Sky Klkmn

I’ve been making a lot of ice cream recently, spurred on by acquisition of a second freezing cannister. I’m no expert, but I do have some advice for new ice cream chefs. These are those tips.

1. Keep your freezing cannister in the freezer at all times. It takes at least 24 hours to freeze solid, and 48 hours is optimal. If you’re anything like me, two days warning for ice cream is way too long.
2. Make sure your freezer is cold enough. This is especially true when you travel and bring your ice cream maker with you. Nothing worse than driving 7 hours to Cape Cod in July and not having your pina colada ice cream freeze. (I know, bourgeois suffering.)
3. Don’t cheat and plan to eat the ice cream immediately after churning. Even just 2-4 hours in the freezer will significantly improve the texture. And take the ice cream out 5-15 minutes before serving to soften slightly.
4. Store the ice cream in a glass container that you pre-“froze” — this will prevent the just-churned ice cream from melting immediately where it touches the container.
5. Buy a second cannister and keep it in the freezer. This is somewhat of an advanced move, but you can make twice as much ice cream at a time (say, for a party) OR you can make two flavors at once. And it’s relatively cheap.
6. Some flavors require no heating of the liquid. These are great and quick. Some do require heating, especially custards. Learn to temper eggs (it’s both easy and easy to screw up.) If you don’t want to wait 4-8 hours to cool in the fridge, bring down to 40 degres using an ice bath (bowl in bowl or baggie in bowl.)
7. Make two complementary flavors at once (one right after the other, actually — this requires two freezing cannisters) and “swirl” them before storing in the freezer. Better yet, pick two different types of “ice cream” — coconut ice cream and mango sorbet, chocolate ice cream and raspberry sherbet, vanilla/chocolate/strawberry ice creams, etc.
8. Check out The Perfect Scoop by David Lebovitz for a ton of great recipe ideas.
9. Get creative with flavors — deconstruct other favorite foods and turn them into frozen desserts. My wife had the idea to start with a basic lemon ice cream and add poppy seeds and shortbread bits to mimic a lemon poppy seed muffin. It’s awesome.
10. Some other of our favorites: blueberry buttermilk sherbet, toasted coconut ice cream, and strawberry balsamic ice cream (you don’t taste the balsamic, it just makes the strawberry taste more strawberry-y.)
11. Simple flavors with quality ingredients is a great route to take, too. “Plain” vanilla made with vanilla beans is way better than you can buy at the store.
12. Try some “savory” flavors, like olive oil ice cream or herb sorbets (basil, thyme are good starting points). With how much sugar is involved they’re still sweet, but it’s a whole other category of frozen dessert.
13. There’s a whole science to texture that involves ratios of fat/sugar/water/alcohol/eggs which can be a lot of fun if you’re into that sort of thing. But you can also follow recipes without missing out on great results.
14. Once you get really into making ice cream, think about homemade sauces, add-in, toppings, etc. Really kicks things up a notch. I just made a mulled red wine and rhubarb sauce that would be great with tons of flavors.
15. Some handy equipment: blender, stick blender (for when pouring scalding liquid into a blender doesn’t sound fun), wire mesh strainer (for removing fruit seeds, coagulated bits of eggs, and other unwanted bits), instant electronic thermometer (for making sure you heat liquids to the right temperature and making sure you’ve cooled liquids low enough before churning.)
16. Buy a small cooler, a bit of dry ice, and mail me samples of your favorite concoctions. Don’t forget to check the appropriate safety boxes on the shipping form.

This post has the goal of explaining a comment I’ve made on Twitter a few times: good scouting writing is a lot like good role-playing-game design. Be forewarned, I’m not a scout and the last RPG I played was probably Final Fantasy III. But as someone who enjoys reading about both as entertainment, let me see if I can explain the parallels.

First, by “scouting”, I actually mean a couple things:

• One, scouting doesn’t just have to be human visual observations, it can include stopwatch times, radar gun readings, video analysis, or a lot of the spiffy new Pitch f/x data. Basically anything you can observe about a player over a small sample (a game or two). We’re not dissecting performance results, but performance methods (I almost typed processes but that’s a loaded word.)
• I’m also broadening the usage of scouting a bit — not just the current state of a player’s abilities, but also a player’s potential and avenues of development. That last one is broad, as it nearly requires a full understanding of how baseball works.

That last phrase is probably the most important — at its core, we’re trying to understand how baseball works. I’m not foolish enough to think we can ever fully answer that question or that we’re close. But to make any claims about players, we have to claim some level of knowledge about baseball and talk about players within that framework.

Now, the RPGs. Let me repeat that phrase again: a full understanding of how baseball works. You could substitute a RPG phrase instead: a full understanding of how a RPG’s universe works. Kill an enemy, get 3 coins. Kill a boss, get a rare item. Collect 1,000 experience points, level up. Pair this crystal with that one, cast a special spell. Use a fire potion on a water monster, deal extra damage. Use a special attack, sit out the next round of battle.

