A little OOTP 11 update...

[Vinny P.]

Originally Posted by timlowery

Provided he has decent teammates....

Ah, but of course. And plays in a hitter-friendly park as well.

[Vinny P.]

Originally Posted by mikev

Well, first off, just because you don't look at a stat doesn't mean it shouldn't be included. I never look at play by play, manage single games, play historical leauges, or play real MLB rosters. Should that all be removed just because I don't use it? Obviously not.

The whole point of stats is to have information. As you stated, the players in OOTP aren't real, they are just lines in a database. The whole point is to get stats and numbers. Why on earth would you NOT want more stats and numbers out of a game designed to give you stats and numbers?

To speak for your crowded statistical output: Remove VORP and the current defensive stats. Replace with wOBA, UZR, UZR/150, and WAR. Done.

I dunno about VORP, but I certainly do know that the current defensive stats should stay. I do like to know the hard numbers of how many errors my fielders are committing. Having UZR alone would be kind of like just having BA, without knowing how many hits a player has racked up.

As for VORP, I know it is used quite a bit IRL, but I certainly don't rely on it in OOTP. Having said that, it is just interesting to know a player's VORP. I look at it once in a while purely out of curiosity.

[darkenigma510]

This may be a stupid question, but what is the difference between WAR and BP's WARP? Are they the same? And if not, which is considered more reliable?

[mikev]

Originally Posted by Nihilianth

I dunno about VORP, but I certainly do know that the current defensive stats should stay. I do like to know the hard numbers of how many errors my fielders are committing. Having UZR alone would be kind of like just having BA, without knowing how many hits a player has racked up.

I was talking about Range Factor, not Innings played and errors. Sorry.

As for VORP, I know it is used quite a bit IRL, but I certainly don't rely on it in OOTP. Having said that, it is just interesting to know a player's VORP. I look at it once in a while purely out of curiosity.

VORP isn't hardly used at all anymore actually, since there are much better and more complete stats available.

Originally Posted by darkenigma510

This may be a stupid question, but what is the difference between WAR and BP's WARP? Are they the same? And if not, which is considered more reliable?

Not a whole lot. I think they have slightly different weights given to certain stats like doubles and HRs, but I'm really not a big fan of the BP stats since they won't actually tell you how they're calculated.

wOBA is easy to calculate. There is a simple formula for it. UZR, not so much, but at least it's available. It's a more complex calculation and requires a ton of batted ball data. Still, though, you CAN go get the info and figure it out on your own if you want.

That is, imo, enough to make me use wOBA, UZR, and WAR as the most complete set of advanced stats out there now. At least until something better comes along.

[Vinny P.]

Originally Posted by mikev

I was talking about Range Factor, not Innings played and errors. Sorry.

Oh! My bad! Than I totally agree. Do away with RF altogether. Like I have been saying in this thread, that is a totally stupid and useless stat anyway, which tells me absolutely nothing. So my CF has a lower RF than other CFs in the leage. Who cares? My pitchers are mostly groundball pitchers anyway. In fact, I end up with the highest GB% in whatever league I am playing most seasons anyway. So RF, to me, I really see no use for, and honestly, have no idea how anyone would even find it in the least bit "interesting."

I suppose you COULD use it in some arbitrary way to see just how much of an impact the GB-FO% of your pitchers has on indivual fielders, but even then, it really doesn't say a whole lot.

VORP isn't hardly used at all anymore actually, since there are much better and more complete stats available.

Yeah, I see what you mean. Still though, I have always thought that VORP was just a fun stat. I never really took it seriously, and it certainly does not impact any of my decision-making at all.

[mikev]

Originally Posted by Nihilianth

Yeah, I see what you mean. Still though, I have always thought that VORP was just a fun stat. I never really took it seriously, and it certainly does not impact any of my decision-making at all.

Me too. It's also pretty arbitrary, and it's not really easy to determine who/what is good.

With WAR, it's very simple. Even if you don't calculate it out yourself, or aren't big into stats, it's rather simple to explain to someone the concept of runs above average, and that roughly 10 runs is equal to 1 win.

