rluzinski

rulz, I am trying to think outside the box by catching a trend before it becomes,..you know, a statiscallly measurable trend. Well, I'm not contending that I'm 100% sure Weeks is perfectly fine. If he's hurt, I wouldn't know. My contention is simply that on the balance, 40 spring training ABs does a very poor job of representing a player's true skill. Craig friggin' Counsell has a 1.266 OPS!!!!

but i can't wear the sample size blinders that most of you have on here. You are just disagreeing with us because you don't like our answer, not because you (as far as I can tell) know anything about sample statistics. But hey, if the "sample size" mantra is just used by those who can't handle the harsh realities that the Brewers' spring training is bringing to light, perhaps I can be converted: Counsell: 31 AB, .484 BA, .556 OBP .710 SLG Counsell just must be playing with so much confidence right now that I think it's obvious it's going to carry over to the regular season. The Brewers better find a starting spot for the guy or it won't be long before he's demanding a trade. If Fielder can't improve on his current .389 SLG, maybe we can slide Counsell over there to start the season? Have I shed those ignorance inducing sample size blinders yet? Have I seen the light?

I wish I could be as positive and patient as you but I've never seen a player struggle so bad in Spring Training And when I didn't think Gwynn Jr. was the next Gwynn Sr. after 36 AB last year, people wanted to know why I was being so pessimistic. I may not always be right but I'm just looking at the facts as I see them. I'm patient with regard to analyzing performance because the game demands it.

But for the most part, hitters are trying to hit. Prince is trying not to pull everything, braun might be trying to be more selective. It's not even worth fretting about for so many reasons.

From the comments section, the author writes: "Seriously though, while some of what I wrote was to get a reaction..." That's exactly what I suspected and I HATE when it's done.

What I took from it is that you sometimes have to do things that are not necessarily the best option just to keep the defense honest. I agree with that (have to do that stuff once and awhile to keep the defense honest) but most of the rest of what he says is utter nonsense. Sabermetricians would advocate letting runners steal at will? Sabermetricians don't see the value of Henderson's 81% success rate while stealing? Nonsense. I almost think it's an early April fools joke? Sabermetricians generally don't advocate small ball strategy because the average reward often does not often justify the risk. I'd love to see some tweaks by MLB to change that (small ball strategy can be very entertaining, IMO) but it's never going to happen.

But we have 4 or 5 players who aren't hitting home runs right now. That certainly becomes more statistically significant. Only if we knew nothing else about the players on the team would it become more stastically significant. But in this case, the more players you include in your team sample, the more information you add of prior knowlege. You wouldn't have any expectations of these players if you didn't have prior information (scouting and past performance) telling you to. The more players you add to your team sample, the more thousands of AB from previous seasons you need to consider. Unless we have a reason to believe that the perfomance of teamates are somehow linked (beyond sharing a specific ballpark and having common opponents), a team sample tells us nothing more than simply looking at the samples individually. This is why my in-season series probabilities are currently pretty flawed, even if I do wait until about midseason to start them. I'm trying to estimate a player's true skill from maybe 300 AB, which has a high degree of uncertainty associated with it. I should be adding the in-season sample to my preseason projection as my first step. By the way, has anyone ever actually does extensive regression analysis on spring training stats? Surely there must be some correlation between spring training success and regular season success. That would make for a great article. I vaguely recall a study that looked at the most extreme performances over spring training and concluded that it predicted a small increase or decrease in expected performance for that upcoming season. Something like 40 points of OPS above or below expectations in spring training translated to 4 points in the regular season. It was small but it existed. Spring training starts aren't worthless, but small samples, lower average opponent skill and players not always putting winning first adds a lot of noise to the data.

Considering the relatively difficult starting schedule and the fact that the Brewers are maybe an 87-88 win team with Cameron and Yo, my gut tells me to be happy with anything above .500. Without doing the math, I'll say 13-14 and mass hysteria at bf.net.

"You can preach small sample size all you like - however, as a team (going off of 549 at bats so far) we would be on pace for only 141 Home Runs in a season." That 549 sample (which was compiled from many small samples) is not the same thing as having a 549 sample from one person. For a team, that's still a very small sample. I think that some (not all) people have a tendency to over react to spring training stats because they don't understand the concept of statistical uncertainty as it relates to sample size. There are a few different ways those people can potentialy deal with it: 1. Trust in those that already understand the basic concepts and accept the limitations of small samples. 2. Educate themselves with at leasdt the basic concepts behind statistical samples. 3. Dismiss the entire field of statistics because their findings are contrary to your intuition. To number 1 I say, trust no one! To number 3, not good. I very much prefer the second option and I suggest that anyone interested in learning more about the "sample size topic" start with this article: http://www.athleticsnation.com/2008/3/8/252165/staturday-small-sample-siz Even if you skip over the math-heavy parts, I think the author does a nice job of illustrating the concepts. Without understanding them, it's impossible to appreciate why some people can so easily dismiss the significance of a 30 AB sample.

I'm already annoying that I never know if I'm watching a flashback or a flash-forward. Now Lost is going to use that confusion as a plot device? Another underwelming episode, IMO.

