Diamond Mind Baseball MB game considers many factors to determine the outcome of every play. Our studies of detailed play-by-play data from tens of thousands of real-life baseball games enable us to measure and accurately reflect the playing conditions of the era, ballpark and weather effects, fatigue, infield positioning, offensive and defensive strategies, the home field advantage, and other important elements of baseball. Statistical accuracy does not always mean that if a player hit 30 homers and drive in 111 runs in real life, he should do exactly the same thing in DMB. If he hit those 30 homers in a great homerun park, and you trade him to a team that plays in a pitchers park, it may be quite consistent with his ability for his homers to drop to 22. And his RBI totals might rise or fall if he's put into a different spot in the batting order or traded to a team with a better or worse hitting team than the one he was on in real life. To be considered statistically accurate, we believe a game must separate the contribution made by the talent of the player from the contribution made by the context (league, park, team, weather) in which he compiled his stats. This is necessary to achieve the following goals: - if he is used in exactly the same context in your computer-league games as he was in real life, his statistics should match his real-life totals in the long run. In the short run, of course, streaks and slumps may take him away from the target for a period of time, sometimes even for a full season. But if you play several seasons in the same context, the stats should be very, very close to real life. - if he is used in a different context, his stats must change to reflect the new playing conditions. If a game slavishly reproduces real-life stats under dramatically different conditions, it's just as bad as if it failed to match his real-life stats when playing conditions are the same. The most important factors in any baseball confrontation are the abilities of the batter, pitcher and fielders. DMB player ratings are based on extensive research of all relevant factors that affect player performance, overcoming many of the illusions that can occur in the traditional statistics. This is true of all DMB ratings, but we will use fielding range as an example to illustrate the point. For example, consider two players who handled 300 chances in 100 games at third base. Are they equally good fielders? A simple calculation of chances per game might lead you to believe that they should be assigned the same range rating. But what if one player started all 100 games while the other started less often but was used as a late-inning replacement? What if one was on a team of control pitchers who created a lot of ground balls, while the other played behind a staff that struck out many more batters. What if one player caught many popups while the other handled more tough ground balls? What if one faced many more right-handed hitters, who are twice as likely to hit grounders to third as are left-handed hitters? What if one played on grass and the other on artificial turf? When assigning fielding ratings, DMB examines the number of balls hit into the area each fielder is responsible for and identifies the balls that were fielded successfully. These measurements are not perfect, but they are much better than judging talent based on comments by television analysts or by counting the number of times a player shows up on highlight films. As a result, when you play DMB games, you can be sure that your DMB players will perform to their real-life abilities. Does this mean that a DMB season will produce exactly the same results as were seen in real life? No, it doesn't, for some very good reasons. Real-life players are not totally predictable. The best team and the best players don't always win. Sometimes a great pitch is blooped over the shortstop's head for a base hit. Sometimes your line drives are right at people. Sometimes you get a break on a bad call by the umpire. Most of the time, these things even out in the long run. But not always. It is common to see a player's batting average or earned run average change by 10-20% from one season to the next, even without any apparent change in that player's ability. When you replay a real-life season with DMB, most players will produce statistics that are very close to their real-life totals. But luck will cause a few to vary by 10-20%, just as in real life. In addition, some statistics are heavily influenced by factors beyond a player's control. A pitcher's win-loss record, for example, depends as much on how many runs the offense scores as it does on his own ability. And his earned run average depends, in part, on how well his relief pitchers get out of jams he creates. Beyond luck and external factors, however, there are a number of other reasons why your statistics might differ from real-life. The remainder of this note discusses these reasons. The "Elite Talent Pool" Effect ------------------------------ In real life, a player's statistics were compiled against a large number of players, including everyday players, rookies, and declining veterans. The everyday players are generally very good, but many of the others are marginal players who get a little playing time and are released or put back on the reserve roster. Draft leagues (leagues where all players are released and drafted onto new rosters) can create a strong elite-talent-pool effect if they do not contain as many teams as in real life. Suppose your draft league contains 16 teams drawn from a population of 28 real-life teams. Your rosters will be packed with the 400-500 best players in the game. Your batters will face better pitching than in real life, and your pitchers will face better hitting. DMB will accurately reflect their real-life talent levels, but their statistics will suffer because the quality of the opposition is so much higher. Pitchers will be hurt more than hitters, since the offensive manager has more options (platooning and pinch hitting) for creating great matchups. Left-handed pitchers will be especially vulnerable, since every team they face will have many strong right-handed hitters with which to pack the lineup. Overall, you can expect to see more scoring in a league like this. The Perfect Hindsight Effect ---------------------------- Real-life managers start the season with many questions. Is last year's rookie sensation for real? Will my ace pitcher bounce back from shoulder surgery? Does the aging veteran have one more good year, or is he washed up? Ultimately, the only way to find out is to give the player enough playing time to prove that he can or cannot do the job. Some of these experiments are great successes, while others are dismal failures. The DMB manager has the benefit