It’s one thing to understand all the little rules of a RPG, but they also have to make sense as a whole. Why do enemies carry coins? Why does a new level bring more stats? Why do crystals exist in this world and why does combining them do new things. Why does the special attack take more time? Some of those answers are mechanical and follow from other mechanisms. Some are narrative and follow from the theme and story of the game. Both can be explanatory and both can be entertaining. Especially in board games, there’s enjoyment in the inter-workings of the rules. And a good story can explain away some awkward rules.

In many ways we’re talking about a model — a model of how a RPG works and a model of how baseball works.

A scout on twitter once told me a story about watching Miguel Cabrera bat in A-ball. He didn’t get a hit and his season line was putrid. Yet it was obvious to this scout (and everyone else, apparently) that while Cabrera currently wasn’t any good, he was the best hitter in that league and had great things ahead of him. That’s a sensational lede, now tell me the rest of the story. Tell me about how his swing is fast and compact, but he can’t recognize a breaking ball worth crap. Tell me how someone learns to recognize a curve ball and why Cabrera will certainly pick up this skill. Tell me how his approach at the plate is poor, what a good approach looks like, and how someone learns to change their approach — is it all at once or are their stages of development? Explain why he has most of the tools to drive balls to all fields, but he needs a slight tweak of his mechanics in order to use those tools effectively. (I could go on for a while here.)

Like RPGs, there are a couple levels to good scouting (actually, “good communication of scouting” might be more accurate.) First, explain the individual mechanics better. Two crystals = special spell; doing X let’s you hit a curve ball better because Y. Second, connect the individual mechanics in a meaningful way. This involves pure storytelling, but also requires the individual mechanics to be discussed in a common way with hooks that can be connected. (There are a lot of paradigm issues at play — many ways to discuss the same thing. Picking a consistent perspective is important.) You get the boomerang from the first boss because he’s the bully who stole it in the first place; a slider combines well with a four-seamer because it moves in two different planes, and hitters don’t deal well with multiple plane changes because X, Y and Z. Everything is connected to everything else if you’re willing to take the time to tell a long story.

So before talking about any individual player, have a full world-view of baseball in place. Know how things interconnect, how one things affects another. And when talking about something specific, relate it to other things that connect to it in a variety of other directions. This adds context to help understanding, adds reasons-to-believe, and makes the whole thing more entertaining. Baseball might not be as cut-and-dried as a RPG (nothing is), but by treating it as such, I think scouting writing can be more informative and more entertaining.

Discussion invited.

I don’t understand the difference between command and control. I mean, I could spit back some of the differences I’ve heard spoken about by scouts, but I don’t get it. Now, maybe that’s my fault for being dense, although communication is a two-way street. But if I’m being bold, I actually think it’s a bad way to talk about pitcher skills. In the next few paragraphs, I’m going to outline why I think these concepts make for a bad model and raise some more questions. Afterwards, I’m hoping to chat with others to flesh things out.

As I understand it, control refers to a pitcher’s ability to throw pitches in the strike zone and avoid giving up walks, while command refers to a pitcher’s ability to locate pitches within the strike zone, avoiding fat pitches that can get crushed. My main frustration with these two concepts is that they aren’t independent. Both involve the idea “aim”, plus other differnig skills. Why would we discuss two skills that heavily overlap? To me, it makes more sense to pull out “aim” as a core skill, then discuss what skills lie behind the ability to aim both in and out of the zone.

There’s an acronym I use a lot in my job called MECE, which stands for mutually exclusive, collectively exhaustive. If you think of a model of reality as a jigsaw puzzle, none of the pieces overlap, and together they show the full picture. I realize this isn’t always possible, but it’s often a good goal. Control and command aren’t mutually exclusive (and that annoys me.)

It’s not just control and command that overlap properties. My non-expert hunch is that one of the main reasons pitchers don’t throw more pitches in the strike zone is that their stuff isn’t good enough. They aim for the edge of the zone because the cost of missing outside (a ball) isn’t as bad as the cost of missing over the heart of the plate (a line drive or home run.) A pitcher with impeccable aim but mediocre stuff won’t rate highly in control. How is that a useful model? Why not, instead, talk about aim and stuff separately? Two pitchers who throw a lot of balls might actually need to work on two totally different things. The one with bad aim should work on aim. The one with mediocre stuff should work on improving his stuff. And how does command fit in? Is it a purer measure of aim? What other core skills are involved/related other than aim.

