Interesting Discoveries Re: Off Days

[kq76]

I know, I know, some of you are instantly going to react, "well duh!". But none of this had occurred to me before.

While using one of my schedules after the All-Star Break I got hit with a bunch of rainout makeups. My team ended up playing 16 games in 15 straight days and a lot of my players were tired or exhausted. I managed to get through it without any major long-term injuries, but I vowed to look into it. I thought I must not have scheduled enough off days or maybe I bunched them up early or late.

I found my % of off days was actually fine (13.8% vs MLB's 12.9%) and they were spread throughout the season fairly evenly. But what I realized was that if I scheduled more off days later in the season (as opposed to evenly spread out) then there are more chances that rainouts can be made up while minimizing the number of days in a row games must be played without an off day (early off days are not likely going to be needed for makeup days while later off days are as there are more chances for rainouts as the season progresses). And even if OOTP schedules doubleheaders instead, you'll still have those more off days late for your players to recover.

Early in the season would be a good time to schedule some 4-game early week series too (in the past I've primarily stuck with 2x3-game series a week).

Here's a tip: before you play with a new schedule, check the schedules of a few teams. If they average about 3 or 4 off days every month with maybe 1 or 2 less early and 1 or 2 more late, then it's probably a pretty good schedule. If you don't see that, then you might want to reconsider using it.

EDIT: I'm not so sure about overloading your schedule early anymore. Read on to learn more, particularly starting with gmo's post. It still might make sense to do that, but maybe not. I don't think you can really go all that wrong with trying to spread them out fairly evenly though. Yeah, I could have scheduled another off day shortly after the ASB, but I got unlucky with rainouts too.

The things you learn making schedules!

EDIT: If you want to easily calculate your off days %, duplicate this spreadsheet I made. The spread of the off days you'll have to check in OOTP.

I switched what SS I linked to as the way the original was made more for planning schedule dates and was therefore needlessly complicated for the purpose of calculating an off days %. This new SS is far more simple and gets right to the point.

[Ty Cobb]

We can all say, "Well, duh," if we want, but it's a darn good point to remind schedulers of. Good tip IMHFO. Thank you.

[Le Grande Orange]

Originally Posted by kq76

While using one of my schedules after the All-Star Break I got hit with a bunch of rainout makeups. My team ended up playing 16 games in 15 straight days and a lot of my players were tired or exhausted. I managed to get through it without any major long-term injuries, but I vowed to look into it.

My questions would be these:

(1) When exactly did the original postponements take place?

(2) Were there any off days following the original postponement dates which could have been used for a make up game but which were not used by OOTP?

(3) How many games were made up as part of a split or regular doubleheader as opposed to using an off day?


I know in the past OOTP didn't really make up games the way MLB does; i presume that's still the case. (And I know it didn't differentiate between major and minor league make up game procedures, as the minors use a simpler process.)

[kq76]

Yeah, it could be worth looking closer at the makeups...

April 10th's game vs Rio Grande (RGT) was made up the very next day with a doubleheader. That was Gulfport's (GUL) only series @ home vs RGT and they didn't have a day off after the series so I don't have a problem with that one. It might be nice if OOTP could schedule a makeup game on an off day after the team was playing close to the team they missed a game with (like if RGT were later playing Mobile, AL, then swung by to Gulfport, MS for one game), but I can't expect that.

GUL's May 17th game @ Cape Coral (CCF) was postponed to... a doubleheader on Aug 14, the stretch where we played 16 in 15 days. And that was the last time they were in CCF. They actually had an off day May 18th that would have made more sense to schedule the makeup game on. CCF also had a day off that day so why OOTP didn't schedule it then I don't know. Maybe you're right, OOTP refuses to schedule makeup games on teams' off days. I'll add that to my wishlist.

The rest of that stretch was my fault though as Aug 5-19 they had a game scheduled everyday. In the future I'm definitely going to schedule an off day in the middle of that.

The game scheduled for Aug 28 @ New Orleans (NOL) was made up as part of a doubleheader on the 29th. That was also the last time those 2 teams would matchup in New Orleans, but they both had an off day 3 days later on the 31st so why not schedule the makeup game then?

