DMB News April 2001
Diamond Mind Email Newsletter
April 27, 2001
Written by Tom Tippett
Welcome to the second edition of the Diamond Mind email newsletter for the year 2001. Through these newsletters, we will try to keep you up to date on the latest product and technical information about the Diamond Mind Baseball game, related player disks, and our ongoing baseball research efforts. Back issues are available on our website.
Topics for this issue
Version 8c patch now available
Injury lengths in version 8
Passive play-by-play feature
Playing head-to-head over the Internet using NetMeeting
Near-term product plans
2001 Projections update
Earlier this week, we posted the version 8B patch to our web site. That patch included a few enhancements and fixed some important bugs, and we recommended that all version 8 users download and install the patch as soon as possible.
Unfortunately, we introduced a new bug (that interferes with a human manager's ability to choose the base to throw to in certain baserunning situations) while making one of the bug fixes. So we immediately corrected that bug and created a new patch called 8C that you can now download.
Whether or not you installed the version 8B patch, we strongly recommend that you download the version 8C patch as soon as possible. The version 8C patch includes all changes since version 8 began shipping in December, so there is no need to install any previous patches before installing this one.
To download the patch, visit our web site and follow the link to the version 8 support page. There you'll find bug reports, trouble-shooting tips, answers to common questions, plus a link to the pages from which you can download the latest patch and read about the changes in each patch.
Here's a quick summary of the most important enhancements and bug fixes in this patch (see the patch pages for the full list):
- it was possible to cause the game to skip a batter in the lineup by pressing the 2, 3 or H key at a time when the game was NOT asking you where to throw the ball; these keys are now processed only when they are supposed to be
- for catchers who are used very heavily (140+ starts per season), the catcher fatigue system was causing them to tire a little too quickly
- when playing league games using the Injury rule of "Injury Rating", the game is supposed to limit any single injury to the number of games that player missed in real life. This limit was not being correctly enforced, so players were sustaining longer injuries than they should.
- the computer manager was not inserting platoon pinch hitters and defensive subs as often as it should. (Please keep in mind that if you are using the "Limit bench playing time" rule for your league, the computer manager won't use a platoon pinch hitter who is ahead of his playing time pace for the year. So if you see a situation where a pinch hitter seems to be called for but wasn't used, this might be the reason.)
- the Team vs Team report has been added
- when exporting a league database at the beginning of a play cycle, you now have the option to export all boxscores or only those boxscores that have been added since the last play cycle, thus providing a way to dramatically reduce the size of the transfer files for leagues that share all of their boxscores
NOTE: Because this feature has been added to version 8C, the commissioner and all league managers must be using the same version of the game. If you receive the message "A required database file (dmbera.dat) is missing from the transfer file", it means that the commissioner has installed version 8C and you need to do so as well.
A couple of you expressed concern that the injury system is too severe in version 8. With the exception of the bug noted above, our testing found that this was not the case. We simulated five seasons with the version 7 and version 8 editions of the 2000 Season Disk, and the results were as follows:
Version 7 Version 8
Avg # of injuries per season 698 704 Avg duration (games) 5.2 6.5 Maximum duration (games) 54 55
At first glance, it appears that the average injury length is up by more than a game, but the change is actually much smaller than that. Version 7 always reported injury lengths without including the game in which the injury occurred, while the version 8 injury logs included the initial game. So the difference is really between 5.2 and 5.5 games per injury, and that's as we intended it, because we made a conscious decision to increase the length of some injuries in version 8.
By the way, the change from excluding to including the initial game was unintended, so the version 8B patch restored the old convention of excluding that initial game. This change has no effect on the actual length of the injuries, just how they appear on the injury log report.
A lot of people have requested that we add back the ability to sit back and watch the play-by-play without hitting the Go button before each play. To that end, we spent several days working on this feature last week. Unfortunately, we ran into some unexpected technical problems, and because it wasn't clear how much longer it would take to finish the job, we decided to release the patch without this feature so we could get the bug fixes into your hands as quickly as possible. We haven't given up on this, however, and we'll be getting back to work on this feature next week.
Some customers have been disappointed to learn that version 8 doesn't work as well as version 7 when using NetMeeting to play head-to-head over the Internet. As a result, we spent quite a bit of time looking into this issue in recent weeks.
