In part two we looked at the data and found that the system fails miserably if the goal is to compensate for homegrown players. In this part, let’s talk about what could be done.

At this point, there are three options:
1) Leave it as is.
2) Get rid of compensation completely.
3) Tweak the system.

The first two seem completely unrealistic, which means tweaking the system is the only plausible solution. One of the better ideas I’ve heard about this came from our own Joe Pawlikowski. Joe’s idea was simply to downgrade the compensation for those players who spend little time with their new team. Bring back the Type-C classification if you need a new name.

Instead of first round compensation, give these teams a pick after the second or third rounds, which does not completely punish teams for employing journeymen or rental players, but it places a tier or two in between those who are losing homegrown talent and those who are losing a player who barely played with them. Billy Wagner with the Red Sox comes to mind.

I’m not sure if there’s a silver bullet answer that magically cures the ailments, but the system appears broken enough that just about any kind of reform should represent an upgrade. The draft compensation system is no longer a market inefficiency used by insightful teams. Over the timeline each team but the Pirates raked in at least one first round compensation pick, and the teams that netted the most were the Braves, Red Sox, Angels, Athletics, Diamondbacks, and Rangers. The bottom five saw a jam of teams with three picks, leaving the Rays as the only team with fewer than two. The Royals, Rockies, Phillies, Orioles, Nationals, Marlins, and Astros held three apiece.

With the exception of the Phillies and Rockies, those lower teams usually do not field contenders. The upper teams do. Is the purpose of the draft not to help raise the lower team’s talent levels? And yet, baseball’s compensation rules are directly contradicting the entire purpose of the draft by supplying the better teams with more high-end picks than the teams that need them. Given the inherent variability of draft picks working out, sometimes all it takes is more ammo to hit the lottery.

We’ll examine the top pick fetchers and who gamed the system best later on, but for now let’s leave it on this note: the Yankees held as many first round compensatory picks as the sum of Rays, Orioles, Royals, and Pirates picks.

For background information on the purpose and methodology please check out part one.

Between the 2000 and 2010 drafts, 171 players netted teams a collective 219 first round compensatory picks; or roughly 15 players per offseason. The average percentage of plate appearances or innings pitched spent with the benefiting teams comes in at 37.9%. Meaning, simply, that fewer than 50% of the playing time is coming with the teams collecting those draft picks.

Now, this data is skewed a bit. As mentioned in part one, this focused only on first round picks. Including second round picks might raise that number towards 50%, but maybe not. This look also makes no effort to separate first time free agents (in the truest sense) from those veterans who qualified for free agency compensation multiple times through the timeline. I do not believe the latter to be a fatal mistake because the entire point is to show how flawed the first round compensation system is, and those free agent lifers play a big part in the proceedings.

I wanted to get a feel for what establishing a playing time threshold on these compensation picks would do to the supply of the picks and the number of players who would qualify. Starting with the nearly impossible number of 95% or higher returns 33 players and 47 picks; note that the picks/players average rises because most of the elite players spend their time with one team before leaving. These are the cases like Jim Thome and Manny Ramirez with the Indians, Alex Rodriguez with the Mariners, and Troy Percival with the Angels amongst many, many others.

Moving the bar down to 75% yields 35 players and 49 picks or about an identical sum to the 95% rate; meaning only two players fell in between 75-95%. Drop it to 50% and you arrive at 49 players and 70 picks. Go further down to 30% and we’re at 70 players and 95 picks. That’s still less than half the original player and pick totals. Flip this to players who recorded 20% or less and the totals are 81 players and 100 picks. That’s right, you get more players and picks from the 20% and under crowd than the 30% and over group.