I wonder, IRL, do teams ever play a makeup game @ the originally scheduled road team's home ballpark? NOL & GUL didn't play each other again, but they are in the same division and so I wonder if it could have been an option to play NOL @ GUL later. I'd rather see a doubleheader, if possible, but if not, then I wonder if that can happen.

[Le Grande Orange]

Originally Posted by kq76

I wonder, IRL, do teams ever play a makeup game @ the originally scheduled road team's home ballpark?

Yes, if playing in the original park is no longer an option. That's been the case in MLB since 1910, according to the notes I have.

A change was made for 2007 which states that in the case of a postponed game being made up at the other club's park, that other club will bat first in the make up game even though the game is taking place at its park. The minors gradually adopted this rule also, and all the affiliated minors use it now.

[Le Grande Orange]

MLB postponement and rescheduling combined figures for 2009 through 2021, excluding the truncated 2020 season. The 2022 season is excluded due to its reworked, compacted nature as a result of the late lockout settlement.

23.8% of games were made up on a common off day as a single game
00.8% of games were made up on a common off day as a day-night doubleheader
01.7% of games were made up on a common off day as a regular doubleheader
44.1% of games were made up as either the day or night game of a day-night doubleheader
27.8% of games were made up as the second game of a regular doubleheader
01.7% of games were not made up

[kq76]

Originally Posted by Le Grande Orange

MLB postponement and rescheduling combined figures for 2009 through 2021, excluding the truncated 2020 season. The 2022 season is excluded due to its reworked, compacted nature as a result of the late lockout settlement.

23.8% of games were made up on a common off day as a single game
00.8% of games were made up on a common off day as a day-night doubleheader
01.7% of games were made up on a common off day as a regular doubleheader
44.1% of games were made up as either the day or night game of a day-night doubleheader
27.8% of games were made up as the second game of a regular doubleheader
01.7% of games were not made up

Interesting! So most games are made up as a part of a dh, but a fairly significant % are not.

Does the bold mean 2 games were made up on the same day?

And what exactly is the difference between a regular dh and a day-night dh? Are the games shortened with the first and 1 ticket gets you into both with very little time in between?

[Le Grande Orange]

Originally Posted by kq76

Does the bold mean 2 games were made up on the same day?

Yes.

Originally Posted by kq76

And what exactly is the difference between a regular dh and a day-night dh? Are the games shortened with the first and 1 ticket gets you into both with very little time in between?

Day-night doubleheaders are separate admission events. There's a day game with its own admission, and a night game, with a start time usually five to six hours after the day game's scheduled start, which has its own admission. Separate attendance is reported for both games.

Regular doubleheaders are the traditional two-for-one event, that is, one ticket is good for both games. The second game starts about a half-hour after the first game ends. One attendance figure is reported for both games, with the practice being the first game listed as zero and the second game having the attendance figure, so that the per game average attendance for the season works out correctly.

[Le Grande Orange]

The year-by-year numbers:

                Common Off Day
             --------------------
       Ppd     1 gm  2 gm  2 gm      DH    DH
       Gms           (D-N) (Reg)   (D-N) (Reg)    Ccd
------------------------------------------------------
2009    37      13     0     0       12    11      1
2010    21       3     0     0       12     6      0
2011    51      15     2     2       22     9      1
2012    21       3     0     0       11     7      0
2013    37      11     0     2       20     4      0
2014    35       7     0     0       18    10      0
2015    38       7     0     2       18    10      1
2016    26      10     0     0       11     3      2
2017    39      10     0     0       18    11      0
2018    55      16     2     2       19    15      1
2019    39       5     0     0       21    12      1
2021    75      13     0     0       27    34      1
------------------------------------------------------
       474     113     4     8      209   132      8

[gmo]

I had written up some notes way back when but found them again and realized I never put them here...

Interesting. I had never thought about the offday distribution in that way. It makes sense though now seeing it explained.

