Sunday, February 24, 2013

Baseball Mogul 2014: Fixing The Game (Part 2 of 2)

As mentioned last week, we have been reworking the General Manager AI so that computer-controlled teams are much more intelligent when it comes to trades and roster management. This also means that the computer-controlled teams won't have to "cheat" on high difficulty levels just to provide a challenge. But the other weakness that gets abused is the player rating system.

Player Ratings

One solution to the problem of accurate player ratings is to make them less accurate. This works, but leads to some really dumb results, like your scouts telling you that Ubaldo Jiminez has a "95" Control rating.

The other option is to simply turn ratings off. Baseball Mogul lets you do this, and it's actually a good option in my opinion. After all, the game is much more realistic when you turn off the ability to see player ratings. Nobody in Major League Baseball has a crystal ball regarding player abilities. You can't just give your scouts more money and magically gain access to the "true" ability level of every player in baseball. If a GM could view the Strat-O-Matic cards for every single player in his organization, the job of talent assessment would be pretty boring.

Strat-O-Matic Cards

So, real front office personnel don't have a binder filled with scouting reports that say things like "Albert Pujols has a 93 Power" and "Derek Jeter has a 67 Range". Thanks to statistical analysis (by everyone from Bill James to Clay Davenport, Voros McCracken and Nate Silver)* we now know that stats are better than talent scouts at predicting player performance. In the year 2013, you are more likely to see a GM consulting PECOTA projections than asking his scouting staff whether Felix Hernandez is on the downside of his career.

But, player ratings have a long history in baseball simulations, with good reason. Ratings are convenient. Without ratings, all you have are stats. For a minor league player, you have to adjust those stats for the level of competition and the ballpark he's in. And then translate it in your head to some general sense of the player's quality, so that you can decide if he should be promoted or traded. That's not much fun. I'd rather just have a few numbers that summarize a player, so that I can spend my time building my team and playing the games.

Realistic Scouting

Thus, the creation of a mode simply called "Realistic Scouting". Baseball Mogul is a computer game, but it doesn't have to feel like a simulation. You should have access to ratings and predictions for each player, but they should only be the ratings and predictions that would be available to an actual general manager.

"Realistic Scouting" includes many of the ratings you have seen before, but they are now derived from a player's actual historical (major league and minor league) stats. And, now that we are including complete real-life minor league stats, we can calculate Major League Equivalencies for every year that a player spends down on the farm.

Sandy Koufax scouting report (click to expand)

More Scouting Options

We've also added an option called "Variable Scouting", where your scouts are better at judging veterans than they are at judging 18-year-old high school players. This just makes sense, and addresses the problem mentioned above where your scouts aren't sure whether Miguel Cabrera or Torii Hunter would make a better cleanup hitter.

Finally, some cosmetic changes. You can change from the 50-100 rating system to the 20-80 scale used by Major League scouts. You can also choose 25-95 (like Art of Baseball) or 1-20 (like Championship Manager). No screen shots yet, but I am using Realistic Scouting in much of my testing and I prefer it to previous versions.

* My apologies for not including the hundreds of researchers, writers and bloggers that have helped us understand and appreciate the sport we love. On a personal level, extra special kudos to Christina, Joe, Rany and all of BP, THT, the Baseball Analysts, the authors of The Book, and "the Seans" (Lahman, Forman and Sullivan).


Tim said...

I've been playing since one of the very first versions and my biggest gripe has been how easy it's been even at the mogul rating. Therefore I'm really excited about the changes you referenced on both posts. Like a previous comment mentioned, I'd love to see more realism (in addition, how can I have 4 people at the same level starting? Would make sense to have the option to form actual teams at each level, but I can imagine how much more involved that would be)like 40-man rosters, options, etc so thank you for answering that question. I'm also curious about when/if we'll see the option to give players contracts prior to arbitration? A lot of the smaller market teams either sign their top players long term (i.e. Matt Moore) or trade them. I know we can always release the player and sign them but then we have to compete with other teams.


Mike H said...

(I posted this on Part 1 by accident, meant to post this on newer post here. Feel free to delete first comment on Post 1)

Clay, I absolutely love Baseball Mogul even if it is too easy to dominate on Mogul difficulty level. And I just wandered over to this blog here and enjoyed many of the non Baseball Mogul posts as well. Really appreciate all the work you put into what is a great, fun game.

