Pricing updates

The pricing stuff is indeed complex. I would argue though that the effect higher caps have on small ball is also. Yes, there are many more bomber suitable players than small ball superstars at the upper end of the pricing spectrum. However, having played in the majority of capless TSN live drafts (hey it has only been 10 months since that critical functionality has been available, but don't worry SOM will raise prices anyway) there are generally a large preponderance of bomber teams, so the competition for the admittedly scarcer small ball guys is much less Salty would be the greatest expert here, he played almost exclusively small ball during the live drafts, but my experience has been that there are plenty of superstars available for small ball teams at these drafts, particularly so after I prevailed upon Bernie (after lots of lobbying) to unleash the 1894 heroes.

It has been my consistent observation that small ball teams, stocked with an appropriate amount of 10M+ players, still don't cut it at live drafts. I am confident this speaks to their inherent inferiority to bomber teams when caps are not present.

Also most caps truncate the player pool. 60 and 80M eliminate much of the higher priced players as 140 and 200M eliminate most guys under 7M except for scrubs.

Oh, before I forget. I like motherscratcher's idea of a built in random distortion of +/- 2% if pricing ever gets updated. I'd even go as high as 5%. That might even make Nev happy. It would certainly make hunting for values enormously enjoyable to most of us on real updated iterations of ATG.

Some of the new ATG 7 card prices are just nuts. As usual, the worst are the pitcher cards. The most extreme example appears to be the $5.19 million bob friend card, which I doubt I would use in any park if it was priced at $3 million. He's just awful.

I was looking at that Friend card and agree, unless the error and hold ratings are worth three million there is something wrong. The Dizzy Trout 2.39 is better. Friend allows a ton of extra base hits while Trout is mostly singles and walks at less than half the price. In a game dominated by hitters because of the ballpark effects, I think pitching is over priced.

In a game dominated by hitters because of the ballpark effects, I think pitching is over priced.

If pitcher prices were lower across the board, in a "game dominated by hitters" I think such a move would have the unintended effect of teams loading up even more on offense by spending less on pitchers. Wouldn't it be better raising the salaries of power hitters to improve the balance between pitching and hitting?

In a game dominated by hitters because of the ballpark effects, I think pitching is over priced.

If pitcher prices were lower across the board, in a "game dominated by hitters" I think such a move would have the unintended effect of teams loading up even more on offense by spending less on pitchers. Wouldn't it be better raising the salaries of power hitters to improve the balance between pitching and hitting?

Probably both need to be done. Or max the ballpark numbers to six instead of eight. I wouldn't want the job of pricing, that is for sure.

Ok,

a mix of comments, some of them while maybe valid points bring up topics and issues outside of pricing. Here are some responses:

Hakra- I actually addressed similar pitcher concerns in another thread. Higbe's price is in error, he must have slipped through the review cracks. For non asteric starters one of the goals in pricing ATG6 was to make them more usable. I think we went a bit overboard and it can be seen the most in the $3 mil and under range. This is a mechanism I would look to change in the next go around.

Rburgh- Pricing for multiple parks. As I have mentioned in other threads we use a mechanism like this already for the 200X game and prices are based off a hitters max value in his best park. This method worked much better however in the limited player pool where scarcity plays a factor. I am going to revisit this type of mechanism for the next update.

The last Druid- Yeah I agree I have to be careful using such a small sample size and probably even more so when that small sample produces results I was expecting. The sample I pulled from the barnstormers was actually only $100 mil leagues and in that environment small ball teams where the most successful. One of the concerns is that at higher caps small ball becomes less usable.

On the topics of salary caps- using static pricing regardless of the salary cap clearly won't meet all of the cap needs. The game pricing has always been centered around $80M leagues. Some of you would argue that a $15M player is priced out of an $80 Mil league, but the output is there. However, since $100 Mil seems to be the cap of preference in ATG, I am trying to look at usage from that perspective to adjust pricing. Caps above that level where created for special leagues the salaries can not adequately accommodate free agent drafts at this level.

Nev- I disagree with the need to put in artificial salary adjustments just to create bargains. It is clear that even using a scientific method to price, their is still plenty of value to find based on parks and other unique variables to each league.

Salty- you asked for me to explain defense, so I will and address another comments about overpriced defense at SS.

I have a chart that computes the incremental team runs that are generated or reduced when you put any rated defender in the field. So SS is based off of 200 X chances in a given year. Each X chance can produce an out, a DP, or some type of OBP/TB event. These events over 200 PA's are converted in runs for each rating.

As an example Phil Rizzuto compared to Luke Appling (1e12 vs. 2e32). Over the course of a season, Luke Appling will generate 12 few double plays, while allowing 39 more hits (error acts the same as a hit on the X chart) and 45 more total bases. This converts to 34 more runs allowed over the season.