The one impression in this area that I had formed over time is that there are a relatively high proportion of offdays very early in the season, i.e., more an average early than across the season as a whole.

My first explanation for myself there was that it was to allow players to more ease into the season. Could also be to make it easier to more immediately make up games - if teams playing an April Mon-Wed series are both off that Thu, then a rainout (or snowout) could potentially be made up right away. Or it might not even be so much for makeups but just because April may be the worst weather month of the regular season so avoid scheduling as many games then.

To say nothing about the reasons I came up with, that impression may be wrong. I have done no analysis. It may be formed merely from the (maybe very new; again, no research) phenomenon of offdays within the very first season series of the season. E.g., March 31, the day after Opening Day this year had only 5 games, and that was a Friday which is hardly otherwise an MLB offday.

How might this affect me as a schedule creator?

First a mention that I am actually playing the game regularly with OOTP24 after having not played much at all in over a decade. I note that only to make the point that the capability of the game to have rainouts and reschedule is a new phenomenon to me. But I have only had that happen once for "my" team in almost 5 seasons because they have a home stadium with a roof. I have noticed a case or two of a game moved to after the originally schedule last day of the season. Anyway, a schedule should not be considered an immutable piece of art.

Will I adjust though? I can see leaning toward slightly skewing offdays more later in the season, but that will not be an easy change to make mentally.

When not requested to have everybody always all off together, I am so keyed into having offdays at least feel spread pretty evenly across the season, though with it being differently uneven among the teams. I like the aesthetics where on a given a day one team has played, e.g., 98 games, another 95, another 96, etc.

The typical current MLB-like 162 games has 14 full weeks pre-All-Star and 11 full weeks post-All-Star. Six 4-game series (and thus six no-offday weeks) fill out the schedule. Now I would probably aim to average 3 of those on each side of the All-Star Break. Since the first half is longer than the second, that means offdays are slightly more common in the first half. I can try to make a point to lean more toward having a median at 4 of those pre-ASG and 2 of them post-ASG, which would slightly lean the offdays to being more common in the second half.

[kq76]

Originally Posted by gmo

The one impression in this area that I had formed over time is that there are a relatively high proportion of offdays very early in the season, i.e., more an average early than across the season as a whole.

I had gotten that impression too, but I never paused to ask why. Or, to verify whether that impression was true.

Originally Posted by gmo

My first explanation for myself there was that it was to allow players to more ease into the season. Could also be to make it easier to more immediately make up games - if teams playing an April Mon-Wed series are both off that Thu, then a rainout (or snowout) could potentially be made up right away. Or it might not even be so much for makeups but just because April may be the worst weather month of the regular season so avoid scheduling as many games then.

Yeah, that's a good point. More scheduled offdays could very well be because of the less than ideal weather in April, and those other reasons.

Looking through OOTP's settings the other day, I was reminded of the early season expanded rosters. IIRC, that was introduced due to covid and only for 1 season. But it did make me wonder, should April have more offdays? And the thing that immediately came to mind was that April seems to have more players out due to injury than other months (not in shape coming off the off-season, coming back from injuries, playing hard to make a roster, who knows). I can't remember it exactly, but I seem to remember once watching a video essay (by Brian Kenny maybe?) breaking down the # of games missed by players due to injury by month. If anyone knows where we can find this data, I'd be really interested in looking at it from a schedule creator's viewpoint. Maybe there is a good reason for April to have more offdays.

Originally Posted by gmo

The typical current MLB-like 162 games has 14 full weeks pre-All-Star and 11 full weeks post-All-Star. Six 4-game series (and thus six no-offday weeks) fill out the schedule. Now I would probably aim to average 3 of those on each side of the All-Star Break. Since the first half is longer than the second, that means offdays are slightly more common in the first half. I can try to make a point to lean more toward having a median at 4 of those pre-ASG and 2 of them post-ASG, which would slightly lean the offdays to being more common in the second half.

