rluzinski

Atlanta ALMOST lost $1.2 mil but instead lost nothing. I agree that offering arby can be a risk but has to be the worst example ever.

I know the exact calculation to predict how many heads I should get in 10 flips but I'll be wrong more often than I'm right. For IP, we can't possibly quantiy all the variables involved. All you can do is look at historical data and create a model. Add whatever relevant scouting and medical information you feel is relevant and make your best guess.

I think the idea that replacement level is evenly split among all organizations at all positions is false. Replacement level is low enough that it's likely many teams actually have better than replacement options. It's quite conservative. It's so low that they are the kind of guys you can get for a PTBNL. At least, that's the theory. "Freely available talent". What the exact ERA of "freely available" starting pitchers probably changes somewhat and is hard to define in the first place, though. The Brewers failed to find a number of guys that were replacement level last year. And, perhaps more to the point, it wasn't at all obvious at first glance what they had that was the best option and they lost games while cycling through their options. Neither Villanueva nor Suppan were replacement level as starters last year and although regression to the mean is likely, it's by no means guaranteed. I know you know this but just because you have replacement level options doesn't mean you are going to get replacement level production. It's a pretty good bet that both Suppan and Villy are better than replacement level (Suppan barely). The problem is when a true 5.5 ERA pitcher get's 4 starts and ends up with a 6.0 ERA in those starts. Often, teams get impatient and randomly try someone else. But really, if Melvin didn't have 5.5 ERA starting pitchers to throw into the lineup last year, it's because the choose poorly, IMO. I certainly don't have the data, but I would estimate that a (projected) 75 win team would make the playoffs something like 5-10% of the time, an 80 win team 10-20% of the time, and an 85 win team 25-30% of the time. Maybe the 85 win team number is low. If the Brewers had a 20% chance to make the playoffs every year, I think I'd be happy. Even with a projected 80-win season, I'm just hoping this will be an exciting year (as in, playoff push). If you treat the whole season as a Bernoulli Trial (flip a coin 162 times), you can calculate the odds of a 75 team winning a certain amount of games just by luck. Try this calculator: http://www.stat.tamu.edu/...pplets/binomialdemo.html n=162. For a 75 win team, p=75/162=.463. Odds of a "true" 75 win team winning at least 85 games is 7%. At least 90 wins = 1%. Of course, we never really know "true talent", so there's error in our projection as well.

IP Range Odds Ave IP 230 - 211 7% 220.5 210 - 191 26% 200.5 190 - 171 30% 180.5 170 - 151 12% 160.5 150 - 131 5% 140.5 130 - 111 5% 120.5 110 - 91 4% 100.5 90 - 71 3% 80.5 70 - 51 3% 60.5 50 - 31 2% 40.5 30 - 11 2% 20.5 10 - 0 1% 5 100% 164 At least 180 IP: 63% Less than 91 IP:: 11% If I told you Wolf had 2 in 3 odds of having 180+ IP and 1 in 10 shot at having less than 90, would you call me a lier? How about if I told you the most likely Ip range was 190-171? My assumptions above have a weighted average of 164 IP. This is all made up of course. It would be interesting to look at all 30+ pitchers who accumulated at least 360 innings the previous year. What did their IP distribution look like?

That's right, all it took to bring on one of the best relievers in the game (career 3.31 FIP), was Jesse Chavez (4.79 career FIP, for comparison). You havbe to look at his projected salary as well. Isn't Soriano going to roughly get whatever his FA value was in previous year? That's the whole idea of post arby arbitration (must eb a better way to say that), isn't it? Look for comparisons from recent years, adjust for inflation and pay the man. If other teams are going to be paying 70 cents on the dollar for FA this off season, arby guys are going to make out. That may be why so few teams offered arby this year. Soriano probably had negative trade value.

Do any of the modelers really try to do accurate projections for IP, or is it just the normal calculations for the other pitching stats? I don't know what goes into the CHONE IP projections. Marcel just looks at the previous two years, regresses it and adds an age adjustment. It's crude, to say the least. You would hope that a major league club could do a better job of projecting IP by considering all the available variables. It wouldn't surprise me if a healthy 33 year old ML starting pitcher does, in fact average 160 IP. The mode might be higher (more often than not, they finish with between 170-180 IP, for instance) but the 0-50 IP drags the average down? That would be my guess.