There’s more complexity, too. Pitchers have varying amounts of stuff and aim on each pitch type they throw. And there are other fundamental skills that could be part of the MECE model of pitching, such as pitch sequencing (could be mental and/or how much repetition a pitcher needs to get “feel” for a pitch), stamina, consistency of break, consistency of velocity, consistency of delivery (not tipping pitches) and how well a repertoire complements itself.

Anyway, that’s my overly aggressive “I’m right and everyone else is wrong” explanation. I’m actually interested in a dialogue, so if you’re interested, help me understand.

I sent this to someone earlier today. If you’re trying to figure out the similarities and differences between the above stats, it may help you.

You can think of wRC+ as wOBA+ (it’s called something different for a pretty technical reason.) OPS : OPS+ :: wOBA : wRC+

OPS and OPS+ are based on OBP and SLG. wOBA and wRC+ are based on linear weights.

OPS+ and wRC+ are park and league adjusted. OPS and wOBA are usually not. (Although you certainly could park-adjust them, they usually aren’t.)

OPS+ and wRC+ are set to 100 as average. OPS and wOBA are not.

*** ***

wRC+ is what OPS+ was designed for. wOBA is what OPS was designed for. The OPS options are just plain less accurate. Some may still prefer to use them because they are more familiar with their scale, but they don’t get you anything else.

*** ***

Sure, a 1.000 OPS is always impressive, even unadjusted. But you don’t need any of these stats to tell you that someone who put up a season that good had a good season. You can see that in HRs, OBP, etc. When you need more accurate measurements, then you should use a more accurate stat. But sometimes color is more important to a story than accuracy, so you choose more descriptive stats. OPS seems to be caught in a middle-ground: too advanced for many, not much color, and has less accuracy.

Yes, you can watch every game for free on CBS and other stations it owns. Yes, you can stream every game for free if you’re a cable subscriber. Yes, the same service was free last year. But I’m excited to pay $4 for the streaming privilege. Why? Well, for a few reasons…

Functionally, because I have to, not being a cable customer nor getting a broadcast signal at my house. Oh, and that whole wanting to watch basketball while at work thing.

But mostly, I’m excited to pay for this service because I want many more a la carte content services like this to exist. HBO GO for a monthly fee? Yes, please. ESPN and NBC Sports streaming for a monthly fee? Yes, please. All MLB games over the internet? Yes, please — well, that one already exists. I’m willing to pay a good price for content I like, I just don’t want to pay for stuff I don’t. Basketball in March is something I want. I’ll support it and I’ll enjoy the heck out of it.

Plus, let’s be real, it’s only $4. FOUR DOLLARS. I just blew that much on one beer.

Addendum: How about some math. Assume cable costs $80 per month and you watch 2 hours of TV a day. That’s 60 hours a month, or $1.25 per hour. I’ll probably watch March Madness a total of 10-20 hours. For $4, that’s $.20 to $.40 per hour. A steal!

• bob walk
• jay bell
• bobby bonilla
• barry bonds
• darryl strawberry
• nolan ryan (looks like roy scheider)
• jose canseco
• deion sanders
• bo jackson
• albert pujoles
• alex rodriguez
• sammy sosa
• mark (the home run guy)
• michael jordan
• babe ruth
• joe dimaggio
• mickey mantle
• reggie jackson
• hank aaron
• lou gehrig
• andy van slyke
• that tex-mex guy
• jesus guy who pitched for the red sox during the 03 or 04 world series
• cal ripken

This is a really fun trade. My scattered thoughts:

• While I’m a big proponent of judging trades at the time they occur, these are all young players who can go many directions. There’s likely to be a clear winner eventually and I think it’s more ok to judge this trade based on results than others.
• To judge this trade right now, you need to have strong knowledge of these players from a scouting point of view. I don’t have that, and most people out there don’t have that.
• Generally, I’m a big fan of trading pitchers for hitters.
• The one year of service time already used up by Pineda is a big deal. If you assume 0%/0%/0%/40%/60%/80% salaries relative to free agent value for years 1-6 of team control, he’s used up 23% of his value even though only 17% of his service time is gone.
• Due to their ballpark, Seattle’s expected arbitration payments to Montero will be suppressed relative to Pineda, for similar value.
• Baseball people I trust don’t think Pineda’s as good as his numbers were last year. His second half certainly wasn’t as good as his start, although just because a storyline is plausible doesn’t mean it’s true.
• On a completely irrational level, I’m a believer of the “Jesus was bored at AAA” story last year.
• Hector Noesi, as a likely 4/5 starter, has more value to teams that aren’t the Yankees.
• Jose Campos is exactly type of pitching prospect I’d target in trades as “throw-ins”. B-grade pitching prospects, as shown by Victor Wang, are the sweet spot of the roulette-style gamble that are pitching prospects, more valuable than their hitting equivalents. They flop a lot, but still have the upside to make the gamble worth it, relative to expectations.

