tells you how often a player attempts a steal when on first base and how much you can expect them to run.
4. LD%
This one is slightly less concrete as the others, but still something to keep in mind. I heard on The Fantasy Face-off Show last week that one of the guys if given only one stat to draft a team would ask for BABIP. I have to say that is quite ridiculous, because he will end up skipping a bunch of players who deserve their high BABIP due to high LD%. Players like Chone Figgins can maintain a LD% over 24% and that results in an xBABIP of .360. If you ignore him for his BABIP of .338 this year you may miss out on the best hitters.
PS…I would pick BB% for the one statistic I would like to see on offense. Anyone with an eye for getting league leading walks is bound to help. This includes leagues without OBP as well. The leaders in BB% usually have more power and get on base for steals and or runs.
Pitcher
1. K/BB
Speaking of one stat to choose, this would be my end all pitching stat. You can tell almost everything you need to know in this stat. If you don’t have the K/9 you won’t tell how good his K’s will be, but if you pick 4-5 starters with K/BB over 2.5 this year you should be leading in all things except saves. Closers will be better with a better K/BB, but saves obviously can’t be predicted. Luck will still have a say in all stats, but this stat can control your ERA and WHIP and put you in line for a good number of wins.
2. K/9
This stat is still an important factor in helping you win your strikeout category, but you don’t want to go for the guy just on his K/9. He may strikeout more than 8-9 per 9 innings, but if he walks 4-5 per 9 innings then he negates that value. Think Dice-K as he will get those ratios and you can tell right away with his 1.56 K/BB that he is going to hurt you eventually.
3. GB%
This is a great statistic to watch on pitchers who have borderline K/BB ratios. The more grounders you get the less fly balls that can turn in to homers at around 10-12% of the time. The number I generally target is a GB% of 50% or higher, but look for year-to-year changes. If a pitcher adds a new sinker or other pitch this year and goes from 35% to 45% then he is improving and that will help. This should give a few less hits and less fly balls to become homers.
4. Luck
Every year pitchers have extreme luck and you should review all of their stats to see how much helped their success. The following stats can insure you are viewing a true value: LOB% (~72%), HR/FB (10-12%), BABIP (.300) or LD% (18%). These stats should be near these numbers as they are out of the control of the pitcher. You can also check FIP to see if his ERA was repeatable based on current skills. Careful with FIP though if the pitcher has a HR/FB that is not 10-12% as it may vary from true value.
I hope this has helped many of our readers. I will be using these values to process players and project values for next year. Also use these to make evaluations on potential keepers for next year.
K/BB is great, but the only thing I caution is that given 2 pitchers with the same K/BB ratio, the pitcher with the higher K/9 should have the better ERA. Because of this, I think I lean toward K/9 or K% as the best stat, since I think control is the easiest skill to improve upon.
ReplyDeleteRight when they are close you need to go with the higher K/9, but you should look at K/BB first since that is more telling.
ReplyDeleteDice-K and King Felix have almost the same K/9, but I would take Felix with his higher ERA because he is a better K/BB.
I think we are saying the same thing, but getting to the same conclusion.
troy,
ReplyDeleteeach yr. I use your stats and others to build a very solid, but undervalued pitching staff. Anywhere to easily find the rankings of K/BB and K/9 stats for pitchers? Also, who are some undervalued late SPs to complete staffs with?
I use Fangraphs.com for my stats and if you click on leaderboards you can rank them on K/9 or K/BB.
ReplyDeleteI am starting my sleepers this week and will include many late round pitchers who have quality K/9 and K/BB ratios.
i've got some issues with the analysis here. and yes, i'm quite late to the party.
ReplyDelete1) my first is with saying that you'd choose BB% as the stat you'd choose if limited to only one. what you said about high BB% guys is generally true, but i feel it would not hold as well for the guys within one SD of the mean. further, i feel it misses a lot of very productive players on the low end of the spectrum
here are the ten lowest BB% players in 2009 (using batting title qualification as the AB cutoff on fangraphs) with their positional rank according to ESPN's fantasy player rater:
bengie molina (9)
miguel tejada (8)
christian guzman (22)
jeff francoeur (50)
jose lopez (13)
yuniesky betancourt (44)
robinson cano (3)
aj pierzynski (7)
kurt suzuki (4)
ichiro (5)
a csse could be made for every one of those players except betancourt to be rostered for a significant portion of the 2009 season.
ichiro and cano were straight up studs last year (not a shock, either), and suzuki put up a great year for a fantasy catcher
molina, tejada, and pierzynski all put up seasons worthy of starting at their position in a 12-team league, with lopez as a fringe option according to ESPN.
overall, that's 7 of 10 players that were very relevant in shallow fantasy leagues. only betancourt was largely irrelevant.
though far from a perfect solitary stat, my one simple stat that i'd look for would be Rs. not only is it one of the 5 standard batting stats, but it's correlated strongly to lineup position, which is itself correlated to all 5 batting stats.
