Pricing updates

I decided I would start a thread to see what good feedback the community might have on updating pricing. Now for this to work comments need to follow this type of format:

1) Don't provide feedback based on a usage model (that is something to take up with higher powers). The pricing model is a static linear RC model. Usage might be a good indicator of players that are values, but I am interested in understanding why players are used more than others.

2) Please try to identify groups of players that you feel undervalued and to the extent possible the variable that you think is incorrectly priced. (i.e. I think all hitters with good defense, high walks, and high slugging are over priced).

3) It is also helpful if you can group players that you think should be of similar value. This combined with #2 above allows me to play with the value of inputs to tweak pricing on the overpriced players while keeping the correctly priced cards static.

4) Help me identify generalities in the game that pricing needs to do a better job of handling.

5) Feel free to ask about specific variables and the values that I use to model. There are assumptions inherent in pricing, refining these assumptions is what I continually try to do with each iteration of pricing.

Mark

I am going to start off with some generalities that I am currently tackling:

Based on the success of team in various ballparks using the barnstormers as a data source I have determined the following (this might not be an eye opener to some):

If I group ballparks into 5 classes (RH favorable, LH favorable, high HR environment, low HR environment, and neutral there are clearly ballparks that are easier to win in:

1) low HR environments (HR chances no greater than 4) are the easiest parks to win in. On average home teams in this environment outscore their opponents by 23 runs over a season.

2) LH favorable parks are also a good place to play. the average team outscores their opponent by 17 runs.

3) RH favorable parks are break even

4) High HR parks are slightly unfavorable as teams are outscored by 9 runs on the season.

5) The big loser is neutral parks. Teams playing in neutral parks were outscored by 57 runs in a season.

Fundamentally the way the game is currently priced this outcome actually makes sense. The entire set is based on an average park environment. Putting the right players in the right ballpark is very the biggest advantages come. So in a neutral park everyone is priced relative correct so no advantage.

In another thread Adrian had posted an idea about averaging pricing based on multiple park environments. This is how the 200X games are priced to help create bargains in all park environments and reduce the benefits of putting players in their primary park.

However, this methodology didn't not work for ATG. I tried it back in either III or IV and the size of the card set screws things up.

So for the next version of pricing I am considering adding a premium to cards that have extreme BP HR results (i.e. zero or one and seven or eight). By adding on this artificial modifier, these players will be overpriced in the average park environment, which hopefully encourages usage of other players in those parks. These extreme BP cards would still be good values however when moved to the right park environment.

Looking for thoughts and comments,

Mark

Also one of my cardinal rules to pricing is that I adjust over all variables and assumptions, never individual players. That is why comments about one player are usually not helpful, because the variable I adjust to reduce one players price trickles through all of the other prices in the set.

Mark,

I do not think it is possible for you (nor anyone else) to do an excellent job on these card prices, because different managers all have different opinions based greatly on the parks and players they already favor.

I also have some difficulty offering specific suggestions without knowing the "black box" calculations already in play.

But I'll throw a fish into the frying pan with my own observations.

I believe that many managers use Adrien's Diamond Dope site to review certain criteria for players, and I know several also pay to get Joe the Jet's ratings. These ratings and metrics might be useful in price adjustments.

Here is a case study -
Looking at NERP for pitchers as presented in Diamond Dope (which is different than the more recent NERP formulas presented by Dean Carrano) there are a few highly used pitchers that have skewed (lower) NERP values than their price cohort. (and this is regardless of park used)

lets look at 5 parks and a couple of key players -

a)Polo 41 0,0,20,20 B)Griffith 41 7,7,0,0 C) Vet '75 10,10,10,10 D) Dunn '20 16,6,10,1 E) Minute Maid '05 3,11,3,16

Some cheap and imo underpriced and therefore highly used pitchers -
Kirby Higbe 3.34 - NERP scores A) 4.93/4.73 B) 5.93/5.78 C) 6.35/6.21 D) 7.20/5.64 E) 5.36/6.35

compare him to similar priced cohort pitchers, and you really need to get into the 4 or 5 million price range to find similar NERP scores for the same parks

Comparisons- (I chose RH SP*/RP with the same or more SP* innings and an E balance)
Firpo Marbery 5.48 NERP scores A) 6.88/6.73 B) 4.83/4.69 C) 6.78/6.64 D) 7.63/4.70 E) 4.72/7.70
* commentary - Firpo is better where HR numbers are low, but overall is the $2.15 mil diff. appropriate?
Jim Bagby, Sr. 4.72 NERP scores A) 6.25/5.85 B) 7.24/6.84 C) 7.67/7.27 D) 8.52/6.70 E) 6.68/7.41
* comment - Bagby, Sr. is worse in every park and his slightly better hitting and WP rating hardly is worth $1.38 M

There are many more similar RH SP*/RP more pricey but less value based on this metric - check some out for yourself

