HELP VORP - Fantasy Baseball Cafe 2015 Fantasy Baseball Cafe
100% Deposit Bonus for Cafe Members!

Return to Baseball Leftovers

HELP VORP

Moderator: Baseball Moderators

Postby GotowarMissAgnes » Thu Feb 15, 2007 8:59 pm

Ender wrote:Yes because Nick Johnson who gets hurt every year, Freddy Sanchez who had never played a full season, Estrada who was hurt the yaer before and terrible and Lowell who was coming off a horrible year are as easy to predict as early rounds.

Now your just being obtuse for arguments sake. Of course the later round players are harder to predict accurately than the earlier round. Also you can't just look at where players ranked in hind sight, you have to look at where they ranked before last year and what they ended up doing. I gaurentee you the average person is less sure of their predictions for the 15-25th best players at a position than they are for the top 10.

Using thundermatt's rankings as an example...

http://thundermatt.blogspot.com/2007/02 ... lah-1.html

15th player on each list.

C-Gerald Laird - I can't predict him with any certainty at all.

1B- Todd Helton - This is a bit easier because 1B is so deep

2B - Ray Durham - Not an easy case, does 20+ HR guy show up or injured half the year guy

3B - Edwin Encarnacion - A 2nd year player who was hurt last year and has lost playtime time because of his fielding.. not an easy read.

SS - Stephen Drew - Rookie with no clue as to how much playtime he'll get.

OF - Ryan Freel - ok he's an easy read, but Chris Young is next and he certainly is not.

RP - Brad Lidge - could get between 5 and 45 saves, completely up in the air.

Later round picks are usually young high upside guys with playtime issues, injury concerned vets and guys coming off of bad years. These are the hardest types of players to project accurately.


The question isn't whether people are less sure of their predictions.
The question is whether those players ranked 15-20 or so are less accurately predicted.
Neither your opinion or my opinion about one player matters one bit.

I'm looking at TSN's rankings for last year and I see Varitek as the third ranked catcher and Javy Lopez as the 4th. How'd those projections work out? I see Derrek Lee and Todd Helton in the top 5. How’d that work out? At 2B...Cantu, Kent and Giles, where are you??? At 3B Chavez, Blalock and Ensberg in the top 10 in March, were AWOL in October. Should I go through the rest of them? There’s a lot of misses in the top 10, too.

Projections are made each year.
Players ranks are available each year.
As I've said, I've seen NO evidence that suggests those players around 15-20 are projected less accurately.
I'm always willing to base my opinion on evidence.

Provide the evidence---not random names--showing that guys ranked 15-20 are less accurately projected and I'll change my opinion.
GotowarMissAgnes
Hall of Fame Hero
Hall of Fame Hero

User avatar
Fantasy Expert
Posts: 5516
Joined: 12 Dec 2003
Home Cafe: Baseball
Location: Happy Valley

Postby Ender » Fri Feb 16, 2007 12:26 am

My point












your head



I give up, you go ahead and base your rankings on players who are on the FA list if you want, nobody is stopping you. If you can't agree with the fact that more established players are a bit easier to project I'm not going to convince you of anything.
Ender
Hall of Fame Hero
Hall of Fame Hero

CafeholicFantasy Expert
Posts: 7733
Joined: 30 Apr 2004
Home Cafe: Baseball

Postby Chrisy Moltisanti » Fri Feb 16, 2007 3:17 am

Ender wrote:My point












your head



I give up, you go ahead and base your rankings on players who are on the FA list if you want, nobody is stopping you. If you can't agree with the fact that more established players are a bit easier to project I'm not going to convince you of anything.


