What will it take for you to be 'okay' going into the season?

[TwistedLogic]

Again true talent or expected vs. actual.

 

If I can invent a load of ******** like "truGrit" and come up with, what everyone here will agree with is a more accurate result for the Astros 2014 season, how does that prove the projection system's legitimacy any more?

 

Looking at this chart, true talent tells us that the Astros should win 70 games, the Reds are expected to lose 84 games and that the Royals and Blue Jays are the 5th and 6th best team in the major leagues respectively.

 

I'm sorry if you don't agree with my opinion here, and I really respect and appreciate the work and research you've accomplished, because most of it is incredibly good. But this particular concept is complete ********. It's a waste of time. It'll never be ditched, because as I said, sabermetricians are desperate to be able to predict the future, but it holds absolutely no merit. Just another thing to do in the off-season, no different in significance than rosterbation or trade ideas.

[GD]

http://persephonemagazine.com/wp-content/uploads/2012/10/kingofpopcorn1.gif

[TwistedLogic]

http://persephonemagazine.com/wp-content/uploads/2012/10/kingofpopcorn1.gif

 

This f***ing picture never gets old LOL. Oh and screw you.

[dineke]

When I brought home a C on my report card, I told my parents my true talent is an A.

[Abomination]

But this particular concept is complete ********. It's a waste of time. It'll never be ditched, because as I said, sabermetricians are desperate to be able to predict the future, but it holds absolutely no merit. Just another thing to do in the off-season, no different in significance than rosterbation or trade ideas.

 

No, it isn't. It's an evaluation tool to help us see past personal bias to determine what a given team should be able to attain all things being equal. It can't account for things like freak injuries (or freak good health), or trades that haven't happened yet for example. Using them in hindsight is a misapplication of them, as is believing that they are designed as some sort of absolute predictor of the future.

[TwistedLogic]

When I brought home a C on my report card, I told my parents my true talent is an A.

 

Yup I just underperformed and got unlucky. But don't worry because Steamer says I will bring in two A's next semester to compensate.

[Abomination]

When I brought home a C on my report card, I told my parents my true talent is an A.

 

Might have been smarter to say you over-performed, otherwise they probably told you that you were grounded until you starter working hard enough to realize your true talent...

[NorthOf49]

Yup I just underperformed and got unlucky. But don't worry because Steamer says I will bring in two A's next semester to compensate.

There are a lot of players left unsigned, in effect lots of wins up for grabs.

 

Your projection of 108 losses banks on the Astros acquiring a disproportionately small amount of those wins. You're saying such since they're likely to use 2014 as another rebuilding year. While that may be true, it would be misleading for projection systems to already have that baked into their sauce. If you dole out the up for grabs wins among all teams evenly, you'd probably get close to Steamer's 74.

[John_Havok]

If I can invent a load of ******** like "truGrit" and come up with, what everyone here will agree with is a more accurate result for the Astros 2014 season, how does that prove the projection system's legitimacy any more?

 

Looking at this chart, true talent tells us that the Astros should win 70 games, the Reds are expected to lose 84 games and that the Royals and Blue Jays are the 5th and 6th best team in the major leagues respectively.

 

I'm sorry if you don't agree with my opinion here, and I really respect and appreciate the work and research you've accomplished, because most of it is incredibly good. But this particular concept is complete ********. It's a waste of time. It'll never be ditched, because as I said, sabermetricians are desperate to be able to predict the future, but it holds absolutely no merit. Just another thing to do in the off-season, no different in significance than rosterbation or trade ideas.

 

Every front office uses projections. THey may use differing formulas, gut feelings or a Ouija board, but they all use projections to assess their teams true talent level. I suspect it's not projections that you take issue with, but moreso how Steamer presents theirs because you disagree with how they arrive at their numbers.

[Abomination]

THey may use differing formulas, gut feelings or a Ouija board, but they all use projections to assess their teams true talent level.

 

I bet Friedman could have some real fun with an Ouija board and a camera on April fool's day...

