Comparing cards from different eras

I have a question about how to factor in era normalization for my drafts.

I have always assumed that cards are set up to perform realistically against the pitching cards of that year. So Joey Votto hits 305 in a 2013 league where the league's pitchers allowed a batting average of .251.

Obviously that means if you were to insert a Votto card into a league consisting of pitchers cards from 1930 he would hit higher against them and hit lower in a league consisting of pitching cards from 1968.

In other words Votto's card stays the same but his numbers are also determined by the pitchers he is facing.

So in my shorthand calculation, Votto who's .305 avg when normalized for the 1930 NL is 368, would hit about .333 in a 1930 STRAT league because of the 50/50 rule.

But thats not how things are working out for me. Dick Allen's monster 1972 year, during one of the worst hitting years ever, isn't hitting any better than any other guy with similar numbers.

And 1930 Dazzy Vance, a great pitching year in the best hitting year ever, doesn't see to be pitching much better than any other guy with a 2.60 ERA.

The records listings also seem to bear this out.

Have I been drafting all wrong? Should I just draft off raw numbers? Is there something more subtle going on?

The Answer is COMPLICATED. Yes cards are basically normalized to the year they represent, and if a season is replayed with that years cards using inning and AB limits, players should perform similarly. However if you play a draft league using one years cards and no limits, the pitchers will be playing with a different team behind them therefore different defenses, facing players who hit well in limited AB's more times, etc. Now, in this case you must factor in ATG players with very few "poor" seasons, and very limited scrubs, usually teams built to the ballpark they are playing in, and virtually no limits (injuries almost never cause a player to have less AB's than his card is built around). Also a player like Votto has to play in a hitter's park to perform even close to as well as he does in Riverfront. And a player like Vance, although he pitched in a bandbox (Ebbets Field) and had a great season in spite of pitching to a league that hit over .300. That kind of a performance will not translate well to ATG players who routinely hit over .300 one to nine!

All this to say, the pitching cards face the hitting cards and in general if the hitting cards are better than the pitching card they are facing, in this game they will out perform the pitching, almost always. Strat in general is slanted to offense and outlier pitching performances can really not be duplicated without limits of some sort, low ballpark numbers, salary limits for hitters, or AB limits for fringe or low AB players.

Another flaw in your logic of Votto now verses 1930 is the card now is built assuming a league full of pitchers who give up some HRS. Thus it is expected he will get a certain number off pitchers. He likely loses those HRs if he is plopped in to a 1930 league. Thus his power and BA may go down.

Another factor in ATG is you frequently play in a league with unbalanced parks. One league may have mostly hitter friendly parks while the next may be full of pitcher friendly parks. Or there may be multiple lefty or right favoring parks. From one league to the next you may see a substantially different mix of pitchers/hitters, especially in the lower caps. So results can vary a lot from league to league.

You need to look at real life stats as a broad indicator of a player's profile. So if you want to build around power then sort by HRs. But once you get that list on screen look closer at what the individual cards among the top choices look like. Depending on the real life stats too much overlooks the subtle things that go in to a card makeup and assumes Strat is always right on with setting the card up.

Go to www.diamonddope.com, select pool from the menu, select "zero park effects" from the drop-down menu, and then copy and paste the web page that comes up (it takes a while) into Excel.

You will now have the card chances for every player in the ATG set. This is the only way to compare cards from different eras. Trying to play this game based solely on the player's stats is just going to drive you crazy - there are park effects to consider, defenses, and league norms.

Hey guys thanks much for you feedback.
I guess when it comes to comparing players from different eras, don't.
And thanks for the link, I hate counting chances.

But not to belabor a point, but it still seems counterintuitive to me, that all things being equal, a card that is designed to have a 1049 OPS in a league with a 674 OPS and in a pitchers park (Frank Robinson 1966) isn't tearing things up.

Or Allen 1972, or Stargell and Torre 1971, or Schmidt 1981.

Even given the context of a ATG draft league, all the other guys are playing in the same league too.