It's also easy to just look at wOBA for a player, because it's scaled to OBP, and everybody is already conditioned to know what league average OBP is, and what really really good OBPs are, and that sub .300 is pretty awful.

Now, me, I'm kind of a stat geek so I like all the stats I can get my hands on, but at the same time I think it's absolutely fantastic that guys like Tom Tango are coming up with advanced statistics that they scale towards common stats that have been around forever. It's brilliant.

[fhomess]

To get back to whar Markus actually posted:

Originally Posted by Markus Heinsohn

when feeder leagues are used, undrafted/unsigned HS players may go to college.

I'm wondering if we can get some more detail on this. Currently, players are eligible for the draft after they leave HS or College. What this means is that people of often setup multiple HS and College leagues to get a variety of ages into their drafts, rather than just 18 and 21 year olds. So if I've got 4 HS leagues and 4 College leagues, which college league will the undrafted HS players go to? Can the HS players who are undrafted go to independant leagues if none of the College teams want them?

[thbroman]

Originally Posted by fhomess

This is where I think we might disagree slightly. I don't think we should only ask Markus to include the stats that are real by your definition (impacted directly by ratings, vs being the result of random chance not related to the underlying algorithms). I believe we should consider including any stat, whether real or arbitrary, if it increases immersion. Immersion increases when the game looks more like what we get in real life. Immersion also increases when a GM or manager might reasonbly be expected to change their behavior because of the analysis of a player based on a stat.

I also think it's worth remembering that the game has four different levels of stats display. I would not expect to see most of the arbitrary stats on the lowest stats display level, and would expect to see most of them only on the very high detail setting.

All that said, if we've got two stats we're asking Markus to add, and he only has time for one, I'd generally go with the stat that is real rather than the arbitrary one, as long as the arbitrary one isn't in significantly more common usage in real life.

Nice summary of the discussion, Frank, although I actually wasn't objecting to the inclusion of any stats. But you are completely right. Stats are there to be enjoyed or not, as the player chooses. I guess I was more thinking out loud about stats as "symptoms" of a player's ability and their differing roles in the game and in real life.

[lighthousekeeper]

Markus, all that I ask is that when you do implement wOBA, that you publish the approach that you are using to calculate the statistic. That way, all of us know exactly what we are getting when we see the stat. wOBA is derived from a much more complex source of data as opposed to something like OPS, so there's a good chance that this stat will be incorrectly determined.

After reading Tom Tango's "The Book" (which is the source of wOBA), I see how a truly accurate calculation of wOBA would not use the same fixed linear run weights for each event (e.g. HR = 1.397, etc) to simulate the wide variety of different run-scoring eras in the history of baseball.

The commonly cited formula for wOBA is something like this:

wOBA = ((0.72 x NIBB) + (0.75 x HBP) + (0.90 x 1B) + (0.92 x RBOE) + (1.24 x 2B) + (1.56 x 3B) + (1.95 x HR) / PA

But this is based on an a modern MLB environment in which ~0.555 runs are scored per inning. But since OOTP is so flexible and capable of simulating a variety of baseball environments (from a deadball era game, to an extreme live-ball era game), you have to anticipate that there will be situations in which wOBA needs to be calculated for seasons where ~0.2 runs are scored per inning or ~0.7. The run-scoring environment has a huge effect on the relative run values of each event type (so for example, in a deadball era type environment, a homerun's value relative to a single is much higher as opposed to the modern MLB game - which is intuitive).

Here's a chart from Tango Tiger's blog which shows the run values for a variety of run scoring environments.
Sabremetrics 301: Custom Linear Weights

Again, please don't simply use the oft-cited wOBA formula or else you will actually be doing a disservice to those that simulate deadball era games.

[lighthousekeeper]

Originally Posted by mikev

wOBA is easy to calculate.
...
That is, imo, enough to make me use wOBA, UZR, and WAR as the most complete set of advanced stats out there now.

wOBA is also easy to calculate incorrectly (see post above). I can see wOBA as being a good substitute for OPS but that is it.

