CVRC Question

[aburt19]

In setting the player pool to a single pool, it says too enter OF into F2 (QPOS). But QPOS is column D, not F. Do I change
the QPOS (column D) or column F (PPOS)?

[Todd Zola]

DAMMIT!!! I am so grateful this is only early January and you are extremely patient and understanding, you're like my personal beta-tester

I did so much copying and pasting myself when noodling around, I left the column headings all messed up.

Column F is read by the CVRC to do the calculations. It should he PPOS (Pool Position). This is where you do the manual changes.

Column E is used for positional rankings. It is MVPOS (Most Valuable Position). You can change it if you want the player to be listed at a different position that our designation. For instance, in some leagues Gordon Beckham already qualifies at 2B based on a "common sense" rule.

Column D is all the position the player qualifies based on the Waggoner rule book and is called QPOS. It is basically aesthetic and is only useful for leagues that use the original qualifications.

[aburt19]

That's because I'm going to be out of town on business for the next three days and need something to do during the plane
rides. It also doesn't hurt that I have no life .

[aburt19]

One other thing that I wondered about. Victor Martinez is listed at only 1B and not catcher. I would think that any player
with positive value that could qualify at catcher would be put at that position.

[Todd Zola]

yeah, should be catcher, not sure how that changed

[aburt19]

I was messing around with CVRC for pitching. The spreadsheet comes up as a 15 team mixed league. The default setting
for a 15 team mixed league for ERA is 4.424 and for WHIP is 1.414. When I change it to a 12 team AL only league the
default setting for ERA is 4.292 and for WHIP is 1.384.

Given that a higher percentage of available pitchers are rostered in a 12 team AL only league compared to a 15 team mixed
league, I would think that the ERA and WHIP for a 12 team AL only league would be higher than the 15 team mixed league
due to having to roster some lousy #4 starters in order to reach minimum innings requirements (1000 IP).

My league is thinking about going to 12 teams so I don't have any historical information to go on. That is why I thought
that I would use the default settings.

Am I wrong in this assumption that the default settings should be higher for a 12 team AL only league?

ETA: It appears to be a problem with the 12 AL only. When I change it to a 10 AL only league, the default goes up to
4.377 for ERA and 1.406 for WHIP. I would figure that a 10 team would have a lower default than a 12 team.

[aburt19]

One other question on the pitching CVRC. When I enter an ERA and WHIP above the default value that is shown, that
number does not change to the same as what I entered. But on the hitting CVRC, when I enter a BA different than the
default, the number changes to what I typed in. Should the default number change on ERA and WHIP when I enter a
different number?

[Todd Zola]

I was messing around with CVRC for pitching. The spreadsheet comes up as a 15 team mixed league. The default setting
for a 15 team mixed league for ERA is 4.424 and for WHIP is 1.414. When I change it to a 12 team AL only league the
default setting for ERA is 4.292 and for WHIP is 1.384.

Given that a higher percentage of available pitchers are rostered in a 12 team AL only league compared to a 15 team mixed
league, I would think that the ERA and WHIP for a 12 team AL only league would be higher than the 15 team mixed league
due to having to roster some lousy #4 starters in order to reach minimum innings requirements (1000 IP).

My league is thinking about going to 12 teams so I don't have any historical information to go on. That is why I thought
that I would use the default settings.

Am I wrong in this assumption that the default settings should be higher for a 12 team AL only league?

ETA: It appears to be a problem with the 12 AL only. When I change it to a 10 AL only league, the default goes up to
4.377 for ERA and 1.406 for WHIP. I would figure that a 10 team would have a lower default than a 12 team.

I'm hoping this is just a artifact of a still incomplete player pool. The baseline calculation is strictly an empirical formula used to emulate the ratio of the last place team as best I can from the internal data. After I complete the weekend update, I will take a good look at the readings to see if it looks okay.

[Todd Zola]

One other question on the pitching CVRC. When I enter an ERA and WHIP above the default value that is shown, that
number does not change to the same as what I entered. But on the hitting CVRC, when I enter a BA different than the
default, the number changes to what I typed in. Should the default number change on ERA and WHIP when I enter a
different number?

You'll note the pool size changes when you enter in your own baseline, so it is using your number. I just have to fix the code so that it is more aesthetically pleasing and also changes in the other cell.