Same thing with Vance 1930. Here is a card designed to have a 2.60 era and a 10.3 WHIP while pitching in a hitters park in a league that batted .303 and had a 4.97 era!

After all his card should be LESS effected by a league where everyone hits .300

Well thanks much anyway, time for me to pull out my Calculus text book.

Ok, your point finally sunk in gravity.

Here is a partial answer.

I know the hype is that Strato does exhaustive research and designs cards to reproduce seasons based on opponents, parks, etc for a given season, yada, yada. yada. And there may be some truth in that for recent seasons.

But keep in mind that for historical seasons this exhaustive research may be as much mirage as fact. Many seasons were computer generated with the stats fed in to a computer. This is especially true the further back you go both with real life seasons and when the year was first added to the Strato offerings. Simple truth is at the time some of these were first offered detailed stats needed for truly producing the kind of accuracy the hype implies simply was not available.

Now some of the seasons have been updated but generally they were updated in order of popularity and/or expectations of demand and sales generation. So can be a bit tricky to pick exactly which years are which. Some have gone by the prices on the product website and referred to various seasons as chevy and cadillac seasons.

Chevy seasons are totally computer generated. Stats were just fed in and cards produced using a computer formula. You can recognize these not only from the cheaper price on the site but also by some common characteristics of the cards. For example you will only see gb(C) on pitcher cards and no gb(A). Double plays were not factored in to the pitcher card. All double plays were accounted for on the hitter cards. Thus hitters from these seasons when tossed in to something like ATG will hit in to more double plays than they should. Pitchers from these sets who induced a lot of double plays will lose that quality and be less effective than real life. Note the difference between the 30 Vance and the 34 Vance. 1934 was one of the first seasons to be upgraded.

Both hitters and pitchers will have balanced or even identical L/R cards because reliable L/R stats were not available at the time of generation. Note for example the 1901 Cy Young card. The left and right sides are identical except for the placement of outs. The computer generation algorithm reproduced strikeouts but out distribution was probably randomized to some degree.

Some seasons have been updated for super advanced rules. These seasons are referred to by computer gamers as "Cadillac seasons". On these you will see gb(A) on most pitchers with hitters tending to have fewer on their cards. Cards will be unbalanced to reflect what they actually did against L/R opposition. But it is my opinion that even these Cadillac seasons are not all created equal. Officially it is said that extensive research went in to them but of course exactly what that research was has never been specified. I doubt they went back and analyzed boxscores. More likely they used purchased data. But if a season were upgraded in say the mid 90s the depth of the data that was available is simply not even close to what is available now. So no way a season upgraded then could have the same detailed data available for upgrading one today.

Some of this opinion is based on things I read in Strat Fan Magazine years ago, some from brief time I spent on strato bulletin boards, and some just my opinion based on my intuition and what I believe could and could not have been done at the time it was done.

Wow! The most important thing to consider is not the era a player played in but the era in which Strat created the card.

hey gravity a lot of those players cards with monster years are very one-sided. those type cards are easy to counter with like sided pitchers cards. why would I throw left handed pitchers against a lineup with allen, Schmidt, bagwell, etc.? give a pitchers park with a very low #hr rating, there's several pitcher's under 4 mil that will shut them down.

Still, I've noticed (and the diamonddope chances seem to back this up) that pitching cards from hitters years and hitting cards from pitchers years, not just monster years but good and mediocore years as well, don't seem to do as well as expected. Not just guys with huge L/R inbalances - but almost all of them.

I mean why doesn't Allen 1972 with a OPS+ of 199 have a much better hitting card thnt Foster 77 with a 162 OPS+.

I think Valen probably has it right. The cards were not created using the same data and methods so they don't work so well, in terms of accuracy, cross era as in their designated season.

Strat's ballpark factors often have little relation to the sabermetric park factors used to generate OPS+. Crosley 56 was rated as singles 8/8 HR 2/8. So Big Frank's 1956 season was accomplished in a relatively neutral park.