I think this is where some of Doug Melvin's philosophies lie. He is not willing to assume a great deal of risk, mainly due to his financial and market restrictions. That's funny because I see the Wolf signing as much riskier than Texas's Harden signing. Signing a 33 year old pitcher to a 3 year deal assumes a singificant risk. Hell, signing any pitcher to a 3 year deal is risky. Even knowing a pitcher's injury history, it's terribly difficult to predict future IP. And the less predictable something is, the less you should pay for it. Also, tying up more value in one player increases the risk variance. Think about 5 players that are each worth 1 win above average and 1 player worth 5 runs above average. You either take the sum of 5 little risk coin flips or one big risk coin flip. Which on has more risk variance? When the two 4.25 ERA pitchers don't provide an upgrade to what you currently have. Or when the 3.75 ERA pitcher will sign with you and the two 4.25 ERA guys won't. My point is, most teams have the available upgrade slots. Find me all the teams that has four starting pitchers that project to have a 4.25 ERA or better. Looking really quickly, in the NL, Phillies, Atlanta (damn they have a good rotation) and Cards? Now, out of those 3 teams who couldn't mecessarily get much of an upgrade in aquiring two 4.25 ERA starting pitchers, which ones also don't have a postional spot they could easily upgrade via FA? Now, I'm not trying to convince you that all players are always worth a linear salary to all teams. But generally speaking, it's the market as a whole that sets the price of FA's, not the exception to the rules. And for every 500 words I write in this thread, I feel compeled to remind everyone that I don't think this is a terrible deal. I can't really have too strong of an opinion either way, until we see how much the rest of the FA's get. If you believe in the linear WAR model (and I understand and appreciate the skepticism many still have for it), teams have been paying about $3 mil for each win above scrub so far this year (compared to $4+ last year). If that holds up, this will look like a much worse deal.

For the record, CHONE now lists Wolf's projected 2010 ERA at 4.17. Now that he's on the Brewers, his projection isn't league and park neutral. His RAR (runs above replacement) is still the same, though (all players are compared to each other in a neautral environment). So my comparisons between him and Harden, using CHONE projections are still valid. Of course, that means Harden's projected ERA on the Brewers might be closer to 3.7, which gives us something else to argue about. http://forum.brewerfan.net/images/smilies/smile.gif Using Wolf's new ERA in our discussion and also adjusting Davis's ERA a bit: Gallardo: 3.60 Wolf: 4.17 Davis: 4.50 Bush: 4.64 Parra: 4.75 Repl: 5.50 4.52 ERA. Better but still not quite average (at least in comparison to last year's NL starting pitchers). Even if we assume that pitching/defense will be average overall, does the offense really project to be good enough to make this a projected playoff team?

I woudn't say it is irrelevent. There is no use in going out and signing two $4MM a year pitchers and getting $4MM worth of production from each vs. paying $9MM and getting $8MM worth of production from 1 guy. This is again my problem with always trying to compare every dollar spent with a linear production line. There are/were plenty of pitchers available at or around Wolf's worth this year: http://www.baseballprojection.com/2010/freeagent2010.htm According to that list, Wolf isn't anything special and I happen to agree. I don't believe production/dollar is a linear function. There are steps up in that pay for the next level of talent. Could you provide me with the evidence that supports that argument? While I appreciate that it would make sense to pay more for better players, the contracts dished out over the last decade or so doesn't appear to support that. Most teams (liek the Brewers) are not short of positions that can be upgraded via free agency. Why should a team pay an exponentially higher salary for a 3.75 pitcher when two 4.25 ERA pitchers can be gotten for less money and give the same overall upgrade? Unless you are the Yankees, it's not an issue.

Why is a 3 year deal a long-term deal? Is it because of his age? It's long compared to 1 year. That's all I meant. OTOH, if I have 200+ IP to pitch, I think there's plenty of evidence that Wolf > Harden + replacement. Especially since I'm not convinced that the Brewers will get replacement level out of their AAA reserves. First, I find it hard to believe that the Brewers can't find a couple of 5.5 ERA pitchers to fill in if needed. The definition of replacement level suggests it's an almost certainty. Off the top of my head, Villy, Soup and vargas fit the bill. Probably one or two guys in AAA as well. Second, if we are only concerned with 2010, I agree that a reasonable projection can make Wolf > Harden + 5.5 ERA pitcher. Wolf at 4.2 ERA and 175 IP vs. Harden at 3.95 and 120 IP, for instance. But as I said earlier, if you also factor in their salaries, they would be even value for 2010. After after you consider Wolf's 2011 and 2012 part of the contract, it makes harden's contract look quite attractive to some. To address another person's comment, no, there is no guarantee Harden was interested in the Brewers at $7.5 mil. Same could be true of Penny. Even if the going rate had been $8.5 mil, I think I still would have preferred it over Wolf's deal. If most teams end up paying Wolf's rate to buy FA pitchers this off season, as big of an offense. If most end up paying the Penny/Harden rate, I will be disappointed.

