Players 2007 projections

[cyranoqc]

I was comparing the "baseball forecaster" projections with those on the site and have some difficulty understanding some important discrepancies. I do understand that those site projections are adjusted considering changes like trades or lineup changes (Juan Pierre being an obvious example even if i don't concur with your drastic changes).

However i don't understand the changes for some other players. For example, "Boof Bonser". I tried to find the reason for the lowered projections and could find nothing explaining it.

Hope i am on the correct forum.

Could you briefly explain what elements generally affects projections at this date and time, other than trading and linup changes. It would help me pursue my analysis for player performance in 2007. Thank you!

[RON@HQ]

From the Baseball Forecaster FAQ page

http://www.baseballhq.com/books/bffaq.shtml

There seem to be large variances between the dollar values that appear in the book and those that appear on Baseball HQ. Why is that?

The values that appear in the Baseball Forecaster are general estimates calculated without regard to MLB league or fantasy league size. At the time the book is published, it is too early to determine what team each player will end up on, how the pool of stats will shake out in each league or how playing time will be allotted. As such, the values in the book are best used only for player-to-player comparison purposes. The dollar values that appear in the Baseball HQ data files are more precisely generated from a 12-team American League or 13-team National league environment, with the best available information we have at the time of publication. For the most accurate estimates of Rotisserie value, you should use HQ's Custom Draft Guide engine that calculates dollars based on your own league parameters.

[cyranoqc]

Thanks for the "cut and paste" answer, but i had already read that and it does not answer my question. I guess an acceptable answer would require me to identify the specific players whose stat changes i don't understand. And i know you don't have the time to answer that kind of question.

So i will go on with my analysis and hope to find what provoked the changes. Thanks anyway.

[RON@HQ]

Not sure how the cut-and-paste doesn't answer the question. Look at Boof Bonser:

Forecaster: 8-7, 4.27 in 116 IP
HQ: 6-5, 4.24 in 87 IP

The only thing that has changed is the IP, which is explained in my note above. For further detail why the IP has changed, you can always go to the team depth charts (in this case here: http://www.baseballhq.com/members/to...ink.php?tmid=8 and see how his PT fits into the MIN rotation. The Forecaster can not go into that level of detail.

And as the MIN rotation roles further shake out, his IP will likely continue to change.

[helloneumn]

To add to what Ron has said, there will likely not be a specific "blurb" which explains the change in the projection but the vast majority of the changes will be based on PT. Of course, given that its only January, there are many more changes to occur. That said, you're free to read the tea leaves differently and disagree. There are often threads around this time of year questioning the HQ wisdom regarding PT. So, if you disagree with the Boof projection, feel free to post a thread and see if others agree/disagree.

[Mr X]

I'm not sure if this is the correct Forum, but along these lines I see that the Forcaster has different R$ values for 06 vs the end of the year values I pulled off the system.

Example:

Arroyo has a 06 R$ in the Forcaster at $26 and the end of the year values that I pulled of the system has him at $21. I use the R$ for 06 as well as the projected R$ for 07 on draft day so I'm wondering if you can you explain why this is?

Thanks and I love the data don't get me wrong...........

[RAY@HQ]

I think Ron's first post above addresses that point above, Mr. X.

[TyCobb]

Follow-up Question

I have a related question. The dollar values of the top players in particular are just way too low. I checked the 5x5 AL values and only one player is over $30 (Carl Crawford). No way will only one player go for over $30 in any reasonable league, even without inflation. Are you saying Johan Santana is not worth over $30? If I used these projections as my guidelines for drafting, I would have no stars on my team and would be left with huge amount of money to squander at the end of the draft.

I love all the analysis, forecasts and insight (clearly the best around), but I also buy Alex Patton's software because his $ valuations seem to be more on target. What is the explanation for why the dollar values seem out of whack?

P.S. I apologize if this has already been covered elsewhere.

[RAY@HQ]

Ty, I think you need to draw a distinction between what the projected statistics are worth in terms of dollars (which is what our R$ indicates), and what the marketplace will pay for those stats (which we have no control over).