[aburt19]

One other question on the pitching CVRC. When I enter an ERA and WHIP above the default value that is shown, that
number does not change to the same as what I entered. But on the hitting CVRC, when I enter a BA different than the
default, the number changes to what I typed in. Should the default number change on ERA and WHIP when I enter a
different number?

You'll note the pool size changes when you enter in your own baseline, so it is using your number. I just have to fix the code so that it is more aesthetically pleasing and also changes in the other cell.

DOH, I should have noticed that. Sometimes I ask a question before I think it through.

Thanks.

[shif6]

On the cvrc hitting, if I input American league, the National League players drop out of the player pool, as they should. But if I try to set up a National league the American league players are still in the player pool.

[Todd Zola]

Please double check, the NL should drop out -- but what you do need to do is make all FAA (free agents) be the opposite league. If you are doing AL only, make sure there is an N in the league box and vice versa.

[shif6]

I see. My mistake was to assume that players in the excluded pool would not be in the player pool. They are, but players from the excluded league will not show up on the values sheet, and that is what is important. I see the point about the free agents and I see the importance of changing the dh's to a different position in the F column when running the values.

I notice that the instructions do not show up in the 1/16 CVRC. Is that because they are being revised? One question. Last year after doing the C/OF sort we would create a worksheet and figure out what revisions to make in order to get enough catchers with plus values. We would then put the revisions in the set up sheet in running the values. Are you suggesting this step is not necessary? That is what you appeared to me to be saying in the instructions at the end.

[Todd Zola]

I see. My mistake was to assume that players in the excluded pool would not be in the player pool. They are, but players from the excluded league will not show up on the values sheet, and that is what is important.

If you expected the players to disappear from the player page, that does not happen. But the AL/AL/MIX pulldown sets the pool within the black box of the calculations.

I see the point about the free agents and I see the importance of changing the dh's to a different position in the F column when running the values.

I take responsibility here, we need to do a better job of explaining the FAA change. This is as early as we have ever rolled out projections so there are usually much fewer unsigned players. When I taught and even as a manager when I had direct reports, I used to say it was okay to make a mistake once, the problem comes when you make the same mistake again, so I will make sure the instructions to deal with FAA in single leagues are readily accessible.

I notice that the instructions do not show up in the 1/16 CVRC. Is that because they are being revised?

I have not yet looked at the posted material as I am focusing on getting the written review of playing time completed, but if you mean the primer that was included last week, it may just be that I sent it in a bulk folder last week thus it got included in the zipped file. I will ask JP to post it as a stand-alone download. It has not been revised. That said, I am going to be including a means for MAC users to utilize the CVRC and you may note the addition of a couple of new buttons to give a slightly different method of rankings which can now include the actual dollar value.

One question. Last year after doing the C/OF sort we would create a worksheet and figure out what revisions to make in order to get enough catchers with plus values. We would then put the revisions in the set up sheet in running the values. Are you suggesting this step is not necessary? That is what you appeared to me to be saying in the instructions at the end.

Yes, that is my recommendation, only because the pool is such that the values are really close enough without the extra effort. But please feel free to go ahead and do it the old way. Bit please realize I said "close enough", not "exactly the same."

Yes, the players jump around a bit -- but within a small range that is pretty insignificant when you are putting together the puzzle that is your team.

It is a bit of a philosophical switch for me since my industry oats were sowed on being the "foremost export on the science of player evaluation", but I found that did me minimal good at the draft/auction table.

And finally shif -- I am guessing you were pleased to see the return of old friend Rob

[viper]

Todd/others
Looking at the projections, I am sure you must be using CVRC to do the dollar amounts. Are you using 1's across the board for the categories or are you using some type of variation?

Side question is what do you find to be the best variation(s)?

[Todd Zola]

I have altered the settings for hitting and discuss it in the accompanying primers.

I have always made HR = 1.2 and SB to be 0.8 to account for the distribution within the categories, with SB being top and bottom heavy while HR is a bit more evenly distributed. This also ties into the evidence from my studies that show year after year, the category most champions to the best in is HR.

This year, I have upped that to 1.25 and 0.75 as what has happened is the replacement of SB has risen so even though there are more raw gross SB, the net of useful SB is down as more are being subtracted out of the pool based off higher replacement. So the top SB guys are WORTH even more in reality, but the adjustment is to temper their value just a tad.