I have converted every error rating into a run value for each position.

I used these two players because another poster suggested that defense was overpriced. That may be the case and I need to review that, however Diamond Dopes NERP is looking at offensive value only so you would need to modify those amounts for defense, possibly significantly. I agree with your comments about usage, and while Appling is used Rizzuto is rarely used, so it is something to reconsider. I have levers that I can use to dampen the impacts of defense.

The value of speed is currently priced as a function of speed rating and OBP (i.e. speed is no good if your not on base, and really slow players have been penalized for high OBP and no speed). While all of the variables on the card have been included in pricing, some are not translatable to a sabermetric runs created model and therefore subject to a lot more debate over their true value. This is one that does need to be revisited, but in truth 1-17 running is probably only adding $200-$300k to the highest OBP guys.

I agree that you can not perfectly price the set due to all of the variables. All I am trying to do is come up with a pricing mechanism that makes all parks viable and all players worth using. In that regard I think improvements can be made.

There were some comments specifically about some of the new pitchers released so I will take a look at those mechanics. I think there are some aspects of my pitching model that need to be revisited. Another project I would like to tackle is making defensive ratings variable with parks. In theory a high singles BP environment produces more PA's over a season as well as more DP opportunities. It makes sense to assume that in this environment good defense is more valuable, but at the moment my model is static on defensive value. It is not park specific like other aspects of the model are.

Mark

Friend is priced appropriately compared to other S8* starters in the same price range and he is priced about $1 mil per 12 run increments above S8* at the $4 and $3 mil range. However, I will re evaluate my pricing slope for Pitchers, especially in this mid priced range.

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

Why place a premium on cards that have extreme BPHR's on their cards? You may have an added edge on a specific home park but pay a penalty on the road if a league has a preponderance of pitchers' parks. It would be up to the saavy manager to find the correct mix for both home and away and bypass the cards too heavily dependent on BPHR's. It would seem to me that in this type of environment, the value of such cards decreases

Conversely, in a small ball environment, the value of Cobb, Speaker and the like would increase. Maybe the pricing structure should take into account the type of league one is playing in and the cap? Maybe the prices on the drafted players could be adjusted post-draft based on the configuration of the league? A script could be run post-draft using the avg BPHR's, BP Singles of the league, divided by two (home/away) and use that as a multiplier vs. a players "avg" price tag - just a programmers dream I bet!

I don't think the 200x model would be appropriate in pricing the ATG cards. The ATG cards are a collection of the greatest seasons in MLB history whereas the 200X cards are based on a specific season. It's two different games for all practical purposes. I suspect both sets use avg. league totals as a baseline for determining what a specific player is worth depending on how many standard deviations above the norm that player is. I think a good place to start would be using the completed data available in Diamond Dope, establish the avg baseline and establish the prices based on how far above or below the avg those cards are.

Friend is priced appropriately compared to other S8* starters in the same price range

Horsepucky!

I downloaded the raw card data for all 128 S8* RHP in the card set. I then sorted them by price, and for the middle 108 cards by price I computed the average of on-base chances, number of XBH, card total bases from hits, ballpark singles, and ballpark homers for each side of the player's card and the ten adjacent cards by price both above and below.

I then took each of those ten variables from the individual card, and subtracted the corresponding average. I then summed the "error" for each of these 108 cards. Now, if you have a "better" card than the local average, this number should come out negative. If you have a "worse" card than average then this number will be positive.

The 5 "best" cards by this methoology are:
Bob Turley '58 5.61M
Bob Buhl '57 4.55M
Earl Moore 1901 4.72M
Van Mungo '36 5.83M (one of the new cards)
Steve Blass '71 4.46M

Their "rating errors" as described above range from -14.87 to -18.2. 81 of the cards have "ratings" between +10 and -10.

Friend '56 is +47.49. The next worst card is Jack McDowell '92 at +24.

So apparently your card data that you used for developing the card price does not agree with what I see on the card image. This is very disturbing.

Either:

1. You are being given incorrect card ratings from strat-o-matic (in which case playing the game on this site is a freaking waste of time).

2. Your pitcher pricing model is hopelessly messed up (in which case newbies have no chance until they become sophisticated enough to start doing comparative evaluations of this sort for themselves and the rest of us are wasting our time).

3. There are errors in the trasncription of the card information from SOM to this web site (9n which case playing the game on this site is a freaking waste of time).

Which is it, huh? And when can we get it fixed? Yesterday would not be too soon.

Marc, thanks for the hard work and detailed answers here.

I wasn't suggesting artificially creating value cards it just seems to me attempting to get everything as close to some perceived perfection possible, is unnecessary and in some cases less fun.