Value calculations question

[Hammer]

Ive purchased rotochamps version of draft dominator and with my settings as a roto 5x5 and came across these projections


player IP W S ERA WHIP K
Chapman 55 2 14 1.31 1.07 80 $14.28
Kuo 58 2 10 2.33 0.88 69 $13.05
K-Rod 57 4 31 3.16 1.14 59 $12.70
Gregerson 74 3 2 2.80 0.95 87 $11.79


When I inquired about what I preceive to be inaccurate values since K-Rods save total far exceeds the others, rotochamp's CEO said, "Kuo, Chapman, and Gregorson all have outstanding ERA and WHIPs that overcome their lack of saves. People tend to focus on saves over ERA and WHIP when it comes to relievers, but if a guy saves 35 games with a 1.35 WHIP and 3.85 ERA he will be less valuable than somebody like Gregorson who will have 1 or 2 saves but a low ERA and WHIP."


I cant seem to wrap my mind around it. Any input here would be appreciated.

[Erik]

ERA and WHIP are generally very volatile categories for relievers. Saves are not. That seems to be absent from his algorithm.

[The Dane]

I cant seem to wrap my mind around it. Any input here would be appreciated.

Crack is bad.

My guess is that we have a stat geek who doesn't play fantasy. He's right that we do over-value saves, but that's the way the game is built. Same with SBs. Perhaps in real baseball, a guy like Kuo is more valuable than a guy like KRod, even without the saves, but in fantasy this is not the case.

[Hammer]

the funny thing is, his algo seems to over-value speed as Ichiro is the 5th OF in overall value and Elsbury and Pierre are valued higher than guys like Hart and J Upton.

[schlesinj]

the funny thing is, his algo seems to over-value speed as Ichiro is the 5th OF in overall value and Elsbury and Pierre are valued higher than guys like Hart and J Upton.

It seems like BA is very valuable just like the pitching ratios.

[Todd Zola]

I took their projections and plugged them into my calculator, set it for NL only, 13 teams, 10 pitchers, 69/31 split and it told me the following, with my values in ():

KRod $15 ($16)
Kuo $14 ($10)
Gregerson $13 ($9)
Chapman $9 ($4)

My take is that while there is something amiss with his algorithm, likely in the manner it uses IP (AB) as a multiplier when you convert ratios to counting stats, they also do not regress BABIP enough in their projections, which speaks towards the volatility comment above.

[Hammer]

thanks guys. I guess I will have a lot of manual work between now and then.

[rhd]

Ive purchased rotochamps version of draft dominator and with my settings as a roto 5x5 and came across these projections


player IP W S ERA WHIP K
Chapman 55 2 14 1.31 1.07 80 $14.28
Kuo 58 2 10 2.33 0.88 69 $13.05
K-Rod 57 4 31 3.16 1.14 59 $12.70
Gregerson 74 3 2 2.80 0.95 87 $11.79


When I inquired about what I preceive to be inaccurate values since K-Rods save total far exceeds the others, rotochamp's CEO said, "Kuo, Chapman, and Gregorson all have outstanding ERA and WHIPs that overcome their lack of saves. People tend to focus on saves over ERA and WHIP when it comes to relievers, but if a guy saves 35 games with a 1.35 WHIP and 3.85 ERA he will be less valuable than somebody like Gregorson who will have 1 or 2 saves but a low ERA and WHIP."


I cant seem to wrap my mind around it. Any input here would be appreciated.

Ignoring for the moment the questionable projections, here's what I get w their projected stats using my formulas for a 5x5 league:

Chapman - $ 9.18
Kuo - $ 6.58
K-Rod - $13.91
Gregerson - $ 6.25

Many people do tend to undervalue superior set-up guys, especially those that get a lot of Ks, but how can a guy w 2 SV be worth the same as a guy w 31 SV when the main advantage the first guy has is 28 more Ks? A Save is worth a lot more than a K; you dont need to know much about fantasy valuation to see that. I have K-Rod's SV value being almost $14 above Gregerson's. The difference in ERA and WHIP value doesnt nearly make up the difference.

My problem w the projections have mainly to do w Chapman. He may be a great talent, but what reason is there for anyone to believe he'll have a 1.31 ERA when very few relievers have ever done that over an entire season? His ERA in the minors last year was 3.57. And his WHIP isnt likely to be as low as 1.07 when he still has such control issues. OTOH, his K projection seems very low; he'll have over 100 easy if he's healthy all year. Also, they project K-Rod to be somewhat in decline while they project Gregerson's ERA to be lower than it's ever been. It seems like K-Rod has been around forever but this is only his age 29 season. He's only 2 years older than Gregerson; they're both still firmly in their primes. Then again, I dont know what the extent of K-Rod's off-the-field issues are.

[Lucky]

I can understand this a little, especially in an single-league format where there aren't as many pitchers available. In those formats, I have always tried to find a high-K, low Whip, low ERA middle reliever over a fifth starter. The reason is that a super MR can get you six or seven wins, a very low ERA, a microscopic Whip and 70-90 strikeouts. What these numbers do, essentially, is to take one of your "decent" starters and turn him into a Cy Young candidate. You get serious help in 4 categories (as compared to a mediocre starter who might not be much better in the Wins and Ks departments). The overall effect on your team might exceed a closer who gives you help only in saves and actually hurts you in ERA and Whip.