[TwistedLogic]

There are a lot of players left unsigned, in effect lots of wins up for grabs.

 

Your projection of 108 losses banks on the Astros acquiring a disproportionately small amount of those wins. You're saying such since they're likely to use 2014 as another rebuilding year. While that may be true, it would be misleading for projection systems to already have that baked into their sauce. If you dole out the up for grabs wins among all teams evenly, you'd probably get close to Steamer's 74.

 

Bud, the Astros were projected at 74 wins last year, going into the season, after those wins had been distributed.

 

Every front office uses projections. THey may use differing formulas, gut feelings or a Ouija board, but they all use projections to assess their teams true talent level. I suspect it's not projections that you take issue with, but moreso how Steamer presents theirs because you disagree with how they arrive at their numbers.

 

No my problem isn't with trying to make an educated prediction of the future by using informed estimates. Saying "Adam Lind is going to continue to be poor against left-handed pitching, we should account for at least one disabled list trip for Morrow, and we should have a serviceable backup for right field because it's getting to be a habit for Jose to miss time". That's a legitimate "projection". My problem is then sabers throwing their usual OCD-twist on it by trying to assign number values to every player based on this information, and then ridiculously combining it into a total win projection for a team. It becomes useless at that point.

 

Coming on this board and using that information to tell someone "we shouldn't worry as much because based on this, the Jays look to be an 87-win team", you yourself might know that what you just told this person is useless rubbish, because it is based on the fallacy of "all things being equal" (which is true 0% of the time in the real world) but the other person will likely take it to mean something else. It's grossly misleading and it is used in the wrong context almost 100% of the time on these boards. "Steamer projects this vs this, so we should go after this guy".

[John_Havok]

Saying "Adam Lind is going to continue to be poor against left-handed pitching, we should account for at least one disabled list trip for Morrow, and we should have a serviceable backup for right field because it's getting to be a habit for Jose to miss time". That's a legitimate "projection". My problem is then sabers throwing their usual OCD-twist on it by trying to assign number values to every player based on this information, and then ridiculously combining it into a total win projection for a team. It becomes useless at that point.

 

How do you think these projections are formed? By using exactly what information you just said. Adam Lind will likely perform poorly against LHP and pretty good against RHP, so using past performance as a guideline, you can make a very educated guess as to what type of value he will provide the Jays in 2013. Ditto for Morrow and Jose. Just because there's a WAR number attached to it doesn't make it voodoo, magic or uninformed guesswork.

[TwistedLogic]

How do you think these projections are formed? By using exactly what information you just said. Adam Lind will likely perform poorly against LHP and pretty good against RHP, so using past performance as a guideline, you can make a very educated guess as to what type of value he will provide the Jays in 2013. Ditto for Morrow and Jose. Just because there's a WAR number attached to it doesn't make it voodoo, magic or uninformed guesswork.

 

Attaching a WAR value to that specific player is not the problem. If someone says "based on all of this information, Adam Lind looks to be a 1.2 WAR player in 2014". I'm completely on board with that. That is an educated guess. It's not accurate, because it doesn't count for injuries, breakouts or regression, but we can conclude that we can all see him at or around that mark, while nobody will be surprised if he goes back to being below replacement level, or even posts twice as much WAR.

 

The problem is when you then take that very fragile number you just gave to Lind and put it into a calculator and hit the + sign and then enter equally fragile numbers that you assigned to Rasmus and Morrow. Now you're trying to put a predicted value to a guy who's been a different, but similar player in every year he's been in the majors (Lind), a guy who's sandwiching two horrible seasons with two excellent ones (Rasmus) and a reliever-turned starter that oozes with potential if he can only manage to stay on the field. You're already treading into the "virtually impossible to predict the outcome" waters with just these three players. Now you throw in the other 22 on the active roster, including the entire category of volatile relievers, the other 15 on the 40-man, the entire farm system, developmental staff, coaches, in-season trades, unforeseen circumstances (tu ere maricon), all of the other intangibles and pure dumb luck. You're telling me you can give me a number of wins a team is expected to accumulate in 2014, with a straight face?