Then I did some looking into batting average and I believe since it is such a volatile, if not luck-driven category, it should be devalued a bit, so I took that down to .8 and distributed the .1 to HR and RBI -- rendering the weights:

HR 1.35
RBI 1.1
SB 0.75
BA 0.75
R 1.0

For 4x4, I just take out the 1 for runs.

I originally made RBI 1.05 and runs 1.05, but the difference was negligible and this made it easier to switch from 4x4 to 5x5.

Of course, the user is able to enter whatever weights they desire.

Pitching is still 1's across the board.

And quite frankly, the best variation is the one that gives you the best jumping off point for your league. I have chosen one that combines the theoretically correct and practically useful as best I can.

That is, I did not change SB because it looks better. While it might, it is based in the distribution of SB in roto scoring and how power usually wins championships.

[shif6]

Todd
Thanks for the response. I am glad that you will make the cvrc adaptable for the MAC (I have both PC and MAC, but the MAC is more convenient). And, of course, it is great to see Rob back in the fold. It is the advice given here that allows me not to be always killed by him in the three leagues we are in together. I must admit, however, that in one of those leagues last year he trounced me and my co-owner.
2010, however, is a new year.

[Todd Zola]

For the record, usability will be limited with the MAC and will require the user to have some Excel acumen to get maximum utility.

[aburt19]

Using CVRC with a 12 team AL only league with the custom settings of 24 catchers and 144 OF, I end up with only 20
catchers with positive value. There are four or five catchers that are shown with $0, so it's not a big deal. I just
wondered if there might be something that I am doing wrong.

There aren't many players that don't have positive value and some of the ones that do show a positive value, like Mike
Aubrey, are only shown with 131 AB. I figured this was due to some positions that are still unsettled on some teams,
such as the Yankees left field situation, but I wasn't sure.

Thanks.

[Todd Zola]

Using CVRC with a 12 team AL only league with the custom settings of 24 catchers and 144 OF, I end up with only 20
catchers with positive value. There are four or five catchers that are shown with $0, so it's not a big deal. I just
wondered if there might be something that I am doing wrong.

There aren't many players that don't have positive value and some of the ones that do show a positive value, like Mike
Aubrey, are only shown with 131 AB. I figured this was due to some positions that are still unsettled on some teams,
such as the Yankees left field situation, but I wasn't sure.

Thanks.

The catcher quirk is explained in one of the primers -- it isn't a big deal and is an artifact of the marginal pricing mechanism that sets the pool size. I was going to rework it this off-season, but then we merged and with the plan to fold the CVRC into the Zen Engine, I figured we could get by one more season as is.

As for the AL only/Aubrey, perhaps it is because there are still 15-20 unsigned players that will end up in the AL, but guys like Aubrey that hit for an average that will help you (last place in average is usually about .258-.262 in 12 team AL only), will earn a buck or two and actually make great end gamers for $1 because if they happen into more playing time, their additional counting stats help.

[aburt19]

I'm having trouble figuring out the logic behind the CVRC when making changes. Just a couple of examples.

Using the positions shown in the set up page, changing the BA from the default to the average of my league, and with the
SB at .8, HR to 1.2 and everything else 1, when I change one player from 1B to the OF pool (either Cuddyer or Swisher), it
changes the value of Paul Konerko from $11 to $16. My problem is that if I leave the player at 1B and someone in the
league uses one or both of those players at OF, then my values for the remaining 1B is wrong. I have changed to using
custom and recording it as only two pools C and OF to solve that problem, but that changes dollar value for everyone.

Second scenario: I run a 10 team using only two pools, C and OF, with changing the BA default to the average of my
league. I run values. I change it to a 12 team league and change only the BA default down a little bit and run values.
The value of Miguel Cabrera stays exactly the same at $29, but the value of lower value 1B goes up by a considerable
amount (for example Konerko goes from $12 to $15). I could see that being possible if
one of the players was getting a lot of SB because the "replacement player" for SB has changed. But both players
derive a good portion of their value from HR and RBI and any change in the stats for those in the replacement
player should effect both. The value of catchers goes down considerably with Mauer going from $26 to $23. I figure
that is due to the fact that the replacement player for catcher didn't decrease that much, but the replacement player
for the OF did decrease considerably.

Thanks.

[Todd Zola]

This is EXACTLY why my crusade this off-season has been to seriously discount the notion of a "DOLLAR VALUE" and look at the player in terms of relative worth to other players in his position and to other positions, which can still be gleaned from the information. Basically, I have seen pretty much everything you are talking about over the years which has led me to that realization.