One suggestion to make scouting a bit more realistic/harder would be to add a "floor" rating along with a "peak" rating, which is consistent with real life scouting. Like in real life, some of the prospects with a high ceiling/peak also have a low floor and therefore a greater risk of completely failing out. The guys with lower peak ratings could have higher floor ratings and therefore be safer choices, very much like real life. Of course, not every player with a high peak has to have a lower floor, like high first round picks (eg. Steven Strasburg), but players that tend to get drafted at the end of round one with high peaks also usually have a high "bust" or floor potential as well (eg. Andrew Brackman).

A more obvious fix would be to limit how good your scouting staff and farm can be. Like in real life, no matter how much money you throw at scouting and the farm, there is only so much better you can get compared to the rest of the league.

Just something to think about, and again, love the game. This may even be somehow incorporated into the game already but can't hurt to put it out there anyway. Thanks for all your hard work and looking forward to future versions of the game.

Anonymous said...

If realism is what we're after, then I would love to see the added ability to hire/fire scouts. Each scout would have ratings based on positions, some strong scouting pitchers, infielders, catchers, outfielders, etc.
This was something used in a game I still love to play on the PS3 called Football Manager 2009. This allows you to make it less about how much money you throw at scouting (it's stupid how you can throw millions for half a season then cut it in half and still keep an A+. This also makes it less tedious to monitor the allocation of money to scouting.
You sign a scout for each level MLB, AAA,AA, etc. each with strengths and weaknesses, some with more or less of both depending on what you are willing to pay. It's really a brilliant system. Sign them for #years much like a player contract.

Anonymous said...

Believe it or not, I use Baseball Mogul's rating system to catch the up and coming baseball player. Jose Reyes was one example. If the SF Giants had listen to me and Larry Krueger, they would have plug the hole at Short for last eight years and maybe winning three consecutive world series. Anyway keep up the good works and please allow the rating system to be an option to turn on and off so I can get a sense of this year's baseball prospect and young players around the league.

Victoria said...

This is cool!

Anonymous said...

I want you to know that I am buying BB Mogul 2014 (the first BB mogul I've bought in about 5 years) solely based upon what you've done with scouting and the minors. I like the way you're thinking.

I love bringing players up through the minors and watching them develop, it's my favorite part of a baseball sim. But the only sim that made me really feel that I was doing that in the past was OOTP - Puresim and BBMogul made it too obvious when you should promote a player and how good he was.

However OOTP always crashes a lot for me, their tech support is non-existent, and I've been waiting for years for one of the other franchises to step up and compete. I'm downloading as I type this and I hope I'll be playing this new and improved Mogul for a long time to come.

Anonymous said...

This is excellent stuff. Great work by the best in the biz.

One suggestion that I have is for you to create a different rating system for prospects and MLB players. After all, in real baseball, you don't look much at 20-80 (or 2-8) ratings for a handful of skills for MLB players, but naturally those are on the forefront for prospects.

The idea is that we don't want to be rid of being able to see some simple ratings for prospects, but you want something more realistic than overall/potential. So for prospects, have the usual tools with a 2-8 rating, plus an overall on the same scale. You can have variability for each, but the overall would give you a sense of general quality (so only the top prospects are 7+, and they are indeed top prospects,etc). This would also allow you to introduce separate budgets for minor league and major league scouting, something that also makes more sense to me (why not have "draft prospect" scouting too, which affects what you see in the draft window?).

Another option you could add is an "industry consensus", something that seems to be missing from the game. In real life, there is a broad consensus ranking of draft and minor league prospects, but in this game you focus primarily on your own scouts. How about creating an industry consensus that every team can see, and then have scouting provide you with the variable nuggets that can help you make better/worse decisions? This would add even more realism, plus provide you with a narrative bonus, as you could finally include "top 100 prospects" lists and top draft prospect lists. Real life fans eat such lists up, so I can't imagine that this wouldn't be eaten up by fans of the game.

It would help if the variability between consensus and scouting were a bit more dynamic too. In other words, you're not going to see much variance regarding certain prospects (Buxton), but others naturally are bigger question marks.

Finally, if you were to implement such a system, it would be key to decide when players transition from the prospect rating system to the MLB system. It could be determined by playing time in the majors, seasons in the majors, end of rookie status, or some such.