Go even lower, to 10% or less, and you’ll find 41 players and 51 picks. Less than 5%? 20 players and 25 picks. Remember, we only had 35 players and 49 picks from 75% or more, and here, on the very end of the scale, the numbers match up fairly. Here’s a chart that should convey some of these numbers in a more digestible and comparable form (Note: players are counted in each appropriate group so the total sum of the 95%+ , 75%+, etc. will exceed the actual sum):

If the league required that players had to spend at least 25% of their career playing time up through the previous season with a team to receive first round compensation, we would’ve essentially halved the actual player and pick pool. That is incredible and signals that the system fails miserably if the goal is to assist with homegrown losses.

After reading Buster Olney’s column yesterday morning, I had a brief conversation with Jonah Keri about draft slotting and compensation. I half-heartedly threw out the suggestion that perhaps compensation should be altered so that players have to spend a certain amount of time with their last team to qualify for compensation; the thought being that those compensatory matters only exist to help teams replace exiting homegrown talent. I think I threw out a number along the lines of 60% of their plate appearances or innings should come with the team benefiting.

The ramifications of such a move would alter the trade value of those players while also raising the value of the draft picks as the supply dwindled. Curious, I decided to look into just how much time most of the players are spending with the teams being rewarded with draft picks. Beginning with the 2000 draft and ending with the 2010 draft, I went through and noted each player who returned a first round draft pick and/or first round supplemental draft pick. I did not – and this is important to note – include players who returned picks later than that; meaning I focused on Type-A free agents.

From there, I took each of the teams that these players returned draft picks for and figured out the percentage of the player’s career plate appearances or innings pitched (to that point in the player’s career) came with the compensated team. That explanation might be a bit confusing, so let me write through a real life example.

Take the 2007 draft. Carlos Lee brought in two first round picks for the Texas Rangers. Lee had racked up a total of 260 plate appearances with the Rangers the season before out of 5,029 plate appearances for his career. Two sixty divided by 5,029 is roughly 5% of Lee’s total plate appearances through 2006, and yet, the Rangers are the ones who landed those picks.

Now, obviously, it’s not that simple. The Rangers did give up talent in order to secure Lee’s rights and significant talent at that, but the point is: if the compensation rules are designed to replenish teams losing homegrown talent, then examples like this one prove that it doesn’t work. At least, if these examples are in the vast majority, which is what we’ll examine later today.

Last night on Twitter, a follower (the unfortunately named MrNegative1) alerted me that 13 first round picks were starting for their respective teams. The talent and names ranged from Matt Garza to Justin Verlander to Mike Minor and so on. Curious, I looked up the remainder of the starters and was reminded that the Rangers drafted Tommy Hunter in the supplemental portion of the 2007 draft as compensation for losing Mark DeRosa. A closer examination yielded that the Rangers made more first round selections that year than I could recall.

In the first round alone, the Rangers selected Blake Beavan, Michael Main, Julio Borbon, Neil Ramirez, and Hunter. Later on in the draft, they would select Mitch Moreland, Josh Lueke, Matt Lawson, Evan Reed, and Ryan Tatusko. With the exception of Ramirez, each one of those players is playing or has played some role in the Rangers’ fantastic season. Moreland, Hunter, and Borbon are obviously with the big league team and starting on some nights. Meanwhile, Main went to San Francisco for Bengie Molina; Lueke, Lawson, and Beavan to Seattle for Cliff Lee; Reed was part of the Jorge Cantu deal; and Takusko part of the Cristian Guzman trade.

Farm systems are multi-purposeful and the Rangers have used each of the phases well. They can be used to groom young players to eventually star on the big league team, trading chips, and role players alike. You can argue that the Rangers overpaid in trades or that the various players are nothing special, it’s still somewhat impressive that the Rangers’ 2007 draft, which produced 54 players total, has been such a talent well for them this season.

Let’s talk about how the new draftees with major league deals in hand and options work — starting first and foremost with 17-year-old Bryce Harper. The precocious talent very well could reach the Nationals before being legal drinking age. Such is a rarity and even more so when a player who will not turn 20 for another two seasons is given a major league deal.