I've been meaning to re-do my schedules with some 4-game series, but that's a big project (at least for me it is) and I was kind of stumped over the best way to go about it. And by your post I think you just helped me wrap my head around it! I'm not sure what exactly I was stumped by (how many to have, where to put them, what type of opponent to give them to, keep the Mon-Thurs and Fri-Sun windows or overlap them some), but reading your thoughts made me look at the Jays schedule (I typically like to come up with my own solutions in the hopes of maybe coming up with something better, but if something isn't coming to you it's probably best to look at how others do things) and I noticed they have had 1 against the Royals in April, 3 (!) in May (vs the Red Sox, Yankees, and Rays), 1 vs the Astros in June, none in July (no surprise, the ASB can be tricky to schedule around), 1 in August vs the Guardians, and only 1 in September vs the Rangers, starting today actually.

So, 3 intra-divisional 4-game matchups and 3 inter, no interleague. Also, 4 before the break, like you were thinking. That all sounds good. The key thing, for me at least, is to see that not once did they break from their Mon-Thurs and Fri-Sun windows. I also just checked the Yankees and they break those windows right off the bat with 2 of their 4-game series starting on Thursdays in April this year (it looks like 3, but 1 included a makeup game). Then another 1 of 2 in May, 0 of 1 in July (5 before the ASB), 1 of 2 in August, and 0 of 1 in September (4 of 8 start on a Thursday). So I was correct, MLB does break those windows, but I don't think I should feel like I should try to too. If keeping them unbroken is easier, just go with it. Same thing with the # of 4-game series, they don't all have the same #, but I don't think I necessarily need to aim for that level of detail. Simply integrating some set # of 4-game match-ups would be a realistic improvement. Thanks!

I'm sure you understand what I'm getting at, but if someone new to schedule creating is reading this and asks, "what's so difficult about breaking these windows?". Well, the problem is if the Yankees are starting their next series against the Orioles on a Thursday then you have to also make sure that the Orioles don't also end their previous series on a Thursday. And if you have these set windows (that early and late week series are played in), then it's pretty easy to slot and move matchups around. But if the series are split over the 2 windows, then it can get rather complicated fast.

[Le Grande Orange]

MLB regular season postponed/cancelled games by month for the 2009–2019 and 2021 seasons.

                     MLB Regular Season
             Postponements/Cancellations by Month

        Mar                                Sep
        Apr    May    Jun    Jul    Aug    Oct     Total 
----------------------------------------------------------  
2009     12      8      6      3      2      6        37
2010      2      7      3      1      3      5        21
2011     17     15      2      2     11      4        51
2012      8      2      2      0      2      7        21
2013     18      8      6      3      0      2        37
2014     18      5      4      2      3      3        35
2015      6      5      8      4      1     14        38
2016     11      7      1      2      2      3        26
2017     12     15      1      2      6      3        39
2018     28      7      4      5      2      9        55
2019     15      9      4      4      3      4        39
2021     22     15      8     17      7      6        75
----------------------------------------------------------
Total   169    103     49     45     42     66       474

       35.7%  21.7%  10.3%   9.5%   8.9%  13.9%    100.0%

57.4% of postponements over these seasons occurred in the first two months. 2018 had the most in Mar/April. 2021 had an usually wet July, while 2015 had a wet September.

The next question would be the number of games made up by month.

[kq76]

Interesting. Even taking the fact that the Mar/Apr and Sept/Oct include partial second months, it's still way more than you'd expect given the other averages.

As for "man-games lost" (that's the term I was trying to remember), I found a few studies.

One study says:

In conclusion, in the evaluation of MLB injuries resulting in time on the disabled list, there were a number of noteworthy findings. First, upper extremity injuries are predominant in pitchers while lower extremity injuries are more common in position players. ... Third, the highest injury rates were seen in the early months of the season and rates declined as the season progressed. ...

But, it does point out earlier that:

We elected not to include preseason exhibition games in our injury and exposure calculations, while clearly these injuries may result in regular season disabled list designation. In order to best control for this, we excluded players placed on the disabled list before and including the first day of the season; however, it is possible that some prevalent cases were included in our data that may have contributed in part to the higher observed incidence rate in April. Finally, players who are injured during the month of September may not be placed on the disabled list because of MLB rules permitting teams to expand their rosters to include 40 players. This may in part explain the much lower injury rate observed during the last month of the regular season in the present study.