of knowing how these experiments turned out, and can choose to give playing time only to the players who performed well. After all, who wants to play a .180 hitter or a pitcher with a 6.29 ERA? In doing this, the DMB manager is creating another form of the "Elite Talent Pool" effect. The other batters and pitchers do not get a chance to fatten up their statistics against the weaker players, as they did in real life, so their DMB statistics will suffer, even though their talent is being accurately reflected. The DMB manager also has the benefit of accurate information on how batters performed versus left- and right-handed pitchers. This information can be used to construct devastating starting lineups and platoon pairings. Again, because the offensive manager has more flexibility than the defensive manager, overall levels of run scoring will rise if the DMB manager uses this tactic. In real life games, we frequently find players who attempted many steals despite having an average to below-average success rate. The DMB manager can use perfect hindsight to refrain from stealing with these players. By reducing the number of outs given up by being caught stealing, DMB managers can generate more big innings and a higher number of runs. The Ballpark Effect ------------------- Ballparks have a significant impact on real-life statistics. It is not unusual for a power hitter to hit half as many home runs at home as on the road if his home park favors pitchers. If that player is traded to a team with a good hitter's park, his home run totals can be expected to rise significantly. Of course, players traded from a hitter's park to a pitcher's park will experience a decline in their home run and other batting totals. The same is true in DMB games. If you are playing in a draft league, most of your players will be playing in a different ballpark than the one in which they compiled their real-life statistics. This affects their statistics in DMB games just as much as it does in real life. If your draft league does not use all of the real-life ballparks, you may find the overall level of run scoring going up or down, depending on the mix of ballparks you are using. The Manager Effect ------------------ Your managing style may also affect the results. If you use the bunt or hit and run much more or less than the real life manager did, you may see very different results. If you use your pitching staff differently, it could also have an effect. Changes in the lineup can also have a major impact. For example, runs scored and runs batted in totals depend much more on where a player bats in the lineup than on skill alone. If you move someone from fourth to second in the batting order, you can expect a large drop in runs batted in and an increase in runs scored. The Injury Effect ----------------- DMB's injury system gives you three options. Whether you choose None (no injuries), Injury Rating (injuries based on each player's injury rating) or Random (all players have an equal chance of getting hurt), there is no guarantee that your DMB players will be hurt for exactly the same number of games as in real life. As a result, it is possible for your team and league totals to vary from real life simply because your best players had more or less playing time in your DMB league than in real life. The Weather Effect ------------------ DMB's weather system is designed to produce weather that is very similar to the weather in real-life games. In real life, the weather is not always the same from one year to the next. Some years are warmer than others and sometimes the wind blows out more often. You will find similar variations in DMB seasons, and you might see a small difference in results because of changing weather patterns. The Designated Hitter Effect ---------------------------- Consider two identical pitchers, one playing in a DH league and one playing in a league without the DH. The DH-league pitcher can expect to allow more hits, have a higher walk-to-strikeout ratio, and have an ERA that is, on average, half a run higher. If you are playing in a draft league that uses the designated hitter, your non-DH pitchers will see all of their numbers go up accordingly. And your overall league ERA could go up by a quarter of a run, since half of your pitchers will suffer the half-a-run increase. Achieving the Highest Levels of Statistical Accuracy ---------------------------------------------------- The previous sections have described several situations in which you should expect to see differences between real-life and DMB statistics, even when DMB is accurately reflecting real-life player abilities. The following paragraphs present some ideas for you to consider in choosing how to conduct your league. If you want to replicate real-life statistics, use the same schedule, rosters, lineups, strategies, and ballparks as in real life. For information on roster moves and starting lineups, pick up one of the many books and magazines that include a complete set of boxscores for a season. If you can, use DMB's automatic transaction log feature and game-by-game lineups to get as close as possible to real-life matchups. However, if you want to change some of these variables (and much of the fun of DMB is the ability to do just that), use the previous sections to understand how your choices might affect the statistics. If you want to compensate for some of these effects, consider the following ideas: - in a draft league, make sure the talent pool includes weaker players. For example, if your league includes only 12 teams, consider using the players from only one real-life league. You may find that the challenge of handling a roster with some weaknesses compensates for leaving out some popular players. - make sure your draft league uses a mix of hitter's and pitcher's ballparks - if you are using the computer manager to replay a real-life season, and you cannot take advantage of the new automatic transaction processing feature, set up your starting rotation, saved lineups and spot starters to approximate real-life playing time as much as possible. Make a few extra players active, so reserves are available to fill in for injured starters. And consider playing the season a week or month at a time, so you can make important roster moves and adjust manager profiles to reflect real-life changes in player roles. - consider using a different Era than the one from in real-life. Suppose you are playing in a draft league that has an elite talent pool, uses the designated hitter, and encourages managers to take full advantage of perfect hindsight. But you do not want to see the league earned run average go up as a result. To compensate, use an Era based on one of the real-life seasons (such as the mid to late 1960s) in which pitching dominated.