I really wish people would stop mentioning that Wolf was supposedly the second best FA pitcher. Even if it's true, it's irrelevent. I want to know projected production over the life of the contract vs. cost.

I don't think the 132 IP projection for Harden is realistic at all though or the 163 for Wolf. They are both being regressed towards a mean in opposite directions. I would think a projection of something like 120 IP for Harden and 175 IP for Wolf would be way more realistic. Harden averaged 145 innings the last 2 years, so 132 doesn't look unreasonable to me. You think 12 less innings for Harden and 12 more for Wolf? That's such a small difference that I wouldn't be foolish enough to call you wrong. Really, there's a huge amount of uncertainty in playing time estimates anyway. If I use your IP, they are basically equal for 2010. I'm also not sure I buy the 4.42 ERA for Wolf given the fact his xFIP and FIP have been better than that over the past 3 seasons. Looks like it is pulling way too much weight from the ERA in 2007, an ERA that peripherals didn't agree with. At first glance, I had the same thought. I don't think CHONE is relying on ERA too much, though. I think it's the aging adjustment. A 5/3/2 weighted average of Wolf gives us: 184 IP 4.17 ERA 4.03 FIP Even if you just start with his 3 year weighted average FIP, you need to adjust for age. In 163 innings, that could be 4 extra earned runs, which is an extra .21 ERA. In reality, it's really a combination of more earned runs and less IP, though. (As an aside, keep, in mind, that's a park neutral projection. Wolf is a flyball pitcher going to a HR park (1.05 HR factor)). <-- nested parenthesis are OK, right? http://forum.brewerfan.net/images/smilies/smile.gif Using your IP guesses and a 4.2 ERA for Wolf, he's worth about 20% more than Harden and getting paid about 20% more in 2010. So this could easily cost them $11M if he plays most of the season and gets hurt late. Like a 3.95 ERA in 180 IP? Sounds worth it to me. Rangers aren't making the playoffs anyway, right? http://forum.brewerfan.net/images/smilies/smile.gif

We have Wolf for three years which mean Doug doesn't have to go out and build a new rotation every offseason. Instead, he'll have a starting pitcher that's projected to have something like a 4.7 ERA in 2012 for $11 mil. Now, there's lot's of uncertainty in even projecting his 2010 performance, so I won't pretend that 2012 projection doesn't have a huge uncertainty. But when you sign a guy to a long term deal, that's exactly what you have to do. Unless Melvin knows something about Wolf that suggests he won't get worse from age 33 to 36, above average production is not something he can count on. Well, I can't prove anything, I can only provide you the data and methodology I used to come to that conclusion. I first started with the starting rotation's CHONE ERA projections: Gallardo: 3.60 Wolf: 4.42 Bush: 4.64 Davis: 4.69 Parra: 4.75 For an NL team, there's about 960 starter innings per year (6 IP x 162 starts). It would be foolish to presume no injuries, so I'm giving each 5 starters 160 IP and the remaining 160 innings to a replacement level pitcher (Some AAA scrub with 5.5 ERA). Indeed, if you add up the CHONE projected IP for the 5 starting pitchers, you are left with 165 innings, so it works well enough. Averaging the 5 starters and one replacement level pitcher gives us a 4.6 ERA. The NL average starting pitcher ERA last year was 4.32. It wouldn't be a bad rotation (worlds better than last year's) but still below average. I really don't know what to make of this statement (Harden as valuable as Wolf for 2010). I guess I just disagree with it outright. Again, I started with CHONE projections: Harden: 3.95 ERA, 132 IP Wolf: 4.42 ERA, 163 IP We want to know which performance is the more valuable one. We obviously can't just compare runs given up, since it would reward less innings, not more. We need a baseline. The lowest you really can go is the AAAA scrub reject replacement level starting pitcher. The assumption is that at that performance level, the supply outstrips demand, so you never have to settle for lower production. A full season of a replacement level starter is something like a 5.5 ERA and 195 IP (assuming 1/5th of the starts at 6 innings/start). He'll give up 119 earned runs. So to measure earned runs above replacement, we simply see how many more runs a replacement level pitcher would have given up for the innings the starter pitched: ERAR = (5.5 - ERA)*IP/9 The assumption is, if the pitcher in question isn't pitching, the replacement level pitcher is. So if Harden "only" averages 5-1/3 IP/GS, we assume the crappy long reliever picks up the additional 1 IP, or whatever. If a manger wants to instead put the load on the better part of his bullpen, that's his decision. To roughly convert to runs above replacement, you can just divide by .92 (something like 92% of runs are earned on average, if I remember right). So, the lower the pitcher's ERA, the more valuable he is. And provided he's better than replacement level, the more IP, the more valuable. Anyway, this is what we get for the two pitchers in question: Harden: 24.7 RAR Wolf: 21.2 RAR Now, I'm sure some are crying foul over Wolf's IP projection. "He's an innings eater" they'll scream (god, this reminds me of the Suppan signing). I don't know what goes into CHONE's IP projection but I assume age, previous IP and regression to the league mean are the primary variables. If you assume 600 IP for Wolf over the contract, I can almost guarantee you that you will be disappointed. But let's say 163 IP is too pessimistic. We have to set Wolf's IP to 190 innings before his 4.42 ERA is as valuable as Harden's 3.95 ERA is in 132. If you just think Harden's projection is wrong, that's fine. I won't say you are wrong. Which pitcher would you rather have at 3 years and 30 million? Wolf or Harden? An answer of neither isnt acceptable. What is the utility of even asking that hypothetical? I don't WANT either because BOTH are risks long term. Harden is more injury prone and Wolf is barely above average at age 33. I'll take the equal odds of Harden pitching 50 or 180 IP for a $7.5 mil commitment over comitting $30 mil to Wolf.