Also, if you want to adjust the dollars to your league parameters, use our Custom Valuator tool. (See 'Custom Draft Guides' on the left side menu.) For instance, if you run a 12-team AL 5x5 thru that to match what's in the projections files, but select the "Top Players" option (which moves dollars to the top players in a "Stars and Scrubs" kind of approach), Crawford jumps to $39 and Santana to 35, and something like 6-8 guys climb to $30 or above. If you think that matches your league better, use it.

[Howie M]

Ray, I still think the "what the marketplace pays" is only part of it.

This is a reach because I don't have evidence for it to present, but hypothetically, if every year there were 5-10 players in each league who earned more than 30, and HQ for 2007 projects only 1 or 2, the result is that slavishly following HQ prices means you'd never get any stars - and perhaps, not be able to spend all your money. Does that make sense?

I never noticed this issue in previous years, and admit I'm just offering an anecdotal sense this time.
I guess I'll just try the "stars and scrubs prices," although frankly my league doesn't veer too strongly that way.

[TyCobb]

Thanks

Thanks for the heads-up on selecting the "Top Players" option. I've used the Custom Draft Guide in the past and found this didn't really solve the problem, but I don't believe I've ever tried this option before.

I realize that projecting dollar value is difficult, particularly given that leagues work differently. But for standard leagues, I have noticed that Alex Patton's valuation methods do tend to line up closer to market value (not just for my league but for other leagues I know about). I know you use some variation of Art McGee's method, of which I admit I'm not vary familiar). I'm more familiar with Alex's method, and it makes sense to me. I'm guessing Ron is familiar with Alex's method, but it may be something worth investigating.

Here's an example. Alex's forecast and HQ's forecast for Johan Santana are nearly identical. Alex's formula projects those stats to be worth $39-$40. HQ says $29.1 I would guess most of your readers who play AL only and close to 12 teams (no inflation) would say it should be closer to Alex's value. Sure, we can say, "Well, I know HQ's dollars are too low because of XYZ," but why should we have to? I'm not trying to sound critical or get too nit picky because I LOVE your work and this isn't a major problem, and I'm glad there is a work around with the Custom Draft Guides, but the "face value" R$ on the Web site seem to be one of the least useful numbers offered except, of course, to compare players, which I realize is the main point.

P.S. This might be related to what cyranoqc is trying to get at, because it just so happens that the R$ for Santana in the book is $39. Yes, the stats projections for Santana in the book are a little different, but they don't seem to be $10 different.

[timg]

You would probably be interested in this thread which discusses some evaluation issues and talks about the potential trouble with passing on all the over-priced studs and waiting for the bargains to follow.

You may also find this thread regarding the AL Mock Auction of interest.

Thanks for the heads-up on selecting the "Top Players" option. I've used the Custom Draft Guide in the past and found this didn't really solve the problem, but I don't believe I've ever tried this option before.

I realize that projecting dollar value is difficult, particularly given that leagues work differently. But for standard leagues, I have noticed that Alex Patton's valuation methods do tend to line up closer to market value (not just for my league but for other leagues I know about).

Price does not set value. The fact that Alex Patton's valuation more accurately predict price, does not mean that they more accurately predict value.

I know you use some variation of Art McGee's method, of which I admit I'm not vary familiar). I'm more familiar with Alex's method, and it makes sense to me. I'm guessing Ron is familiar with Alex's method, but it may be something worth investigating.

The method in Art McGee's book is referred to as Standings Gain Points (SGP) and was first written about by Alex Patton. In the beginning of the SGP chapter in his book, McGee says that "[m]any of the concepts I cover in this and the next chapter were originally presented in [Alex Patton's] book." Unless Patton has changed his methods recently, I believe he is still using a process similar to SGP.

BBHQ uses Percentage Valuation Method (PVM). There are lots of similarities between the methods.

[Doc]

I think Ron's first post above addresses that point above, Mr. X.

With all due respect Ray, I don't think it does answer the question. I asked the same and really wasn't given an answer. The question does NOT have to do with proj 07 $R which I know is not generated the same way or is as good a prediction as the proj 07 $R values that show up on the site. the past excel files on the site, use a method involving a "standard" 12-team AL or 13-team NL , to come up with a $R, correct? How do you calcuate the same past $R values for the forcaster? Is it the differnce between SGP or PVM methods? Or something else. I know the past values in the forcaster stay the same, by looking at past editions. This question does not relate to the Forcasters projections at all, just the historical data. Thanks Doc