None of these players, Harper included, will use options because the signing deadline does not occur until the middle of August. One of the provisions in using an option is that a player must accumulate at least 90 days of service time within that season for it to qualify. In 90 days, the baseball season will be over, meaning nobody who signed last night is going to come close to qualifying. This is true for any player placed only in short season ball during a given season.

How about a hypothetical. Say that Harper could reach those 90 days, what does it mean for the Nationals? They could circumvent the option usage by making sure his minor league stint lasted for fewer than 20 days before recalling him to the majors, thus ensuring no optional assignment burned. The only catch there is that his minor league time would be applied to his major league service time. Instead, teams are given an extra year of padding based on a rule which allows for four option seasons if the player has fewer than five professional seasons. That comes into play with the aforementioned 90-days rule.

Pretend that Harper impresses the Nationals so much in spring training next year that they option him to a full-season league at the end of spring. After a brilliant month in A-ball, he tears his ACL while running the bases and misses the remainder of the season. The Collective Bargaining Agreement protects the Nationals here, too, as a player must have 60 days on an active list before disabled list time counts towards service time. If Harper played until, say, August and then tore his ACL, he would be credited with a professional season and hence have an option used.

An important distinction to note is that pro service time and major league service time are not the same. The latter – which is undoubtedly more popular and well-known – is the clock that determines arbitration and free agent eligibility. While there are some rules about options that revolve around the major league service time none of those come into play with these signees for the time being.

In summary, here are the key points:

– Players who signed major league deals last night will not use an option because they will spend fewer than 90 days in the minors.
– Players will likely begin burning options next season, assuming they play above short-season ball and stay relatively healthy
– In some cases, these players will have four option years, not the standard three, if those options fall within their first five professional seasons – dictated not by draft date, but by the above 90-day rule.

All of this adds up to mean that Harper does not have to be a fulltime major leaguer before he hits age 21. Assuming the Nationals push his development at an advanced pace, they will have at least four seasons beginning next year for him to prove ready.

First I’d like to acknowledge Baseball Prospectus’ Rob McQuown’s comment about needing to league-adjust the Clutch wins total. Somehow, the pitcher batting aspect slipped my mind, but it is a legitimate concern. At this point, I do not have the rectified team numbers although I’ll see if I can’t come back to that sometime in the near future.

This post is a little more on the trivial/fun side than the others because it concerns the players on the Cubs during those seasons. With the assistance of Jack Moore, I gathered each of the players between 2000 and 2010 who recorded at least 300 plate appearances with the Cubs and also played significant time elsewhere during their careers. I used careers rather than just 2000 onward because I wanted a larger sample size. Unfortunately, the total came in at 36 players, which is not huge, but not bad.

As for the results, it was 50/50 in terms of improving/declining Clutch scores with the Cubs as opposed to elsewhere. The most egregious of declines skews the mean and other data. Believe it or not, Sammy Sosa. I believe Sosa falls into the trap in which most of the criticism about the Clutch stat would reside. Namely, he was too good in regular situations. So, being anything shy of Superman in Clutch spots killed him. His career OPS in more than 2,000 high leverage spots is only .820 though, opposed to a .878 OPS overall.

Everyone else is tossed around randomly. Juan Pierre and Jacque Jones are second and third. Ramon Martinez and Michael Barrett are third and fourth. Mark DeRosa and Derrek Lee are fifth and six. And so on. The guys who actually have better Clutch scores with the Cubs are more interesting. The top two are Mark Grace and Fred McGriff. Grace might not be a surprise, since he played the majority of his career with Chicago and had his best seasons there. McGriff, though? A bit odd. Rondell White and the Bartman game goat, Alex S. Gonzalez, are next up, with Alfonso Soriano rounding out the top five.

There is no correlation between the non-Cubs and Cubs’ Clutch scores, which probably isn’t too shocking, given the variance in plate appearance totals, but … well, other than Sosa skewing the numbers down a bit, I have no explanation for why the Cubs are so random. Therefore, I’m blaming the goat.
Read the rest of this entry »

Earlier I detailed the Cubs’ clutch-hitting woes over the past decade-plus. I’m back with more stats on a league-wide level to support the insanity of the situation.