Some things that you need to keep in mind regarding the limitations of wOBA so that you can set the approriate expectations for this stat:

1. It doesn't incorporate baserunning at all (like OPS)
2. It doesn't incorporate the players position or fielding contributions (like OPS)
3. It isn't park adjusted at all (like OPS)
4. It isn't intrinsically era adjusted (like OPS) (Markus will need to code some custom cross-era adjustments)

These aren't criticisms but simply letting people know what they are (and aren't) getting with this stat. In this light, wOBA can't tell us some stuff that VORP or OPS+ can.

I do believe that wOBA is superior to OPS though, as it provides better weighting for each offensive contribution.

[mikev]

Uh, no. There's nothing incorrect about that, and there's nothing hard about calculating it.

[lighthousekeeper]

Originally Posted by mikev

Uh, no. There's nothing incorrect about that, and there's nothing hard about calculating it.

If you think wOBA is so easy to calculate, then answer this: what is the wOBA of a batter who goes 2 for 20, with 1 double and 1 home run?


.

...it's a trick question because you can't answer the question with the information given. You need to know how many average runs per game are scored in the league. So the first question becomes: how to determine what the average runs/game are for the league? Does Markus use historical data from the ootp league or does he estimate based on the league modifiers in the League Setup screen (remember that we can use the League Setup screen to create any variety of wacky run scrogin environments)? Let's say for argument's sake that Markus codes it to estimate the expected runs/game based on the league modifiers for the league. Let's say this results in estimating 4.3 expected runs/game. Well then what? What run values do you associate with HR, 3B, 2B, etc? You could refer to http://www.tangotiger.net/customlwts.html, but this only provides estimates for the 4 R/G or 5 R/G - nothing for 4.3 R/G.

If you try to apply the wOBA linear weights derived from modern MLB to another era such as deadball, wOBA may very well be less accuate than OPS. Just a word of caution.

[Vinny P.]

Originally Posted by lighthousekeeper

If you think wOBA is so easy to calculate, then answer this: what is the wOBA of a batter who goes 2 for 20, with 1 double and 1 home run?


.

...it's a trick question because you can't answer the question with the information given. You need to know how many average runs per game are scored in the league. So the first question becomes: how to determine what the average runs/game are for the league? Does Markus use historical data from the ootp league or does he estimate based on the league modifiers in the League Setup screen (remember that we can use the League Setup screen to create any variety of wacky run scrogin environments)? Let's say for argument's sake that Markus codes it to estimate the expected runs/game based on the league modifiers for the league. Let's say this results in estimating 4.3 expected runs/game. Well then what? What run values do you associate with HR, 3B, 2B, etc? You could refer to Sabremetrics 301: Custom Linear Weights, but this only provides estimates for the 4 R/G or 5 R/G - nothing for 4.3 R/G.

If you try to apply the wOBA linear weights derived from modern MLB to another era such as deadball, wOBA may very well be less accuate than OPS. Just a word of caution.

Instead of putting actual weights on each single, double, triple, and homerun initially, wouldn't it make sense to first calculate the actual average number of each, and have another formula to then use these averages to then figure out what weights to give?

[yajeflow]

Originally Posted by Markus Heinsohn

We will have a big announcement in a couple of weeks, so stay tuned!

this was written on february 3. i just went through all 19 pages looking for this announcement to see if it came early.

3 + 14 = 17

today is february 17. two weeks have passed. gimme gimme gimme.

xoxoxoxoxo,

yajeflow

[Jeffy25]

well I guess it is being talked about here...Which is awesome. I am hoping for WAR and more sabr stats

[Wolfbane]

Originally Posted by fhomess

To get back to whar Markus actually posted:


I'm wondering if we can get some more detail on this. Currently, players are eligible for the draft after they leave HS or College. What this means is that people of often setup multiple HS and College leagues to get a variety of ages into their drafts, rather than just 18 and 21 year olds. So if I've got 4 HS leagues and 4 College leagues, which college league will the undrafted HS players go to? Can the HS players who are undrafted go to independant leagues if none of the College teams want them?

This is one of my biggest questions also, as a big fan of the player development aspect of the game. Can we select certain high school leagues to feed into several different college leagues, or is it strictly linear where one or more HS feeders can only go into one college. Also, can a HS feeder send unsigned players to more than one college league. Or is it the players select the college, since they are FAs before the draft(I believe they are anyways). I would love to be able to send unsigned HS players from a multiple leagues into one or more college feeders to control foreign player #s more cleanly under one college league in my Euro League.