Why is every other team getting an economic downturn discount but the Brewers? Harden has as much value as Wolf does in significantly less innings. They are roughly worth the same this year. WAR measures against a scrub (5.25 ERA) pitcher, so innings eater don't have some hidden value. And even with a Doug Davis, the starting rotation is below average.

CHONE projects for Vargas to have a 4.1 ERA, so maybe we already have signed him. NL relievers had a 4.0 ERA last year. The fact that Hawkins might beat that by only a quarter run over 60 IP has limited value. That's about a 2 run improvement. It will be in a higher lev. situations, so maybe it's closer to 3 or 4 runs. It will be interesting to see what Vargas was signed for.

I think was more true when the Brewers stunk. It's still probably a factor but what are we talking about here? If it is even 10%, I would be surprised. I don't think any of us really know, though. If Melvin really needs to spend 10% more than anyone else, that suggests that maybe he shouldn't be signing high profile, 37 year old reliever in the first place. I'd rather get a 4.0 ERA/60 IP reliever for $1.5 mil and be done with it, personally. And I think there's a couple people in this thread that are getting fooled by his ERA from last year. He won't touch that this year. I'm thinking 3.6 ERA. I see his velocity is still great, so hopefully I'm seriously underestimating his worth.

I see a 3.65 FIP and 63 IP/year. He also turns 37 this year. It's not like this is a big contract but what's the point? For the Brewers to compete in the coming years, Melvin needed to do something special. Instead, he signed a couple of guys for more than they are worth (IMO, obviously). Anyone can go into a car dealer and buy a $20k car for $22k. Melvin needed to get a $25k car for $22k.

Was melvin tricked into overvaluing Hawkins based on his flukish 2010 LOB% of 91%? He's what, a true 3.7ish ERA reliever? This better not be more than 2/$6 mil. Probably wil be. EDIT: $7.5 mil, huh? Guess I don't see it but owell.

Can you find one person who has said this is the only option? The opinion of those who like the deal seem to be that this is the best option. You posted this about 15 minutes after I had edited my post to fix the misquote. http://forum.brewerfan.net/images/smilies/smile.gif I think his career numbers represent a larger sample and therefore are more relevant and his career OPS is .105 above average. It's a doubel edged sword. A larger sample is preferable but "true talent" is always changing. Projections generally use the last 3-5 years, weighting the oldest years 5-10%. What Wolf did in 2004 doesn't tell us much anymore.