2) i'm also not sold that all pitchers should regress to that park-adjusted 10-12% HR/FB%, either. i'd be quite confident betting that tim lincecum and clayton kershaw will post statistically significant park-adjusted HR/FB rates that are lower for the portion of their careers where they have similar stuff to what they have now. just because xFIP is the best predictor of future ERA across the population, doesn't mean that it's the best predictor for individuals or small sub-samples. i realize i'm cherry-picking players here, but that's my point- though it's not readily apparent, certain pitchers are better at inducing weaker contact, even for fly balls. overall, i could be wrong on this, but i feel sabermetricians are assuming too much when saying all pitchers will regress to the mean in HR/FB%.
3) this one is more a challenge to what i perceive as consensus opinion than a response to your post:
similar to #2, i think there's an overemphasis on GB% by sabermetricians, especially in it's application to fantasy. pitchers that are adept at changing a hitter's eye level (by throwing up in the zone at times) are better at accumulating strikeouts. i'm not saying GB% isn't a useful stat, just that it's only marginally correlated to fantasy success across the population in my opinion
Hey Andrew,
ReplyDeleteYour point is correct that those players deserved to be rostered and I didn't say I would like having only one stat, but I stand by BB% being the one I would use if I could only have one.
Your stat Rs is based largely on 2 things, lineup construction as you said and the ability to get on base. You are looking at a stat that a player only has half control of where I am looking at a large control of that number.
Your second point I really don't understand. You say you don't agree with Park-adjusted HR/FB yet your arguments are Tim Lincecum and Clayton Kershaw. They benefit from nothing under their control. Their low HR/FB is based completely by home park and league. They play a huge percentage of games in LA, San Fran and SD. That is the only reason they can maintain low HR/FB number. Put either in Philly and they top 11%.
On your last point I would again have to disagree. While I think it's important to have a good K/9 and a good GB% I wouldn't say giving up more flyballs is required. Look at Joel Pineiro this year. He just learned to pitch down and gain groundballs and he became a 3.27 FIP pitcher.
Even if his xFIP is right he is a 3.68 pitcher. This is one place a pitcher has been shown to decrease HR/FB as well. It's been shown GB pitchers can decrease BABIP and HR/FB.
regarding runs being my stat, i'd argue that a player DOES have a large amount of control over lineup placement. there are a couple exceptions (the inexplicable use of matt kemp last year), but by and large, managers will put their most fantasy productive hitters in spots where they will create runs. furthermore, runs as a category is a lot broader than BB%. as you said, runs incorporates a player's ability to get on base. BB% really only tells you if a hitter can get on, which as i noted with my unscientific bottom 10 in BB%, will exclude a lot of productive players.
ReplyDeleteregarding kershaw and lincecum, they had HR/FB rates of 4.1% and 5.5% respectively. their team rates were 8.1% and 9.1%. the data doesn't support your claim that their low HR/FB%s are wholly dependent upon park effects. the pattern holds true across each pitcher's career as well. i realize you can argue small sample size in each case, as combined they've got something equivalent to like 4 years of starting experience. my point, however, is that i don't understand why it is assumed that all pitchers will regress to that HR/FB% mean. it seems to me that the sabremetric community has taken xFIP's predictive ability in regards to future ERA to conclude that regression occurs the same for all individual players.
to further elaborate, take a pitcher, such as mariano rivera, that has a consistently low BABIP against and consistently low HR/FB rates. the assumption is that both of those rates should normalize. over the course of his career, however, they haven't. his career ERA is significantly lower than both his career FIP and xFIP (as you'd naturally expect).
it's become accepted that pitchers who can consistently induce a large amount of IFFBs will have lower BABIPs. Why isn't the same considered true for HR/FB%. the delineation between an IFFB and a normal fb isn't a consistent distinction (largely dependent on scorer's decision and player range). the most effective pitchers not only tend to have a lot of strikeouts, but they are able to induce weak contact a large percentage of the time. our current measures, though, (well-hit %s and IFFB%s) have a lot of room for improvement. my prediction, is that as our measures improve (hit f/x????), it will become more clear how some pitchers are better at inducing weaker contact, which would account for a lower HR/FB%.
the reason i cherrypicked lincecum and kershaw, is that i'd hypothesize their explosive FB/CB combinations are better at inducing weaker contact through the air, due to the dramatic shift in the batter's eye level between those pitches. they also maintain similar release points for each pitch, aiding in their pitches' deception. you actually demonstrate my point (albeit through a different type of pitcher) by saying that GB pitchers are able to reduce their BABIPs and HR/FB%s. normalization of HR/FB% does not apply equally.
for the final point, i don't think having a higher FB rate is necessary either, only that there's an overemphasis on GBs by many sabermetricians, especially when applying GB rate to fantasy. it is one of a myriad of factors to look at, and i feel this simplification has resulted in GB rate being overemphasized to where some higher FB% pitchers are undervalued.