A good example of this was Arthur Rhodes in 2001: 8 wins, 1.72 ERA, 0.85 Whip, 83 Ks in 68 innings, and 3 saves. What do you figure that was worth? He had another year almost better, with 10 wins and 102 Ks in 95 innings, never starting a game.

So, I'm not saying I agree with him 100%, but I can get where he's coming from. If I could have Rhodes' 2001 numbers from a MR this year, I'd take them over several closers out there.

[Hammer]

The rotochamp product is not quite up to par when compared to rotolab. In addition to the poor valuations you can't adjust categorical valuations. It also is unable to calculate total profit value of your keepers on the team page.

[Erik]

I can understand this a little, especially in an single-league format where there aren't as many pitchers available. In those formats, I have always tried to find a high-K, low Whip, low ERA middle reliever over a fifth starter. The reason is that a super MR can get you six or seven wins, a very low ERA, a microscopic Whip and 70-90 strikeouts. What these numbers do, essentially, is to take one of your "decent" starters and turn him into a Cy Young candidate. You get serious help in 4 categories (as compared to a mediocre starter who might not be much better in the Wins and Ks departments). The overall effect on your team might exceed a closer who gives you help only in saves and actually hurts you in ERA and Whip.

A good example of this was Arthur Rhodes in 2001: 8 wins, 1.72 ERA, 0.85 Whip, 83 Ks in 68 innings, and 3 saves. What do you figure that was worth? He had another year almost better, with 10 wins and 102 Ks in 95 innings, never starting a game.

So, I'm not saying I agree with him 100%, but I can get where he's coming from. If I could have Rhodes' 2001 numbers from a MR this year, I'd take them over several closers out there.

Right. But the problem is that non-closing relievers don't repeat those kind of numbers from year to year. If they did, they would be starters or closers. So, yes, it's not a bad strategy to roster a guy with such numbers, but it's quite difficult to figure out which guys are going to put up those numbers. (Remember when Chris Hammond came out of nowhere to post an ERA under 1?) For that reason, paying near-closer dollars for middle relievers based only on ERA and WHIP projections is not a reliable strategy. (If their team has a shaky closer and you're convinced a certain MR is going to get a bunch of saves because of that, that's different.) Depending on how tight your league is, you may even be able to find someone who fits the criteria on the waiver wire, if you're alert early on.

[Don Quixote]

Given the choice between an iffy closer with poorer peripherals, a real-life 4th/5th SP with poor WHIP and mediocre or worse K rate, and an effective MR who is first or second behind an established closer, I'll take the latter. When I've stuck to this idea, I've avoided major ERA/WHIP disasters and sometimes managed some surprisingly good Ks.

Hammer, I agree that Rotochamp (which I'm using in Mock One) seems to overvalue MRs. Once I figured that out, I did some mental adjustments to its values as compensation. I like it, but I may still buy Rotolab for my auctions.

[Hammer]

Given the choice between an iffy closer with poorer peripherals, a real-life 4th/5th SP with poor WHIP and mediocre or worse K rate, and an effective MR who is first or second behind an established closer, I'll take the latter. When I've stuck to this idea, I've avoided major ERA/WHIP disasters and sometimes managed some surprisingly good Ks.

Hammer, I agree that Rotochamp (which I'm using in Mock One) seems to overvalue MRs. Once I figured that out, I did some mental adjustments to its values as compensation. I like it, but I may still buy Rotolab for my auctions.

It also overvalues speed and it takes positional scarcity into account. Look at Napoli and Morneau's values on rotochamp.Morneau's are way better, yet Napoli has a bigger money valuation because he is also a catcher. Thats wrong.
I just bought rotolab because I dont have the head to go through every players projections and even if I change them, the valuations are off and I cant change the weights.

[Todd Zola]

It also overvalues speed and it takes positional scarcity into account. Look at Napoli and Morneau's values on rotochamp.Morneau's are way better, yet Napoli has a bigger money valuation because he is also a catcher. Thats wrong.
I just bought rotolab because I dont have the head to go through every players projections and even if I change them, the valuations are off and I cant change the weights.

No, actually, at least in principle, it is not. I don't know the actual projections in question, I am speaking in generalities.

There are two ways to look at it:

1) The same stat line from a catcher and a non-catcher is worth more coming from the catcher

2) The stat line of a $1 catcher is worse than the stat line of a $1 non catcher (actually, the value itself does not matter could be $10 v $10, $17 v $17, whatever).

The reason is you do not want to assign value to something everyone has. The worst player at each position is worth $0, but by convention, we need to pay $1.

I call the stats over and above what everyone else has "useful stats." So you only want to assign value to useful stats. This makes the value equal to $1 plus the value from useful stats.

A catcher and an outfielder both hit 20 homers. But the 20 from a catcher are worth more than the outfielder's since the catcher hit more useful homers. The last drafted catcher may hit 2 homers and the last drafted outfielder maybe 4. These are just examples and are obviously format dependent.

[Hammer]

in this case, morneau has 24 more runs 25 more rbi 5 more hr and 35 points of BA over Napoli. Napoli hass 5 more steals. Napoli is valued 2 dollars higher.