IOW, the early and late season data may not be telling the whole truth.

There's also this study that specifically looks at hamstring injuries and it seems to agree that they're more often seen in the early part of the season. As the article says, "Hamstring strains are the most common injury for professional baseball players and can result in significant time on the disabled list.". Injury rate is the # of injuries per game so the lower the #, as seen in April and May, the more common.

And this study, about shoulder injuries, concurred. It also claimed the shoulder was, "the most vulnerable anatomic location".

The 511 injuries were spread along the months of March September, whereas October–February witnessed no injuries (Figure1). April was shown to be the month of the most injuries with 125 (24%) injuries, while June and May came in second and third with 89 (17%) and 81 (16%) injuries, respectively. Injuries that required surgery were highest during the months of March and April with 24 (50%) and 17 injuries (35%) respectively (Figure2). Moreover, we calculated the average number of days the athlete was placed on DL due to shoulder injury by month (Figure3). A decreasing trend is evident showing March and April to be the months with the highest number of days issued on the DL (97 days and 92 days respectively), whereas August and September showed the lowest number of days issued on the DL (26 days and 31 days respectively).

I'm sure there are more studies on the topic, but at this point I'm willing to just go with there are more injuries in April, whatever the reason, and so maybe it should have an extra off day or two.

[Le Grande Orange]

Originally Posted by kq76

Interesting. Even taking the fact that the Mar/Apr and Sept/Oct include partial second months, it's still way more than you'd expect given the other averages.

Those rates are for modern MLB, which has an overall postponement rate of under 2%, a combination of better park drainage, better grounds keeping, more willingness to let a game start if possible, and the presence of several roofed stadiums.

The early season postponements would certainly be higher in the classic 16-team MLB years, when the overall postponement rate often hit 10% or higher.

[kq76]

I just discovered something big I had no idea about. And LGO, I know you already know this already you posted a zip of it. From 1910-1956, teams could start the season (and of course end the season) at the reserve limit roster size (usually 40). And by start the season I mean anywhere from opening day to sometime in May or even June 15! It wasn't until 1957 that they said the max they could start the season at was the active roster and even then they raised it to 28 for 1957-1967.

Now I knew the roster limits had varied over time, but I thought for most of baseball history it was 25 then 40. I had no idea that for a good chunk it was actually 40 then 25 then back to 40.

I realize they probably made the change to save money, but I wonder if anyone was arguing in the decade or 2 after they made the change that they should go back to the old way starting the season with more players to maybe prevent injuries, or if that's only a relatively recent concern.

Reading all this got me to decide to enable the "expanded initial rosters" setting.

[Le Grande Orange]

Originally Posted by kq76

I just discovered something big I had no idea about. And LGO, I know you already know this already you posted a zip of it. From 1910-1956, teams could start the season (and of course end the season) at the reserve limit roster size (usually 40). And by start the season I mean anywhere from opening day to sometime in May or even June 15! It wasn't until 1957 that they said the max they could start the season at was the active roster and even then they raised it to 28 for 1957-1967.

Now I knew the roster limits had varied over time, but I thought for most of baseball history it was 25 then 40. I had no idea that for a good chunk it was actually 40 then 25 then back to 40.

Bookmark the following: League Operating Rules.

It'll come in handy.

Originally Posted by kq76

I realize they probably made the change to save money, but I wonder if anyone was arguing in the decade or 2 after they made the change that they should go back to the old way starting the season with more players to maybe prevent injuries, or if that's only a relatively recent concern.

If I recall correctly, part of the reason for switching to a 28-player active limit at the start of the season instead of the previous reserve limit (40, although teams typically carried closer to 32) was to help the minor leagues by reducing the number of players the major league clubs would take away from their minor league affiliates during that first month of the season. The minor league teams would thus be able to put a better team on the field, hopefully helping out the box office.