Try sorting by reliability on the marcel ranks.

http://www.insidethebook.com/ee/index.p ... asts_2007/
Chrisy Moltisanti
Major League Manager
Major League Manager

User avatar

Posts: 1536
Joined: 10 Jun 2006
Home Cafe: Basketball
Location: 6 feet under

Postby GotowarMissAgnes » Fri Feb 16, 2007 10:19 am

Chrisy Moltisanti wrote:
Try sorting by reliability on the marcel ranks.

http://www.insidethebook.com/ee/index.p ... asts_2007/


Exactly....there are lots of easy ways for you to provide support for your point AND help everyone on FBC learn something. You could also take the rankings and projections from last year and see whether the correlation or mean errors were greater among the top 5 players ranked or the players ranked 15-20. I understand your point perfectly. I'm simply asking you to provide some real concrete evidence for that point.
GotowarMissAgnes
Hall of Fame Hero
Hall of Fame Hero

User avatar
Fantasy Expert
Posts: 5516
Joined: 12 Dec 2003
Home Cafe: Baseball
Location: Happy Valley

Postby garf112 » Fri Feb 16, 2007 7:50 pm

Ender wrote:Its not even good for that. 2 SB from a SS is not worth less than 2 SB by a C. That makes absolutely no sense. You are just hurting yourself by setting up your rankings that way. If there was some C that hit .200 with 1 HR, 40 R, 30 RBi but managed 40 SB by a VORP system he'd be the most valuable C in fantasy baseball because of the tight STD Dev of SB for C's. There is no way that C is worth more than Mauer or McCann. Positional stats are meaningless, if you are going to run a STD DEV or VORP style system you need to base it on total points earned, not individual positional stats.

I just don't get how people justify this system, especially the one that rates positional players by individual statistics. How on earth do you play fantasy baseball and come up with the opinion that a 10 SB C is somehow better than 10 SB from a SS. A system that rates like that is a losing system because if we have exactly the same players but 10 of your SB come from C and 10 of mine come from SS I still have a small amount of extra points for other stats on my SS and I'll edge you out in overall value.


Joe Mauer is highly sought after because of his steals. Why would you even bring him up? The idea is that if you lose Joe Mauer's stolen bases, the catcher you pick up, most likely will give you 0. If you lose a SS, you can probably pick up some steals on the waiver wire.

Anyway, I use both to make my decisions. By position can show you how scarce certain stats are at a position (or how scarce the position is as a whole). By all players, to see what their ACTUAL values are against every other player. If you don't think STDEV by position or VORP (which are really the same concepts) AND STDEV by total player pool are important you are a fool.

You need to use as much data as possible to make your decisions on draft day.
Join Pennant Wars today! A simulated fantasy baseball game where you are the General Manager.

http://www.pennantwars.com/index.php?r= ... &user=2163
garf112
General Manager
General Manager

User avatar
Cafe Ranker
Posts: 2307
(Past Year: 38)
Joined: 10 May 2005
Home Cafe: Baseball

Postby GotowarMissAgnes » Fri Feb 16, 2007 9:19 pm

I agree. Both VORP and STD are important. I prefer VORP, because you can actually use it to do both, while STD cannot be used to look at replacement.

Anyway, I took Chrisy's idea and looked at the first five guys selected at each position in the most recent SportsFanatics mock draft and the players ranked 15-19 in Yahoo at the end of 2006 (most mocks obviously don't provide that number of guys at each position..I used 46-49 for OF, also).

Here's the average Marcel reliability for each set of 5 players at each position (first # is the replacement players, second is the top 5 guys). Higher numbers indicate more reliable forecasts. The top reliability is .88; A guy like Matt Kemp - .4

1B: .84 versus .86
2B: .81 versus .80
3B: .84 versus .86
SS: .80 versus .87
OF: .83 versus .86
C: .77 versus .75
All players: .83 versus .85

So, the top guys are a bit more reliable, but it's really not a huge difference. I don't see that this difference is really big enough to cast any doubt on using a method based on replacement players.
GotowarMissAgnes
Hall of Fame Hero
Hall of Fame Hero

User avatar
Fantasy Expert
Posts: 5516
Joined: 12 Dec 2003
Home Cafe: Baseball
Location: Happy Valley

Postby Ender » Fri Feb 16, 2007 11:02 pm

The idea is that if you lose Joe Mauer's stolen bases, the catcher you pick up, most likely will give you 0. If you lose a SS, you can probably pick up some steals on the waiver wire


The FA list is so unpredictable, there could easily be a Russel Martin that shows up and gives you the 10 SB back, the field of players who are not picked is just too hard to judge.