I appreciate people some being skeptical that Weeks will have a break out year. I also agree that, while Weeks' performance after coming back last season was spectacular, it's not very strong statistical evidence of great things to come. That said, When you consider Week's recent struggles with injuries and the fact that his lowest OBP projection for 2008 is still over .360, it makes sense to be hopeful about 2008.

The thing is, Lohse isn't even mediocre. He has a career ERA of 4.83 and his ZiPS projection for 2008 puts him at a 4.78. I would probably want Bush and if I was certain he was healthy, Capuano. Of course, Lohse thinking he was going to get a long term deal was probably just wishful thinking to begin with.

Anything I could say on this topic, I've probably already said in the designated lineup thread. Makes me think that this should be there as well. http://forum.brewerfan.net/images/smilies/wink.gif

James developed the speed score, which does a pretty good job of approximating a player's speed. This doesn't really tell us anything about a player's speed.

Okay, how do we make some lineup decisions based on this stat? I don't see why you should.

The way I like to say it is that the spread of the true talent of a pitcher's ability to control H/BIP is very small. In contrast, the spread of H/BIP over 200 innings is very large. As the sample size increases, the sample spread is going to approach the true talent spread.

Projections are regressed to the mean, since true talent is going to be closer to average than the actual performance you are going to see in 162 games. If you take the 10 highest or lowest for any stat in any year and see how those guys did the year after, they're performance is going to collectively be closer to average. BA, HR, strikeouts, whatever. Take a young player who had a breakout year in 2007 and most of them are going to do worse for 2008. That's all a "sophmore slump" really is.

He butchered a number of ball in RF last year in limited chances. Moving to CF doesn't make things any easier. He embarrassed himself on 2 or 3 plays last year. That is 100% true. Trying to use 2 or 3 very good or very bad plays to judge the overall defensive talent level of a player is a very bad idea, however. Throwing out my own personal observation (which isn't worth anymore than anyone elses) but based on everything I've seen of him, I'd put him a little above average in the corner and a little below average in CF.

I think Gwynn is most likely a very good defender, significantly better than Gross. I just don't like assuming his gold glove material without much evidence, is all. The real question is whether Gwynn's superior defense is going to compensate for their difference in offensive value. ZiPS projections for both (BA/OBP/SLG/OPS): Gabe Gross . 264/.358/.426/.784 Anthony Gwynn .260/.325/.328/.653 I value defense more than the average guy but that's a huge difference. I'd be very surprised is Gwynn's defense and prowess on the basepaths could make up the difference. My preference would be to have Hart play CF until Cameron comes back (I think he's more than adequate there). Then use a Gross/Dillon platoon in RF and DFA Kapler. Use Gwynn as a defensive replacment/pinch runner whenever possible.

I think everyone has accepted the fact that Prince will gain weight as he ages. Anyone who thinks he's going to be playing when he's 38 is probably fooling themselves. You can only fight genetics for so long. That said, there's no reason to believe that Prince will have any weight issues while the exclusive property of the Brewers. It will be someone other team's problem.

True, numbers don't mean everything, but the rumors out of ST are that Cappy has not looked great as far as velocity is concerned. Where did this rumor come from and why is it a rumor anyway? Are pitch velocities in spring training a guarded secret?

Gwynn makes more consistent contact than Gross. Ignoring the fact that simply making contact with a ball that results in it being put into play has little value in itself, let's look at the ZiPS projections for both players. Gross: 333 AB, 60 K, 5.6 AB/K Gwynn: 439 AB, 77 K, 5.7 AB/K The difference is even worth noting. I know that we are just supposed to assume that every fast, skinny player must put a lot of balls in play but for Gwynn, it's a bad assumption. This has been pointed out many times on this messageboard.

Tony Gwynn's ZiPS projection for 2008: 260/.325 /.328/.653 I guess that Yost is expecting gold glove caliber defense out of him to justify that stick in the lineup.

If it isn't clear, the books' study of dual leadoff hitter doesn't use 8 identical players but rather 8 average players for their lineup position. You can use ESPN's splits to find out what the average NL leadoff hitter did for 2007, for instance: http://sports.espn.go.com/mlb/stats/aggregate?sort=runs&split=109&group=8&season=2007&seasonType=2&statType=batting&type=reg The Book basically found out the average for each lineup spot and then flopped the average #8 and #9 hitter around.

Brewcrewball interviewed Dave Lawson a few years ago: http://www.brewcrewball.com/story/2006/4/14/1151/39437 He was recently mentioned here for inventing a stat that's basically been independently invented over a dozen times: http://walksaber.blogspot.com/2008/02/beating-dead-horse-pt-4.html An average avoided DP saves about a third of a run. Perhaps, avoiding a double play with the top of the order coming up is worth more. But even if we inflate it to half a run, is Kendell going to avoid 20+ double plays by batting 9th ? He's never had more than 26 GIDPs in a year so I find that basically impossible.. Hell, a third of the time it's going to be a pinch hitter batting before him anyway. Are we missing some other significant advantage? I don't think Lawson is wrong; I'm skeptical that the 25-30 run estimate even came from him. The latest Hardball Times Annual came with an access to a decent lineup analyzer. If I get some free time tonight, I'll play around with it and see what it spits out.