But to address specific points....

Using the positions shown in the set up page, changing the BA from the default to the average of my league, and with the SB at .8, HR to 1.2 and everything else 1, when I change one player from 1B to the OF pool (either Cuddyer or Swisher), it changes the value of Paul Konerko from $11 to $16. My problem is that if I leave the player at 1B and someone in the
league uses one or both of those players at OF, then my values for the remaining 1B is wrong. I have changed to using custom and recording it as only two pools C and OF to solve that problem, but that changes dollar value for everyone.

First, I would recommend using the new category weights I have suggested for 2010. Second, the values change, but by how much? Usually most $1 or $2, some a little more if they have a category in which they excel or struggle. This amount is MOOT.

Second scenario: I run a 10 team using only two pools, C and OF, with changing the BA default to the average of my league. I run values. I change it to a 12 team league and change only the BA default down a little bit and run values. The value of Miguel Cabrera stays exactly the same at $29, but the value of lower value 1B goes up by a considerable
amount (for example Konerko goes from $12 to $15). I could see that being possible if
one of the players was getting a lot of SB because the "replacement player" for SB has changed. But both players derive a good portion of their value from HR and RBI and any change in the stats for those in the replacement player should effect both. The value of catchers goes down considerably with Mauer going from $26 to $23. I figure
that is due to the fact that the replacement player for catcher didn't decrease that much, but the replacement player for the OF did decrease considerably.

Using two pools, all non-C are compared to the same replacement player, so a 2B or OF with the same stats as Konerko would change the same, so it can't be true that all lower 1B change. Players of a certain stat line change. And again, $3 is really not a significant amount.

I would have to have the exact settings to look at the categorical dollars to really explain what is happening in terms of a static number.

[aburt19]

This is EXACTLY why my crusade this off-season has been to seriously discount the notion of a "DOLLAR VALUE" and look at the player in terms of relative worth to other players in his position and to other positions, which can still be gleaned from the information. Basically, I have seen pretty much everything you are talking about over the years which has led me to that realization.

First, I would recommend using the new category weights I have suggested for 2010. Second, the values change, but by how much? Usually most $1 or $2, some a little more if they have a category in which they excel or struggle. This amount is MOOT.

Using two pools, all non-C are compared to the same replacement player, so a 2B or OF with the same stats as Konerko would change the same, so it can't be true that all lower 1B change. Players of a certain stat line change. And again, $3 is really not a significant amount.

I would have to have the exact settings to look at the categorical dollars to really explain what is happening in terms of a static number.

First, I agree to a limited degree with "dollar value" as a static number. I prefer to think to it as a range. A $30 player in my
mind is a $27-$33 player (pre-inflation)

The new category weights that you are talking about: are they the new ones that increase both HR and RBI, and devalue
BA and SB?

As far as the particulars of the spreadsheet, I don't want to make you go to that much trouble. I'll just use the projections
in a spreadsheet that I came up with 18 years ago.

Thanks.

[Todd Zola]

Yes to the weights.

The other thing to keep in mind, and I should have included this in the earlier response, is we are still dealing with an incomplete player pool, especially for single leagues.

If a 1B eligible guy signs, (Branyan, Blalock, Garko, Tracy amongst others are still available), they push the WORST 1B out of the pool, which changes replacement. The fewer the players in the pool, the better the chance of a serious dropoff between players. Adding a player reduces the likely dropoff, reducing replacement and keeping values closer.

But I still favor eliminating positions all together, as the replacement levels for each position are so close now, they can effectively be valued against each other.

One final point -- the manner I do replacement here is a little different than the way I would do it if I were to do it "by hand" as described by the method we have written but I don't think is presently available for download, but I can send a copy if you wish. The "by-hand" method requires multiple sorts, which I cannot automate. The CVRC is a VERY VERY reasonable replacement for the "by-hand", but it is admittedly not as absolutely precise.

But to go back to the original point, I am trying to get away from dollar value "precision" and am more concerned with relative "accuracy" between players.

Keeping in mind I am a scientist by trade, the words precision and accuracy have distinct meanings, even though they are often deemed synonymous by those not in science or engineering.

[aburt19]

Yes to the weights.

The other thing to keep in mind, and I should have included this in the earlier response, is we are still dealing with an incomplete player pool, especially for single leagues.

