Nox

I have a feeling you'll talk about it but you certainly won't be the judge of it. You don't say. Alan Nathan, the guy whose existence you learned of an hour ago, showed teams how to do this 5 years ago with their HitFX data. Of course the harder a ball is hit, the more likely it is to fall for a hit. Having this information at your disposal doesn't make it a slam dunk to figure out who actually has the ability to do this consistently though. It's not a magic bullet like you seem to think. Oh, so n=1 then. Much better. I'd say about 30% of people working in analytics are actually listed on teams' sites. The reasons for that are pretty obvious. There's no incentive to publicly advertise what you're researching. Fearing has been working for Tampa for the last ~4 years. He was put on their site a month ago. Tango has never been listed on a team's site despite consulting almost full time for various teams throughout the last decade. Sure. For all intents and purposes, a CDS is an insurance policy on an underlying asset. Traditional statistics work fantastically when the underlying asset is something that's governed by mild randomness (gaussian, semi gaussian environments). That's why things like life insurance, car insurance etc are essentially licenses to print money for those companies. Their models are tight and they'll never lose over the long term. The frequency of almost all events in this domain are quite predictable. Things change at a fundamental level when the underlying when the underlying asset is something like gov't debt from s***** Country X or a pool of supposedly uncorrelated mortgages. The randomness that runs the show in those domains is of infinite variance and thus can't really be modeled (model error and convexity effects crushes everything here from an analysis standpoint) at least in terms of predicting the frequency of events. Trying to do so leads to monstrosities that do their best to blow up the world. Like this: http://en.wikipedia.org/wiki/VaR (Quote I just read that I liked: "an airbag that works all the time, except when you have a car accident.") or this: http://en.wikipedia.org/wiki/Gaussian_copula#Gaussian_copula (note the inventor of the latter is as disgusted at it's misapplication in finance as anybody). So, if you don't really know what the frequency distribution of the underlying asset looks like and can't actually predict frequencies of outcomes, doesn't it seem pretty silly to try to insure said asset?

Is it possible that Nonis' true master strategy is to put a team together so bad so as to be in prime position for the Connor McDavid sweepstakes in the 2015 draft? Super Potato? Stuffed Potato? Twice Baked Potato?

Upon further review, you're probably right.

Just thought I'd stop by to say how much I'm enjoying this Leafs collapse! Though I'm torn between having them miss the playoffs entirely or have Boston s*** on their faces in R1.

You seem to think that this information becoming public is somehow going to fundamentally alter what we think about DIPS and some of the BABIP based metrics that are currently in use. Despite what your intuition might be telling you, it doesn't. With the information available, turns out they do a pretty damn good job of getting most of the way to where they need to go. Variance does not disappear as we measure things closer to a player's true talent. First of all, sweet cherry pick. Contrast that with Alan Nathan, Doug Fearing or Sig Mejdal. Do they meet your standards? Would you say that they might have a speaking part in a conversation regarding the analysis of baseball? Secondly, how do you know the kid they hired is an idiot? There's a non-zero chance that he's legitimately smart and an even bigger chance that they took him on in some sort of scouting capacity. "Analyst" is just a catch all title for the lowest rung of jobs in the industry. They can be anything from a video scout to a SQL code monkey to a guy working on some enhancements to a pitch classification algorithm. Thirdly, didn't you just get over telling us all that you thought WAR was used as a predictive measure? You have plenty of good insight in other areas (CBA, taxes etc.) so I don't know why you think you need to speak so authoritatively here when it's a clearly not something you've spent alot of time trying to understand.

Teams have batted ball velo, launch angle etc. from both HitFX and Trackman. It's not like any of that instantly stabilizes either which you seem to think it will. It's just another, albeit, closer proxy to measuring a player's true talent which is really all that a team should care about (from an evaluation standpoint). I'm also not sure why you don't think there are smart people working on this stuff full time. The barrier of entry to the top end analytical teams is incredibly high. You basically need an advanced degree in Math, Physics, Engineering, Operations Research or you can get the f*** out.

Ugh.

At least nobody pointed out that he was 2 and 8 last year. I guess that's progress.

He probably found a reliever with a low ERA and some holds and thinks he has a steal brewing.

Why bring on a replacement level guy who's out of options? That doesn't make any sense.

Arencibia and Lawrie are incredibly loathable. Actually, hard to think of any team more hate-able than the one fielded last year.

Fun fact: Mitch Willams received MVP votes. http://www.baseball-reference.com/players/w/willimi02.shtml

Saying you'd rather be a closer seems like a terrible life decision.

I'm Dinger's Paul Beeston and I can confirm that there are no more pieces going the other way. "We're going to have a track team!"

No, I haven't worked as a trader. By the time those jobs became age appropriate for me the markets had already spiraled into the rigged game that they are now (recent anecdote along those lines: http://www.zerohedge.com/news/2014-03-10/holy-grail-trading-has-been-found-hft-firm-reveals-1-losing-trading-day-1238-days-tr). Despite being a completely vulgar profession I can't say the money on the right side of that scam doesn't appeal greatly. But if we're being real, I have a CV that's really good, not the top 0.01 percentile phenomenal that's required to get in the door at the big HFT firms. Not to mention the fact that the recruiting process for the entry level private equity stuff is essentially hell on earth. Back to what you said though, the 2 stochastic environments in question (baseball and free markets) are incredibly different. The flavours of randomness are essentially apples and oranges.

Just a few bad breaks went against the Leafs. Or something.

Mods - Ban and deploy virus.eviscerate.exe please.

Terrible analogy. Sports and free markets operate in completely different realms, governed by fundamentally different forms of randomness. His model may be wrong, I dunno, I haven't looked at it and even JFas would freely admit there's work to be done on it (as there is in all of hockey analytics). But your analogy sucked, just thought you should know.

And then send two gift baskets.

Ah yes, this guy. http://www.fangraphs.com/statss.aspx?playerid=5648&position=P

Probably just an artifact of overfitting. Thanks though! 0.86 seems reasonable. I'm guessing the truth lies somewhere between that and 1.

How intense of an effort would it be to change the cutoff to 29? (This is mostly out of curiosity and not really even a suggestion for increased accuracy or anything like that)

Exactly.

Using service time is for sure the "right" way to do it but it would be really really time consuming to do it based of public data sources. I actually don't even think you could get historical service time per year per player from public sources. Even Cots (which I'm assuming is being fed an MLB data stream by some front office source) only has current service time figures for guys. So from a pure feasibility standpoint I don't blame JFas for trying to do this with a rule of thumb approach. I just wonder if the results are a bit too skewed to be useful because of the aforementioned case. (It may in fact be fine...I don't know).

While you're absolutely right, I think the bigger problem is the perception of the city for American guys much more so than any sort of tangible quality metric. I've absolutely hated Atlanta too when I've visited.