A couple other notes. Usually, undrafted HS players are by far the worst ones. They would only be worse in a college league, with the exception of the few HS players who fail to sign. This would prove problematic for me in that the college feeder would end up being partially filled with undraftable retreads, who will likely go undrafted again. Also, will they develop at all from HS to college in terms of ratings, or no?

One more question, you said there would be "post-draft contract and bonus negotiations". Does that mean there are going to be monetary values attached to minor league contracts now. Maybe the optimist in me is reading into that too much, but I would love to have AAAA/replacement level talent with upside to be stockpiled with $100,000-$500,000 contracts. This might be the only missing feature I desire now with the new feeder system in place.

Finally, I love the implementation of FIP, wOBA, UZR clone, and OPS+. I would avoid putting in WAR yet for there is still ongoing debates over calculating reliever and starter WAR, especially over reliever replacement level. tRA or now tERA's formula is not publicly available yet unless they have recently revealed it. I read Lookout Landing and Fangraphs regularly, yet have seen nothing to that effect besides a post explaining what it quantifies. Furthermore, I want to say I love what you have done with the game. The work with feeders, stats, player development, scouting, individual pitch rating etc... has made the game so much better than when I first bought OTP07.

Also, props to who came up with this quote "Stats cause nothing, explain nothing and predict nothing. But stats are a valueable indication to look for reasons. They are a maidservant for the mind". Whether it is original or not, I wish I had that on hand in the past to deal with countless numeraphobic historians.

I can never make a short post in this forum, dam.

[Wolfbane]

Originally Posted by lighthousekeeper

If you think wOBA is so easy to calculate, then answer this: what is the wOBA of a batter who goes 2 for 20, with 1 double and 1 home run?


.

...it's a trick question because you can't answer the question with the information given. You need to know how many average runs per game are scored in the league. So the first question becomes: how to determine what the average runs/game are for the league? Does Markus use historical data from the ootp league or does he estimate based on the league modifiers in the League Setup screen (remember that we can use the League Setup screen to create any variety of wacky run scrogin environments)? Let's say for argument's sake that Markus codes it to estimate the expected runs/game based on the league modifiers for the league. Let's say this results in estimating 4.3 expected runs/game. Well then what? What run values do you associate with HR, 3B, 2B, etc? You could refer to Sabremetrics 301: Custom Linear Weights, but this only provides estimates for the 4 R/G or 5 R/G - nothing for 4.3 R/G.

If you try to apply the wOBA linear weights derived from modern MLB to another era such as deadball, wOBA may very well be less accuate than OPS. Just a word of caution.

Not to beat a dead horse here. But.... the same issue would come up when looking at high run or low run environment in different minor leagues. wRC+ is in some ways a better metric than wOBA for that reason. My initial though was that Fangraphs wOBA found a way to adjust for league offensive levels, but it does appear to be so after a cursory glance.

[Kobeck]

Originally Posted by actionjackson

If you lack the creativity to come up with 17,500 or so random names on your own , there's a great site here.

my pirate ship is gonna be named The Lustful Whore of the Ocean, apparently

[Goody]

Sales. Its not all about what will make sales skyrocket. That is obvious when you look at the product, because what is obvious is that Markus has created this game because its something he loves.

Now its true that to keep being able to do what he loves and make enough profit to be able to focus most of his time on it...he'll have to reach some certain bottom line...of course.

But if Markus ever came out and said he can't continue with this because of financial reasons...I know for a fact many in this loyal community would step up to help if they have the means.

If it weren't for this product, many many people would have lost out on hours of enjoyment in their hobby. We're blessed to have somebody to talented enough to make it possible for us. I'm sure he could be working for a large software company and pulling in a nice salary for making somebody else's vision come true so they can make a lot of dough for themselves...with less risk for himself but also less reward.

Its been proven time and again. He does what's logical for realism and simulation...and is willing to listen to input. Let's have some trust in this and not just cheapen it to "what brings in the most $".

[Goody]

Oh.... and what the Chone Figgins is "WAR"??