If a 1B eligible guy signs, (Branyan, Blalock, Garko, Tracy amongst others are still available), they push the WORST 1B out of the pool, which changes replacement. The fewer the players in the pool, the better the chance of a serious dropoff between players. Adding a player reduces the likely dropoff, reducing replacement and keeping values closer.

But I still favor eliminating positions all together, as the replacement levels for each position are so close now, they can effectively be valued against each other.

But to go back to the original point, I am trying to get away from dollar value "precision" and am more concerned with relative "accuracy" between players.

Keeping in mind I am a scientist by trade, the words precision and accuracy have distinct meanings, even though they are often deemed synonymous by those not in science or engineering.

With using only two pools, the addition of a Branyan, Blalock, etc would not be significant because the difference in
the replacement player would be small.

I am by trade an accountant. While precision and accuracy aren't quite the same, they aren't far off. In accountancy we
call it an immaterial item. It is something that is small enough that it wouldn't change the reader's decision on the
financial statements even if it were taken into account. But as much as we'd like to, dollar values are at least the starting
point for the discussion. That's why I prefer to use a range (or if you prefer tiers). But you still get to a point where the
range or tiers make a difference and you have to define that in a way that allows a decision to be made.

[Todd Zola]

With using only two pools, the addition of a Branyan, Blalock, etc would not be significant because the difference in
the replacement player would be small.

Which is why I suggest using two pools. My comment was meant for those using positional replacements. The use of CI, MI and UT also makes the 2-pool means more appropriate, though using 3 pools (2B/SS, 1B/3B/OF, C) is a decent compromise for those that still want to get an idea about scarcity.

I am by trade an accountant. While precision and accuracy aren't quite the same, they aren't far off. In accountancy we
call it an immaterial item. It is something that is small enough that it wouldn't change the reader's decision on the
financial statements even if it were taken into account. But as much as we'd like to, dollar values are at least the starting
point for the discussion. That's why I prefer to use a range (or if you prefer tiers). But you still get to a point where the
range or tiers make a difference and you have to define that in a way that allows a decision to be made.

To a scientist or engineer, they are are indeed far off.

Let's say something measures 20 units.

If you get 10, 20 and 30 and come up with an average of 20, your measurement is accurate but not at all precise.

If you get 26, 25.7 and 26.1, you are not very accurate, but you are quite precise.

Accuracy relates to the raw measurement. Precision relates to the degree of error, usually given in +/- the number of units.

I admittedly used a little poetic license in my application of accuracy and precision, but the larger point is a $23 player and a $26 player are the same player in terms of raw value. The considerations with drafting/acquiring in an auction should be the stats they contribute and can you get those stats later, the position and can you get that position later as well as the upside/downside to the projection itself.

[AllstonRockCity]

While precision and accuracy aren't quite the same, they aren't far off.

As a medical professional, my schooling was of the scientific nature. The example we were taught for mentally picturing the difference between these 2 terms is as follows:

You are throwing 3 darts at a dartboard and you are aiming for the bullseye.

A. 1 dart hits the bullseye, 1 hits the outer ring and 1 hits the wall. you are neither accurate or precise
B. all 3 darts hit the 17. you are precise. you have repeated the same results. you are not accurate, you missed the bullseye.
C. 1 dart hits the bullseye, the other 2 miss. you are accurate. you hit your target. but were unable to repeat your results so you were not precise.
D. all 3 darts hit the bullseye. you are both accurate and precise.

not trying to be nit-picky. but to someone with scientific training these words are not really that similar at all, though in lay-terms they are often used interchangeably. just hoping to share some knowledge.

[aburt19]

The February 15th CVRC has a problem with the PPOS. For example, Bobby Abreu and Torii Hunter have a PPOS of 2B.
I noticed it when I went in to change positions to a 2 pool system.

[Todd Zola]

Yes, thanks, this was brought to my attention.

There will be a clean version early evening.

[freddiebeach]

Did Utley's values get mixed up? His stats on the player tab (Hits, R etc) are the same but when I run the app with my league setting he is worth 11 and before he was 35?

[Todd Zola]

The Utley issue is likely the PPOS issue mentioned above. I just cleaned up the files and will see about getting them posted ASAP.

[freddiebeach]

Thanks Todd, yeah I think in one of the column did have C listed.

[Todd Zola]

Uploaded the new file --