 

Single-player projections make sense to evaluate that player in a vacuum. When you put Player A + Player B + Players C-Y = team wins, you've just given me an absolutely USELESS number. With all of the up-and-downs on each of those players (Lind posting half or double of what he's expected to), you can easily end up missing your mark by 15 wins (Dodgers, Astros), making it a flawed and again, USELESS number.

 

We all know this team is a last place team with this rotation

Nope.

Who are we better than

85-87 win team right now.

No we're not? I don't even know what our rotation looks like

Steamer thinks we are:

 

http://www.breakingblue.ca/2013/12/04/offseason-power-rankings-dec-4/

 

This exchange is a prime example of what I mean. Disregarding the fact that MohYou is a retard, this conversation proves exactly how poorly projections are used on here. Even if the people in question might not completely believe it, it is being heavily implied that the Blue Jays really are an 85-87 win team.

 

You are assigning a value taken from the mythical world of "all things being equal" and pasting it on one of the most volatile, injury prone teams in sports right now. That is not accurate, and in JFaS' own words, is not the actual. Yet then you guys portray it as the actual yourselves.

[John_Havok]

Attaching a WAR value to that specific player is not the problem. If someone says "based on all of this information, Adam Lind looks to be a 1.2 WAR player in 2014". I'm completely on board with that. That is an educated guess. It's not accurate, because it doesn't count for injuries, breakouts or regression, but we can conclude that we can all see him at or around that mark, while nobody will be surprised if he goes back to being below replacement level, or even posts twice as much WAR.

 

The problem is when you then take that very fragile number you just gave to Lind and put it into a calculator and hit the + sign and then enter equally fragile numbers that you assigned to Rasmus and Morrow. Now you're trying to put a predicted value to a guy who's been a different, but similar player in every year he's been in the majors (Lind), a guy who's sandwiching two horrible seasons with two excellent ones (Rasmus) and a reliever-turned starter that oozes with potential if he can only manage to stay on the field. You're already treading into the "virtually impossible to predict the outcome" waters with just these three players. Now you throw in the other 22 on the active roster, including the entire category of volatile relievers, the other 15 on the 40-man, the entire farm system, developmental staff, coaches, in-season trades, unforeseen circumstances (tu ere maricon), all of the other intangibles and pure dumb luck. You're telling me you can give me a number of wins a team is expected to accumulate in 2014, with a straight face?

 

Single-player projections make sense to evaluate that player in a vacuum. When you put Player A + Player B + Players C-Y = team wins, you've just given me an absolutely USELESS number. With all of the up-and-downs on each of those players (Lind posting half or double of what he's expected to), you can easily end up missing your mark by 15 wins (Dodgers, Astros), making it a flawed and again, USELESS number.

 

 

 

This exchange is a prime example of what I mean. Disregarding the fact that MohYou is a retard, this conversation proves exactly how poorly projections are used on here. Even if the people in question might not completely believe it, it is being heavily implied that the Blue Jays really are an 85-87 win team.

 

You are assigning a value taken from the mythical world of "all things being equal" and pasting it on one of the most volatile, injury prone teams in sports right now. That is not accurate, and in JFaS' own words, is not the actual. Yet then you guys portray it as the actual yourselves.

 

Once again, people who have understanding of the projections fully understand that 85-87 is not gospel and the margins of error are absolutely applicable. You have to stop putting one or two people's posts as "you guys portray it as the actual yourselves." If you take issue with how one or two people uses Steamer projections in their posts, take issue with the person, not the projections.

 

And yes, the projections do adjust for possible injuries do try to account for declines in performance. Otherwise Joe wouldn't be projected for less than 600 PA. Breakouts are highly unpredictable, but that's part of why the projections themselves have a listed margin or error. If someone claims the projections are an exact science, they're an idiot, but again, as indicated above that's a failure of the person, not the projections.

[The_DH]

How can you not have our ROY in the rotation?