Another popular choice -
Claude Hendrix 2.42 - NEPR scores A) 5.59/5.18 B) 6.58/6.18 C) 7.01/6.60 D) 7.86/6.03 E) 6.02/6.74

let's compare to some other non* starter/RP with an E balance

Grover Lowdermilk 3.03 - NERP scores A) 5.26/5.09 B) 6.26/6.08 C) 6.68/6.51 D) 7.53/5.94 E) 5.69/6.65
*these numbers are slightly better, esp. vs LH bats, BUT he is only S6 vs S8, fields and hits worse so why $0.61 more
Wild Bill Donovan 3.22 - NERP scores A) 5.59/5.42 B) 6.58/6.41 C) 7.01/6.84 D) 7.86/6.27 E) 6.01/6.98
* these look pretty darn close vs LH and worse vs RH. both S8/R5,(Bill fields better and hits worse) but why $0.80 more

These perhaps are less egregious variations than Kirby, but the pricing based on this metric doesn't compare properly IMO


Is this the kind of stuff you are looking for Mark?

Has Bbrool been killed off?

2) Please try to identify groups of players that you feel undervalued and to the extent possible the variable that you think is incorrectly priced. (i.e. I think all hitters with good defense, high walks, and high slugging are over priced).

I agree with this. People are always griping about A-Rod, Joe Morgan, Darrell Porter, Roberto Alomar, Jim Edmonds, David Wright..............formerly Killebrew until he was repriced, and the cheaper Ted Williams and Tris Speaker until their newer more expensive cards were added. These guys have good defense and power (at least power relative to their position), and a lot of their offensive value comes from their OBP. I DO believe that they need repricing, but not too drastically. Sometimes the places A-Rod or Morgan help your team cant be as easily quantified by a cursory look at your stats page...................

So for the next version of pricing I am considering adding a premium to cards that have extreme BP HR results

I offer a simple solution. Charge for ball parks. Minute Maid, for example, would be $100,000, Petco costs $80,000, Cleveland Stadium '66, $0.00. (The pricing is off the top of my bald head.)

Maxie

I am extremely convinced that player pricing needs to be based on the average value in multiple ballparks. Your selection of Minute Maid, Dunn, Griffith, Polo, and the Vet seem reasonable. It might be better to used Forbes 57 as the singles park rather than Polo just to bump the price up on pitchers with no ballpark singles.

But there's no reason you have to use real ballparks, is there?

Why not just price players on the following 5 sets of part ratings:

All ratings 10 (the Vet)
All LH ratings 20, all RH ratings 0
LH 0 RH 20
all singles 20 all HR 0
singles 0 HR 20 (Polo)

Player pricing would be the average of their 3 best ballparks.

Mark,
I would strongly urge that you exercise great caution in interpreting the park data you presented from the Barnstormers Tour. There are several potential problems with this data, not the least of which is small sample size. A serious confounding variable is that your data ignore cap size, which is a huge determinant of park efficacy. It would be prudent, before drawing any inferences, to also examine the ball park data available to the community from Diamond Dope park actuals. The Diamond Dope park data is comprised of all the ATG 6 leagues that were not live draft leagues on the TSN live draft site, up until Gabriel decided to stop collecting this data, perhaps a year and a half or two years ago.
Here are the most frequently used parks from Diamond Dope Actuals across all caps.
Park Uses Win Pct.
1. Minute Maid 437 .526
2. Hilltop 307 .515
3. Forbes ’57 294 .522
4. Fulton County 246 .504
5. Polo ’41 232 .511
6. Petco 201 .500
7. Dunn 194 .525
8. Fenway ’67 166 .499
12. Anaheim Stadium 119 .476
I never play in neutral parks except when required to by the Barnstormers Tour, so I certainly agree with that finding of Mark’s.
However, cap x ballpark interaction effects are very powerful and should be taken into account if one is planning on adjusting pricing based on ball park type.
The most commonly used parks are all extreme parks, for obvious reasons, they tend to confer an advantage except for parks high in both singles and homers (less unique). A maxim I live by, is that all other things being equal, the higher the cap the better bomber teams do and the worse small ball fares. The converse is equally true.
Let’s take a look at win% as a function of cap in several of the above parks. I suspect the general effect is robust across all parks especially for the pure small ball and pure bomber teams. My gut tells me that the lefty righty stuff is more complex…
Minute Maid
60M .507
80M .513
100M .523
140M .536
200M .532
Forbes ‘57
60M .530
80M .531
100M .525
140M .504
200M .495
Fulton County
60M .502 (only 3 uses though)
80M .495 (50 uses)
100M .494
140 .514
200 .513
Dunn
60M .528
80M .522
100M .522
140M .533
200M .531

I've only cited a few examples from the most used parks. Readers are encouraged to explore the data further.