After taking each team’s Clutch batting score from 2000 until the 2010 season and averaging them on a per-team basis, I found that the mean is -7.26 wins and the standard deviation is 12.7. That means that the Cubs (-35.1) are more than two standard deviations away from the mean. The Diamondbacks (-26.2), Reds (-24.1), and Rockies (-22.6) are all a standard deviation away on the same side of the scale as the Cubs. Meanwhile the Angels (16), Twins (14.7), Mariners (11.7), Athletics (8.9), and Royals (8.5) are more than one standard deviation away from the mean on the other side.

No team, though, is as far away from that mean as the Cubs. The data presents a normal distribution, with 20 of the 30 teams falling within one standard deviation of the mean and every team but those Cubs falling within two standard deviations of the mean; making the Cubs a genuinely remarkable story. One that I have no explanation for other than, that is baseball.

I then ran a standard deviation on a team-by-team basis, to see if the Cubs were amongst the team with the most consistent Clutch scores. As it turns out, they are near the middle of the pack at eleventh. The most impressive team, from my perspective, is the Twins. Who have one of the highest cumulative Clutch batting scores and the second lowest deviations. The Angels are on the flip side, with a standard deviation of nearly four wins and the Reds, well, they’re a special story, with a standard deviation over eight thanks to some hot and cold seasons.

Below the jump is a complete list of teams and their cumulative Clutch hitting scores.
Read the rest of this entry »

One has to wonder what fans of the Cubs would’ve said if a little birdie foretold of Marlon Byrd, Starlin Castro, Carlos Silva, and Tom Gorzelanny’s seasons in the springtime. Sure, Angel Guzman would still be out, but with those four contributing so much, the Cubs would have to be contenders. To steal a bit from Carson Cistulli, if the Cubs’ fan on staff, the illustrious Bryan Smith, had his druthers, this Cubs’ season would disappear.

The list of disasters for the Cubs this season is long and stretches from Carlos Zambrano to Aramis Ramirez to Derrek Lee and beyond. We have this nifty little metric on the leaderboards and player pages called “Clutch”. Essentially, it measures if a batter hits better or worse in – you guessed it – clutch situations than they normally do. The Cubs as a team are hitting worse. Far worse. Like -5.88 wins worse.

The most critical offenders are Geovany Soto (-1.96), Tyler Colvin (-1.06), Lee (-0.95) the departed Ryan Theriot (-0.84), Alfonso Soriano (-0.75), and Ramirez (-0.50). As it turns out, those six batters account for nearly 50% of the Cubs’ high leverage plate appearances. Whenever that much of your lineup is having serious issues hitting in big spots, trouble is bound to ensue. At first blush, with the talent on this roster, it looks like a feature of the offense that regression and a fresh slate will ostracize. Slight problem there: that last sentence could apply to eight of the last ten seasons and now nine of the last eleven. Here are the Cubs’ Clutch scores dating back to 2000:

2010: -5.88
2009: -4.98
2008: 0.90
2007: -3.10
2006: 0.10
2005: -2.63
2004: -6.52
2003: -2.15
2002: -5.96
2001: -0.85
2000: -3.79

That, folks, is a total of -35 wins. The Angels have become spectacles of clutch hitting perfection. If you take their Clutch scores since 2000 and add them all up, you won’t reach 20 wins and if you only add up the positive values you fall four shy of 30. The Cubs are essentially the anti-Angels but to a higher degree. If someone were to ink a holy book of unclutch named The Book of A-Rod (ironically, of course), the 2000-2010 Cubs would be all up in that.

There does not seem to be a common thread either. They’ve had four different managers over this span, numerous hitting coaches – including the highly praised Rudy Jaramillo – and even a couple of general managers. Maybe everything, from teaching to evaluation and preparation methods remained the same, but that seems unlikely. This phenomenon drains on players homegrown and free agents signed alike. It appears to hold no discrimination for age or position. The only thing constant is that it has zapped the Cubs for the last decade plus and the answer as to why is mystifying.