That's one less of these theoretical "ace potential injury guys" that Melvin could go after and pretend they're 100% assured to be able to pitch most of the season. No one is assuming Wolf is 100% assured to pitch most of the season, so talk about a straw man argument. You here many suggesting signing a pair of those kind of pitchers, since you could possibly pencil them in for 180+ innings. As an added bonus, Wolf is one of the better hitting pitchers in baseball with a .181 career average. NL P, 2009: .138/.179/.175/.354 Wolf, 07-09: .154/.208/.218/.426 A .071 OPS upgrade over 70ish AB? Better than the alternative, I guess.

Quick and dirty projection. Just taking a weighted average of Wolf's last 3 years (50%/30%/20%): 184 IP 4.17 ERA 4.03 FIP Averaging ERA and FIP gives us a 4.10 ERA. He's also a year older, so you need to add at least .2 ERA (and that's probably too low). So we end up with a 4.3 ERA projection for 2010. A couple of other projections: CHONE: 4.42 ERA, 163 IP James: 3.87 ERA, 200 IP (I think James has lost his marbles http://forum.brewerfan.net/images/smilies/smile.gif) Let's say we are being optimistic and not only assume that over the course of the contract, the .2 ERA/year drop continues but that he'll also average 180 IP/year: 2010: 4.3 ERA, 190 IP 2011: 4.5 ERA, 180 IP 2012: 4.7 ERA, 170 IP If he's averaging $10 mil/year, he's probably worth that in 2010. I can't see a good argument being made that he should be expected to earn that in 2011 or 2012, however. My rough estimates puts Wolf worth something like $9/$7/$5 over the next 3 years. So, 3/$21 mil. As far as I'm concerned, Melvin spent around $9 mil more than Wolf is worth. I hate the argument that Melvin had to do something, as if this was truly the best option left (on December 9th). Even if EVERY above average starting pitcher who wanted a multi-year deal ends up being overpriced, sign 1 or 2 average arms coming off injury for $7-$9 mil per. The Penny deal looks pretty good to me, for instance. Hell, trade for a slightly below average pitcher (4.5ish ERA) and invest the $10 mil on offense (5 extra runs saved by pitching/fielding IS equal to 5 extra runs scored, as far as I can see). There are always options, especially when you are trying to upgrade from replacement level. I don't want to overstate my unhappiness for this signing. "Wasting" $9 mil over 3 years does not begin to compare to the Suppan deal. It makes the 2010 Brewers much better and it's not going to cripple the franchise overall. I just don't think this is the kind of signing a well run franchise with very limited funds should be making. I think Melvin could have done better.

That's just for base advancement.

Of course there's a correlation between a SB and runs scored. Of course it's going to look insignificant when compared to the correlation between runs and a stat that combines walks, singles, doubles, triples and HRs. That said, I share your skepticism with regard to the wisdom of this new team philosophy. I agree, the negative value of a CS seriously cuts into the value of a stolen base attempt. A good running team needs to select their SB attempt situations carefully. Not only should the attempt have a reasonable chance for success, it should occur more often when the positive value of of a SB is higher and the negative value of a CS is lowest. For example, with a runner at 2B, 2 out and down by 2 runs, it's foolish to attempt a SB because the payoff is fairly insignificant. Tie game in the 9th with 2 out, speedy runner at first and a light hitting batter at the plate, that might be a great time to try. If Macha is selectric in his attempts, the offense can get a slight boost. If not, the Brewers will abandon their new philosophy by June. I hope Macha is studying up.

I'm still not understanding why you wouldn't use actual HR count for pitchers. Again, we are not trying to estimate talent, just performance. A pitcher gave up the HR by himself, without his defense (most likely) influencing it in any way. We just use the average value of that HR (just like we do for batters). LOB% is no more random than HR/FB% is. The way fangraphs defines its "value" metric, it doesn't care if something is random or not. It only cares about two things: 1. Is it context neutral? (we ignore base-out situation and instead use the average run value of batting events) 2. Over a season, could the aggregate performance have been influenced by a potentially biased external factor? (batters face an average defense, so no issue; in contrast, pitchers could possibly have a very good or bad defense). Now, I think fangraph's approach yields a number that while not necessarily wrong, has little practical value. It just tells you what that performance was worth, had bought it on the open market. That's it. So it can tell you if a trade ended up working out for a team but not necessarily whether the GM made a mistake at the time of the signing. If a stud pitcher falls apart after being signed, the signing ended up being terrible for the team but might not have been bad at the time of its signing.

We should sign Wolf to a 3 year deal at $1, $1 and $30 mil! http://forum.brewerfan.net/images/smilies/smile.gif