One issue is speed. The version 8 game screen is more graphical and therefore requires more bits to be sent back and forth than was the case in version 7. For those of you who have moved to faster Internet connections in recent years, it may not be an issue. If you have a slower connection, it's entirely possible that version 8 over NetMeeting won't be fast enough to suit you. Unfortunately, there's nothing we can do to reduce the number of bits that NetMeeting has to send back and forth.
The more important issue is control. In version 7, the two players could use the keyboard to drive the menus. It was necessary to get into a rhythm so you didn't get in each other's way, but the sequence of menus that the game puts up (defense goes first, offense second) made it relatively easy to get into that rhythm.
The same cycle of defense-then-offense exists in version 8, and version 8 allows the game to be driven by the keyboard just as version 7 was, but we were surprised to learn that NetMeeting forces the guest to request control (by double-clicking the mouse) of the keyboard each time. This was not necessary with version 7.
After reviewing Microsoft's NetMeeting developers guide, testing other Windows products over NetMeeting, and asking some DMB customers who work for Microsoft, we have concluded that there's nothing we can do about this. NetMeeting appears to have one set of rules for DOS programs (like version 7) and another for Windows applications (like version 8).
As a result of this discovery, we are going to look into what it will take to add direct Internet support into the game. We hope to begin that work soon, but until we get into it, we won't know whether it will fit nicely into our current architecture or require a massive overhaul. So it will be a while before we'll know how this would work or how long we think it will take to add this capability. We'll keep you posted via our web site and our email newsletters.
In the meantime, some of you may find that NetMeeting play works well enough. If you have a reasonably fast connection and the guest doesn't mind double-clicking to request control before each play, it still provides a no-cost way to play head-to-head over the net right now.
To help you learn how best to do this, we'll be posting a technical note on NetMeeting play to the version 8 support area on our web site next week.
Now that the version 8C patch is available, we'll be getting back to work on the 'passive play-by-play' feature and we'll continue to work on fixing bugs. These changes will be made available in another patch when they are ready.
We are also planning to release an updated version 8 Help file that incorporates new tutorials and adds new material in response to the most common questions that have come up in our technical support activity over the past few months. This update will be made available as a free download from our web site.
We have begun work on the 1978-1981 Deluxe Past Seasons and expect to have all four seasons available by the middle of the baseball season. Thanks to the great work of the volunteers at Retrosheet (www.retrosheet.org), we have access to play-by-play accounts for every game from those seasons. That means we can provide left-right splits and use our most advanced analysis programs to help us develop highly-accurate player ratings for those years.
We also expect to add several new seasons to our Classic Past Season collection. The 1920, 1935, 1940, and 1943 seasons are in various stages of development, and we expect to be able to release them before the baseball season has concluded.
1. Only the league commissioner should use the Transfer > Export league database command. Some managers have been using this command to make backups of their databases, not knowing that this command does more than just archive their files, and this interferes with their ability to export game results later. If you want to make a backup of your database, use the File > Backup command instead.
2. Some league managers have been moving or renaming the boxscore files that are stored in the folder containing their league database. If you do this, the boxscore files will NOT be included when you export your statistics and your commissioner will not be able to import your game results. If you want to rename these files or send copies to another manager in your league, copy (don't move) them to another folder and make your changes there.
3. When sending or receiving league databases or statistics files, don't zip or unzip those files yourself. DMB's transfer functions take care of this for you, and those functions also make important database updates. If you go outside the game and zip/unzip the files yourself, you lose the benefit of these updates and risk damage to the integrity of the database.
As many of you know, we used our Projection Disk to project the final standings for the 2001 season. This is the fourth year we've done this, and the methodology was essentially the same as before -- we ran the season 50 times and averaged the results.
Every Monday morning during the season, we update a spreadsheet that compares the real-life results to our projections. I thought it might be interesting to share a few of the things that we found in our review of the first three weeks.
The hottest topic of the winter and spring was the new strike zone. At the time we released the early-March edition of the 2001 Projection Disk, we felt that any attempt to estimate its effect would be pure guesswork, so we released the projection disk with Eras based on the 2000 season results.