Lyle Overbay passed through waivers yesterday which registers as a non-shocking development given the Jays inactivity in moving him to another team over the last year-plus. Overbay has a quiet genius about him in that he was consistent throughout his career.

In 2004, his first full season in the majors, he posted a WAR of 2.4. The next year? 2.5. The year after that? 2.5. Then he had a down (and shortened) season with the Blue Jays, but after that, his WAR have been 2 and 2.4. This season he’s on pace for around 1.5 wins, depending on playing time, but his career WAR should top the 14 mark without hassle.

Overbay’s consistency got me wondering: just how common are two-win first basemen? So, I went back and looked at 2002-2009 using our leaderboards with a per-season minimum of 300 plate appearances at first base. I tallied each year up by the amount of players who had at least five WAR or greater and so on. Each player is only counted once; meaning Albert Pujols is in the 5+ WAR column only and not duplicated in the 4+, 3+, and 2+ columns as well. Here’s the graph:

Last season appears to be a golden age for first basemen, both elite and above average alike. Let’s take this a step further, though, and combine the first basemen into two groups: those who produced 2 or more WAR and those who produced fewer than 2 wins. From there, let’s make it interesting and make the grouping in which Overbay belonged to during each season yellow. Here’s how that looks:

What does this tell us? Well, besides the obvious – that Overbay was often above 2 WAR – that he was in the first base minority in 2008, 2005, and 2004. Maybe, then, it’s not a surprise that he was traded following the 2005 season, and trade rumors heated up for him once against after the 2008 season.

Rare is it when a pitcher’s ERA, FIP, and xFIP are in a state of complete accord. Nearly as rare is Freddy Garcia pitching more than 100 innings in any given season during these twilight years of his career. All together, I suppose that makes Garcia’s 2010 season an exotic bird with colorful – if frangible – feathers. Twenty-one starts into the season, Garcia holds a 4.90 ERA, 4.90 FIP, and 4.70 xFIP with 119 innings pitched.

The last year-plus has been a tortuous path for Garcia. Signed to a minor league deal with the New York Mets, Garcia made starts for their organization before dismissal. Garcia then latched on with the team he spent the 2004-2006 seasons with and eventually made nine starts, pitching well enough for the Sox to exercise his option, worth $1 million with $2 million in additional performance-based incentives.

Garcia’s fastball still lacks life and sits at a career low 87.9 miles per hour. He’s adjusted by using it a career low 30.5% of the time. For comparison, his next lowest usage rate came last season at 43.4%. Garcia is using a combination of his changeup and split-fingered fastball instead and our pitch run values have his changeup as his best pitch and the splitter as his second best – albeit in the negative, as are his other three offerings.

When Garcia gets ahead in the count, his fastball has the tendency to become a rumor. He uses it 15% of the time or less in each of the two strike counts with the exception of 3-2 counts. It’s only when he falls behind that Garcia leans heavily on his heater, which is about as hot as the unthawed. It should come as no surprise that one-third of Garcia’s homers have come when he trails and another 11 on even counts. When he gets ahead, batters have a .373 slugging percentage against him, when he falls behind, batters are slugging .536; league average for those situations is a .306 slugging when the pitcher leads and a .498 slugging when the batter leads. In other words: when Garcia falls behind, his offerings are being smoked, and he is even being hit harder than normal when he does his job.

Those statistics come with the territory for right-handed starting pitchers with Ronald Reagan era fastballs. Not that Tupac Shakur era fastballs guarantee success, but so much of pitching is luck-based already, having a slow fastball just shaves the margin of error a little more. That Garcia himself acknowledges those limitations in an implicit manner is one of the reasons he’s been able to succeed as an above replacement level pitcher this season.

And that’s something all the metrics agree on.