OOTP Injury Study

[RonCo]

This is the result of a detailed study of the performance of the OOTP injury algorithm (discussed in a different thread some time back). It attempts to answer the question of whether the injury ratings make a difference, and, if so, define how big of an impact these ratings carry. The study is made with OOTP 5 data. I believe the injury algorithm has been left intact in v6 and beyond, but treat the study with kid gloves in that sense.

If you're a "Fog of War" kind of guy, and don't want an insight into how the game works, then this report may not be for you.

Of course, maybe I'm all wrong ... feel free to ignore as you wanna.

Since the document is too big to meet the site's 100K limit, I've put it on another site...I'll leave it here for awhile.

http://www.typosphere.com/Injury.doc

[AirDuster]

FYI,

The reason injuries go down as the season goes on is likely that more Prone players get injured early on and are more likely to be replaced by Normal or Durable players as the season progresses, whereas basically no Durable players are getting replaced.

Basically, as a season progresses, the proportion of durable:normal+prone increases.

[sfgiants88]

Very thorough study. Good job

I know you weren't attempting to cover this, but IMO, the main problem with injuries is that there are far too few of them. http://www.ootpdevelopments.com/boar...d.php?t=101183

[RonCo]

Originally Posted by sfgiants88

Very thorough study. Good job

I know you weren't attempting to cover this, but IMO, the main problem with injuries is that there are far too few of them. http://www.ootpdevelopments.com/boar...d.php?t=101183

I don't think you can compare major league DL usage with OOTP injuries. It's an attractive measure, but I don't think OOTP owners use the DL as often as real teams do. My opinion, of course, is subjective.

[RonCo]

Originally Posted by AirDuster

FYI,

The reason injuries go down as the season goes on is likely that more Prone players get injured early on and are more likely to be replaced by Normal or Durable players as the season progresses, whereas basically no Durable players are getting replaced.

Basically, as a season progresses, the proportion of durable:normal+prone increases.

I'll see if I can come up with an easy way to confirm or deny this hypothesis. My initial reaction is that this would only be true if the number of season-ending injuries were much higher than it is. But you could be right.

[sfgiants88]

Originally Posted by RonCo

I don't think you can compare major league DL usage with OOTP injuries. It's an attractive measure, but I don't think OOTP owners use the DL as often as real teams do. My opinion, of course, is subjective.

To be fair, the study I used was actual DL usage by computer run teams, not human teams.

Even with computer run teams, I can't imagine the difference would be that big. If a player gets an injury expected to last >15 days, he goes on the DL, whether it is OOTP or in real life.

[RonCo]

Real life use of the DL has exploded inthe past few decades. I think there are a lot of players in real life with injuries of severity less than 15 days that go on the DL. This would make comparison between OOTP and MLB very difficult.

[RonCo]

Originally Posted by AirDuster

FYI,

The reason injuries go down as the season goes on is likely that more Prone players get injured early on and are more likely to be replaced by Normal or Durable players as the season progresses, whereas basically no Durable players are getting replaced.

Basically, as a season progresses, the proportion of durable:normal+prone increases.

I attacked this from the other direction, measuring the number of active player-sims per sim over the 6-year span. I figured that if prone-normal players were coming out of the player pool at a rate great enough to cause the fall-off of injuries in the last half of the year (as you suspect) then I would see it as a reduction of players active in each successive sim.

What I found suggests that this theory does not account for the decrease in injuries in the last half of the year:



While there are fluctuations, it appears that as the season progresses players get injured and heal at about even paces.

Until I come up with a different theory, I'm going to assume the drop off is the result of something in the game engine itself.

[AirDuster]

What do you mean by "active" players? When a guy gets injured, he often goes on the DL and is replaced. If a prone guy is injured and is replaced by a normal/durable guy, it will still be one active player. I would think you'd want to look at the ratio of prone:durable and prone:normal on opening day and then again at the beginning of each month or something.

[RonCo]

Originally Posted by AirDuster

What do you mean by "active" players? When a guy gets injured, he often goes on the DL and is replaced. If a prone guy is injured and is replaced by a normal/durable guy, it will still be one active player. I would think you'd want to look at the ratio of prone:durable and prone:normal on opening day and then again at the beginning of each month or something.

Good question. "Active" means "non-injured." Hence, it measures the population of players each sim who are eligble to be injured. You can see the slight dip early in the year as essentially all players begin healthy but then get injured. From that point on everything is pretty stable, indicating that as players get injured, others are healing. In fact, in the graphic posted above, you can see a sllight rise in the availability of healthy players at the end of the year, which tells me that some players are healing but that injuries are not occuring to those who were healthy in the first place.

That said, here's a plot of the metric you're interested in: Prone/Durable and Normal/Durable players active (non-injured) each sim.

http://www.typosphere.com/graphics/injury_ratios.png

It seems very clear that changes in the ratio of players of various injury ratings is not the root cause of the decline in injury rates I've seen over the span of a season.

[zonk84]

Originally Posted by RonCo

I don't think you can compare major league DL usage with OOTP injuries. It's an attractive measure, but I don't think OOTP owners use the DL as often as real teams do. My opinion, of course, is subjective.

I think this is spot-on... I know I don't usually DL players unless the injury is 3 weeks or more - this is normally because my 40 man is usually pretty packed, so unless it's a key position (C, for example) where I can't play a man short for a couple weeks, I let the injured player stay active.

The big issue is the inability to DL players "retroactively". I mix between playing games out and simming a week at a time - and sometimes, miss injuries. Consequently, I sometimes end up with hurt players that have already burned a week's worth of injury.

I don't recall this ever being brought up in 2006 discussions -- but being able to retroactively DL a player (as well as transfer from the 15 to 60 day DL -- which I think HAS been discussed) would be a nice option to have.

[disposableheros]

Originally Posted by zonk84

I think this is spot-on... I know I don't usually DL players unless the injury is 3 weeks or more - this is normally because my 40 man is usually pretty packed, so unless it's a key position (C, for example) where I can't play a man short for a couple weeks, I let the injured player stay active.

The big issue is the inability to DL players "retroactively". I mix between playing games out and simming a week at a time - and sometimes, miss injuries. Consequently, I sometimes end up with hurt players that have already burned a week's worth of injury.

I don't recall this ever being brought up in 2006 discussions -- but being able to retroactively DL a player (as well as transfer from the 15 to 60 day DL -- which I think HAS been discussed) would be a nice option to have.

it has been, but as of yet, no word on whether it will ever happen or mot.

[sfgiants88]

Originally Posted by zonk84

I think this is spot-on... I know I don't usually DL players unless the injury is 3 weeks or more - this is normally because my 40 man is usually pretty packed, so unless it's a key position (C, for example) where I can't play a man short for a couple weeks, I let the injured player stay active.

Right, but the study I did was on computer team usage, and the AI doesn't have the foresight or intuition to be thinking about 40-man roster usage.

Regardless, I think the difference is pretty negligible, not nearly enough to make up for the gap between RL and OOTP.