 

Attaching a WAR value to that specific player is not the problem. If someone says "based on all of this information, Adam Lind looks to be a 1.2 WAR player in 2014". I'm completely on board with that. That is an educated guess. It's not accurate, because it doesn't count for injuries, breakouts or regression, but we can conclude that we can all see him at or around that mark, while nobody will be surprised if he goes back to being below replacement level, or even posts twice as much WAR.

 

The problem is when you then take that very fragile number you just gave to Lind and put it into a calculator and hit the + sign and then enter equally fragile numbers that you assigned to Rasmus and Morrow. Now you're trying to put a predicted value to a guy who's been a different, but similar player in every year he's been in the majors (Lind), a guy who's sandwiching two horrible seasons with two excellent ones (Rasmus) and a reliever-turned starter that oozes with potential if he can only manage to stay on the field. You're already treading into the "virtually impossible to predict the outcome" waters with just these three players. Now you throw in the other 22 on the active roster, including the entire category of volatile relievers, the other 15 on the 40-man, the entire farm system, developmental staff, coaches, in-season trades, unforeseen circumstances (tu ere maricon), all of the other intangibles and pure dumb luck. You're telling me you can give me a number of wins a team is expected to accumulate in 2014, with a straight face?

 

Single-player projections make sense to evaluate that player in a vacuum. When you put Player A + Player B + Players C-Y = team wins, you've just given me an absolutely USELESS number. With all of the up-and-downs on each of those players (Lind posting half or double of what he's expected to), you can easily end up missing your mark by 15 wins (Dodgers, Astros), making it a flawed and again, USELESS number.

 

 

 

This exchange is a prime example of what I mean. Disregarding the fact that MohYou is a retard, this conversation proves exactly how poorly projections are used on here. Even if the people in question might not completely believe it, it is being heavily implied that the Blue Jays really are an 85-87 win team.

 

You are assigning a value taken from the mythical world of "all things being equal" and pasting it on one of the most volatile, injury prone teams in sports right now. That is not accurate, and in JFaS' own words, is not the actual. Yet then you guys portray it as the actual yourselves.

 

 

It makes me think one thing: numbers in the hands of some people is a very dangerous thing.

[Jays]

It makes me think one thing: numbers in the hands of some people is a very dangerous thing.

 

Not sure why you quoted me in there.

[xposbrad]

Twistedlogic, I agree with you 1000%. You are backing up everything you say. The proof is in the pudding.

[fatcowxlive]

Not sure why you quoted me in there.

 

Because of them ROY numbers!

[Jays]

Because of them ROY numbers!

 

Stroman for ROY 2014

[TwistedLogic]

Once again, people who have understanding of the projections fully understand that 85-87 is not gospel and the margins of error are absolutely applicable.

No, the margins of error are so vast that it essentially makes the information worthless. It gives you the vaguest estimate from which you could add or subtract 10 wins to get the area in which your team is likely to land (and funnily enough, even then you have a good chance at being wrong; fangraphs projections were off by 10 wins or more on 9 of the 30 teams last year. 13 more of the remaining 21 teams were off by at least 5 wins. 5 wins can be the difference in being the top seed in your division versus sitting at home in October).

 

The volatility of the information makes it so that it barely tells you anything, let alone acting as "gospel". If you average out the amount of wins that FanGraphs was off by last season, you get 7.4. Knowing that I have the number "85-87" and my team is likely to end up anywhere between 7 wins below that mark (78) to 7 wins above that mark (94) tells me nothing. When all is said and done next year, the projected standings based on Steamer or whoever else will have been about as accurate or useful as any ESPN writer's predictions.

 

So when a moron like MohYou comes along and says "this is a last place team" without providing any basis for that statement, the correct rebuttal isn't "no, it's an 85-87 win team", because that is just as absurd of a statement.

[TwistedLogic]

It's not as absurd though, it's unbiased. Do you not understand approximate true talent? If injuries did not exist, and players played their projected amount of time, the projections would have an R^2 of approximately 0.60. Don't quote me on that but that's what I think it is. Why do you keep referring to the Astros? September call ups and tanking aren't in consideration, which obviously affects the Stros.