I guess I simply disagree with you, VORP doesn't do anything for me. Maybe its because I play in deep leagues where the majority of full time players are gone before the draft, dunno. We have 14 teams and 30 player rosters so when I'm looking at a replacement player its player #421 in rankings, that player is almost never a reliable prediction.
Ender
Hall of Fame Hero
Hall of Fame Hero

CafeholicFantasy Expert
Posts: 7733
Joined: 30 Apr 2004
Home Cafe: Baseball

Postby GotowarMissAgnes » Fri Feb 16, 2007 11:43 pm

Ender wrote:
The idea is that if you lose Joe Mauer's stolen bases, the catcher you pick up, most likely will give you 0. If you lose a SS, you can probably pick up some steals on the waiver wire


The FA list is so unpredictable, there could easily be a Russel Martin that shows up and gives you the 10 SB back, the field of players who are not picked is just too hard to judge.

I guess I simply disagree with you, VORP doesn't do anything for me. Maybe its because I play in deep leagues where the majority of full time players are gone before the draft, dunno. We have 14 teams and 30 player rosters so when I'm looking at a replacement player its player #421 in rankings, that player is almost never a reliable prediction.


16 teams and 25 players per team, and I've never had a problem getting a good read on replacement level. The fact is, reliability has almost nothing to do with the issue. All you are trying to do is get a good read on the typical player you'd be able to get on waivers. Player 421 might be difficult to project, so you just increase the sample and look at the average of players 421-430 as the replacement level. There's little difference between the performance of players over that narrow range, so your mean value changes very little as you add to the sample. The reliability of individual players is a red herring, because the law of large numbers trumps it eventually, if you understand what VORP is doing.
GotowarMissAgnes
Hall of Fame Hero
Hall of Fame Hero

User avatar
Fantasy Expert
Posts: 5516
Joined: 12 Dec 2003
Home Cafe: Baseball
Location: Happy Valley

Postby thinkspin » Sat Feb 17, 2007 9:49 am

I wanted to post the other day but I had no time. I see this is still going in the same direction.

Valuing players is what we are talking about. VORP or StdDev.

What I don't understand is how you can take a specific example of a player's stats and say that one or the other method is wrong. It just shows how you do not understand what you are trying to do.

You have a base set of data, wether that be projections or in season stats. You run your analysis to value the players. What you come up with is not only a RANKING, but a way to COMPARE one player to another regardless of position and most importantly the underlying statistics of those players. The VALUE of an individual player is derived from the Population. VORP and StdDev are just methods to get to this point.
Founder of a fantasy sports related website...
thinkspin
College Coach
College Coach


Posts: 167
Joined: 9 Jan 2006
Home Cafe: Baseball

Postby GotowarMissAgnes » Sat Feb 17, 2007 10:40 am

I agree completely, thinkspin. Both methods work fine, imo. I really don't have any problem using STD; I think Ender has some misconceptions about VORP.

VORP was primarily developed so that you could assess a player's performance to the pool just off the draft board. That's useful information, because it tells you what is likely to happen if you lose the player. Appropriately done, VORP also incorporates position scarcity into your evaluation, since replacement levels should be set by position.

I prefer VORP to STD because in addition to doing these things, VORP can be also used to compare the player to the population that is likely to be drafted. This is essentially what STD accomplishes. You simply have to compare each likely to be drafted players to the others or to the average of likely drafted players.
GotowarMissAgnes
Hall of Fame Hero
Hall of Fame Hero

User avatar
Fantasy Expert
Posts: 5516
Joined: 12 Dec 2003
Home Cafe: Baseball
Location: Happy Valley

Previous

Return to Baseball Leftovers

Who is online

Users browsing this forum: No registered users and 9 guests

Forums Articles & Tips Sleepers Rankings Leagues


  • Fantasy Baseball
  • Article Submissions
  • Privacy Statement
  • Site Survey 
  • Contact