Ten days later, when it came time to run our fifty simulations, it had become clear that the new strike zone was having a major impact on spring games -- walks were down 22%, strikeouts were up 9%, and scoring was down 16%. So we adjusted the Eras that we used for the simulations to reflect these changes. But we weren't convinced that the full-season impact would be quite so large, so we muted them a little.
The result was a pair of Eras designed to produce an 18% decrease in walks, a 7% increase in strikeouts, 12% fewer runs, a 4% drop in homers, and league batting averages about ten points lower than last year.
(Those figures represent our estimates of the effect of the strike zone, but that wasn't the only change that could influence the stats for the coming season. The two new NL parks and the moved-in fences in Cincinnati should produce an increase in homers in that league. It remains to be seen whether the strike zone or the parks will have the larger impact in the senior circuit.)
After running our fifty seasons, we continued to collect data on spring games, and the final spring totals showed a 20% decrease in walks, 8% more strikeouts, and a 12% decrease in scoring. That seemed to confirm our hunch that things might moderate a little.
Now, three weeks into the regular season, what are we seeing? Walks are indeed down, but only by 12% compared to the per-game average for the full 2000 season. Strikeouts are up 6%. The league batting averages are down by about ten points. Homers are down 5% in the AL but up 6% in the NL. Runs per game are down 8% in the AL and 10% in the NL.
So far, the spring training numbers have proven to be fairly good predictors of the real season. It's still very early, of course, so we'll continue to monitor and report on these trends as the season progresses.
The other part of our weekly review focuses on the teams. And this is where I get to brag about how we predicted that the Twins, Phillies and Cubs would all be leading their divisions while Oakland, St. Louis, and Atlanta would be among the bottom three in their respective leagues in run differential. Right?
Well, of course we didn't foresee all of those things, and I think it's fair to say that nobody else did either. Because there are still 140 games to play, it would be easy to dismiss these results as being meaningless. But terrible teams don't go 14-3 to start a season very often, and I believe the early successes and failures offer some clues about the rest of the season.
The AL East is going pretty much according to form except that Toronto's pitching has been better than expected. In most of our simulations, the Red Sox took the early lead thanks to a very easy schedule, so their perch atop the division isn't a surprise.
One of the most controversial results of our simulations was Cleveland finishing ten games ahead of the White Sox. The Indians are already five games up on Chicago, but that 'race' is being overshadowed by the Twins' amazing start. When I wrote the Minnesota team review in February, I noticed a whole bunch of guys who crushed AAA pitching last year but didn't hit much in the big leagues. It's possible these guys really are good and just needed more time than most to adjust to the majors. They won't maintain their current pace, but they could be this year's version of the 2000 Marlins.
Last month, a few Oakland fans gave us a lot of grief for failing to project 100+ wins for their boys. Our simulations suggested that their tough schedule would hold them to the low 90s, but even that seems highly optimistic at the moment. Led by Ichiro Suzuki, Seattle has taken 15 of 19 from its division rivals before flying across the country and sweeping a series with the Yankees. I'm not sold on the bottom half of their batting order, but Seattle's 8-game lead has to be taken seriously.
The surprise team of the NL East is the Phillies, of course, who currently rank third in the league in pitching. I don't think they can keep that up, but they did finish 5th in the NL in pitching in our simulations, and that was a big reason why we had them bouncing back from their awful 2000 season to finish third in the division.
When we ran our fifty seasons, we were surprised to see the Cubs rebound from an NL-worst 65 wins in 2000 to finish fourth in the division. So their hot start isn't quite the shock that it might otherwise be. I don't expect their pitching to hold up all year, though.
Houston finished only a game behind St. Louis in our simulations, and with Mark McGwire on the disabled list for a while and Rick Ankiel walking a batter an inning, I now believe the Astros are the favorites to win the division.
Finally, we had Colorado coming out on top of the NL West, and after three weeks they're on the pace they set in our fifty seasons. In those simulations, the Giants were barely over .500, which seemed odd given their strong finish in 2000. But they've been outscored by 19 runs and currently sit in fourth place with an 11-10 record. If not for a 5-0 record in one-run games, they'd be looking even worse right now.