 

Bottom line: If you want to say how many wins this team should win, you should use projections instead of pulling things out of thin air because projections are unbiased. If you think analyzing how many wins they should win is stupid, then don't do it.

 

Do you not understand realistic actuality? Injuries DO exist, and players don't always play their projected amount of time, and so the projections do NOT give you any useful information. Learning that "in utopia, this is what would happen" is fun and all, but it gives us nothing.

 

Bottom line: If you want to say how many wins this team should win, you should not; you have no way of knowing. You can use as many projections and data as you want, you will not get an accurate gauge on it because you can't quantify intangibles. It's the cold hard truth that sabers don't want to accept. You. Can. Not. Quantify. Intangibles. You can't put a number on a coach fixing Bautista's swing. You can't put a number on Rasmus' comfort level in a certain place affecting his performance. You can't calculate how injury prone a player is, and how much more injury prone he is than anybody else. You can't definitively say that the guy standing in the on-deck circle does not come into the mind of the guy on the mound, and ultimately affects the kind of pitch that the guy in the batter's box sees next. Maybe some day we will be able to calculate all of this, right now we can't.

 

I don't need to keep referring to the Astros, I've already mentioned the Dodgers, the Indians-Twins and the numbers for the rest of the league.

 

You are ignoring reality to throw out this number of 85 wins; maybe it would be an accurate estimate in a simulation-based environment. That isn't major league baseball.

 

Estimating how many wins a single player will accumulate in one season has merit, has use. Estimating how much WAR a collective baseball organization will accumulate is useless.

[TwistedLogic]

"Barring injury, Brandon Morrow can be an x-win pitcher in 2014. It is x as likely to happen because he has had x amount of injuries in the past. x amount of these injuries are real concerns because they are mechanics-related injuries in x-amount of ways, so while he might be an x-win pitcher, he can also possibly end up needing x surgery and be out for the rest of the season, providing no value at all."

 

http://www.imwsa.com/img/correct.gif

 

"Barring injury, Brandon Morrow, Melky Cabrera, Brett Lawrie, Jose Bautista, JA Happ, Colby Rasmus, Jose Reyes and Sergio Santos will be worth x amount of total wins".

 

http://jewellerysearch.com/images/button_cancel.gif

 

The first tells you something, the second is useless garbage.

 

You think it's useless, I (and many others) don't. It has uses, especially for a team knowing when to spend. The Mariners should not spend now, because they're not yet at the high impact point of the curve. The Jays should spend right now because they are at the high impact point.

 

So you would use a system that has an astronomically high fail rate, and missed each team's record by an average of 7.4 wins in 2013, to determine whether the most crucial point for your ball club is now here, and you're ready to spend or not?

 

I wonder if I'm the only one here that believes this, but I don't think ball clubs are out there deciding whether to be sellers or buyers based on Steamer, my friend. These teams have an infinite amount of data, they are using every available tool at their disposal, have lengthy injury histories, in-depth scouting reports, medical records, these teams know when to go for a win and when to stand pat. Us fans use Steamer.

[TwistedLogic]

You and me might frown or laugh at intangibles, and choose whether or not to believe it them, but teams do not. Teams take them heavily into account. That right there begins a massive separation between where a team thinks they are and where your Steamer list says they are. DeRosa didn't just get a job because he was a platoon bat, he got a job because he moonlights as a mascot. If a team believes that a good clubhouse chemistry is worth x-amount of wins, and actually has its own research and data to give credibility to that claim, that gives them a significantly different valuation of their ball club than what we are looking at. JFaS thinks the Mariners aren't ready to compete, so Jack-Z should call it off. It doesn't work that way. The Mariners have their own numbers, their own information, they think they're ready to compete. Now we can sit here and start another fruitless debate about why the Mariners are really competing, pretending that we know what's going on in their front office, or we can accept that there may be a projection system out there, or there may be a cumulative host of information, that tells us that the Mariners can compete in 2014.