Sometimes I feel like I’m picking on JC.  But then I realize he’s trying to make money off of people via his book and he’s more repetitious than I am.  And then I don’t feel as bad.  Anyway, he’s done some cool things with sabermetrics (the PrOPS framework was awesome) and his econ-based approach is fresh (valuing players from a revenue-generating point of view could be awesome).  However, he’s made some HUGE mistakes and many misleading claims.  Given that popular sites like SOSH, BPro, and the Freakonomics blog give him a forum to spread bad information, I thought it might be helpful to write a collection of articles I can use to help people understand what’s going on.  Today I’ll start with one of his silliest ideas.

The Claim

JC is not a fan of “replacement level”.  One of his arguments against it is that there are very few replacement level players, meaningful they’re not readily available and don’t deserve the MLB minimum salary.  In JC’s model a -20 run player (think Jeff Francoeur) is worth about $3M, while a replacement level model pegs them at $400k. JC’s “evidence” for the scarcity of replacement level players is based on this:

That’s a histogram of every hitter with at least 100 plate appearances during the 2009 season.  Notice there are a lot of league-average hitters, but not so many below a .600 OPS, which is where replacement level hitters might fall.  Because there aren’t many hitters that low, JC claims they are scarce.

Common Sense

Before I get mildly statty, ask yourself, “Self, does it make any sense to think that there are more good baseball players than bad baseball players in the world?”

No, it doesn’t make sense.  What’s going on is a heavy dose (well, two doses) of selective sampling.  One, JC’s only looking at MLB players — but many replacement level players are in the minor leagues.  Two, JC’s only looking at players with at least 100 plate appearances, but many replacement level players don’t receive 100 plate appearances.  If you ignore the population you’re trying to find, you won’t find them.

JC is correct that MLB production is bell-shaped.  Most of the playing time produces league-average production, with less star-level and replacement-level production.  However, while stars are rare, replacement level players are not.  Tom Tango explained nicely…

The Longer Explanation

Even 30 year-olds struggling in single-A are damn good baseball players compared to the world’s population.  So when we look at MLB talent, it’s going to be the right tail of a huge bell curve, like this:

The superstars are at the far right of that graph.  But why don’t we see as much bad production as the left side of the graph shows?  Because teams don’t need those guys.  They’re the bench players or long guys out of the bullpen.  They see 75 plate appearances or 20 IP.

Playing time is distributed based on talent, something like this:

The production levels in MLB are a combination of the two graphs.  There aren’t many stars, so the right side of the graph is low.  There are more mediocre players, and they see a lot of playing time, so the middle of the graph is high.  And there are even more crappy players, but they get in the game so rarely that the left side is again low.  Like this:

Looks eerily similar to JC’s graph, right?

The Summary

The lower the talent level, the easier it is to find.  In other words, the worse the player, the more abundant they are.  That’s common sense.  Replacement level players are stashed on MLB benches and spread throughout AA and AAA (although most AA and AAA players aren’t good enough to be considered replacement level.)  The reason you don’t see much replacement-level production in MLB is because teams can mostly fill out their starting lineups with better production.

Moral of the story: if you limit your search to players in the major leagues with significant playing time, of course you won’t find worse players.  But that doesn’t mean they don’t exist in droves.

Addendum

Commentors at JC’s site made another excellent point — there’s much more turnover among the worse players in MLB.  If you’re an above-average player, you’re likely to play for at least a few years.  But if you’re a bench player, you’re more likely to get a look and never come back.  So by looking at one year’s worth of data, JC ignores the fact that the middle and right sides of the graph are likely to include similar players year to year, while the left side will see a lot more turnover.  Again, there are more players at the lower end of the spectrum.

If These Guys Aren’t In, There’s No Point In Having a Hall of Fame

• Bert Blyleven — just a joke that’s he not
• Jeff Bagwell — huuuuuge bat, plus more defensive value than the usual huge bat

Definitely In

• Tim Raines — raines30.com pretty much sums it up
• Kevin Brown — wasn’t a stud until age 31, but 1996-2000 were just sick
• Barry Larkin — basically who people think Derek Jeter is, but in a Reds uniform
• Edgar Martinez — has a lifetime .418 OBP and 147 OPS+; played 3B for 1/3 of his career
• Alan Trammell (and Lou Whitaker) — basically Barry Larkin, but in a Tigers uniform
• Larry Walker — the fielding and baserunning get him in, plus two gross years: 1997 and 2001
• Robert Alomar

Have Arguments Either Way — No Right Answer

• Nobody right now, but this is a category worth emphasizing

Nope, Not All That Close, But Worth Mentioning

• Don Mattingly
• Fred McGriff
• John Olerud

Extenuating Circumstances

• Mark McGwire — definitely in, ignoring PEDs
• Rafael Palmeiro — more in than out, ignoring PEDs