 

A few seasons ago if the Athletics or the Oriole's GMs had suddenly decided to become buyers after the kind of seasons they'd been posting, people like us might have been saying that they aren't ready to compete. Well, it turned out that they were. People laughed at the Royals a season ago, saying that they were not ready to compete and would need to improve before making an all-in trade. We were correct, they didn't compete in 2013, but just a season later they've Steamer'd their way up to the #5 spot going into the next season.

 

Who are we to say that it's smart or dumb for the Mariners to go all-in right now? Even if it doesn't pay off in 2014, it doesn't mean it's an idiotic plan. It could easily pay off in 2014 once Cano is settled in and the Mariners have more knowledge of their core of young talent. The end goal is to get a ring. Two teams decided to go all-in last offseason: the Royals and the Blue Jays. People supported the Jays moves, saying they were a team that was on the cusp and needed that push over the edge, while people ridiculed the Royals, partly because they made a moronic trade, and partly because that moronic trade came at a time where they were allegedly not near contention. Twelve months later, as a result of those moves, the Royals are above the Jays in your beloved Steamer's projections, even if it is by just one spot.

[TwistedLogic]

They have better projection systems using hitF/x and PitchF/x, yes. Injury histories mean nothing to me unless someone has a chronic injury or a condition btw. Injuries are the smallest sample size of anything you can find.

 

Yes, I understand that, because I know that with the emergence of advanced statistics, the new and cool thing to do is to treat players as robots.

[TwistedLogic]

No one said they're robots, obviously there are drop offs and breakouts due to intangibles, but they all show up in the stats. With PitchF/x data and soon hitF/x data, we can tell if a "hot streak" is a change in talent or just random variance. Changes in talent obviously happen, and they may be due to these intangibles, but at the end of the day, the numbers don't lie. You just have to know what numbers to use (Jack Z doesn't).

 

Jack-Z isn't the only one working in his office, as much as we all wish that were true lol. And numbers don't have to lie when they can just as easily deceive

 

I guess the projections debate is one thing that we'll have to agree to disagree on. I think projections are a useful tool among others when you're looking at players in a vacuum, I just think they become insignificant when you want to ball an entire organization up into one number. The amount of wiggle room and ability to go up or down on each individual player's projection, when added all together, makes the final number too much of a variable. That is my personal belief, and I think it's a legitimate, sensible one.

 

Beyond that, everyone else has a right to their own opinion and some things make more sense or work better for others. At the end of the day, it's to each their own.

[TwistedLogic]

I do agree that each can have an opinion, though I enjoy numbers, and this might even be my job soon.

 

Also just ran an out-of-sample test on team changers for xxFIP and the results are so sexy, very happy right now, though TIPS may become obsolete. xxFIP wins handily from 0-200IP in this sample.

 

You're making excellent progress on your work if you've already significantly evolved a metric that got you national attention.

 

When you're talking about advanced statistics that give me more information on what has already happened, I'm all ears.

[TwistedLogic]

Holland was a guy I was begging for all the way back Soria was the closer in Kansas, too lazy to check when that was (2-3 years)? Kenley's another guy I always loved. Though I'll admit I was one out of the many (everyone?) that got blind-sided by Uehara. If you look at his numbers, anyone could have seen it coming, the guy's been and elite reliever since he stepped into the majors.

[kcjaysfan]

Holland was a guy I was begging for all the way back Soria was the closer in Kansas, too lazy to check when that was (2-3 years)? Kenley's another guy I always loved. Though I'll admit I was one out of the many (everyone?) that got blind-sided by Uehara. If you look at his numbers, anyone could have seen it coming, the guy's been and elite reliever since he stepped into the majors.

 

I've been up all night doing school work, so I'm going to be a pedantic twat. The Royals do not play in Kansas.

[Angrioter]

When I brought home a C on my report card, I told my parents my true talent is an A.

 

 

Gold dude