Injuries and Starting Players

[Charlie Hough]

Has anyone done large-scale analysis to compare OOTP's injury rates vs. real life for star players and starters vs. others on the roster? And has anyone compared injury rates vs. player usage?

At least with the teams that I'm controlling as GM, I rarely see depth starters and bench players suffer an injury. It's strongly weighted heavily toward starters, and when it comes to historical play, toward players who might have had only had 350 to 450 at bats but weren't necessarily injured in real life. I wonder whether OOTP is calculating injury ratings based on at bats and usage and assuming that lower usage should mean a higher injury rating.

Maybe a player only had 450 AB in a given season, but that may have been due to some platooning for part of the year or other managerial decisions. It doesn't mean necessarily that the player should have a higher injury rating. But it seems that starters and especially players like this get hurt far too often.

There are rashes of injuries for my teams every season, and generally I play historical games. So either the game is generating hidden injury ratings that shouldn't be there or something is going wrong with injuries being heavily weighted toward starters. At least that's how it seems to me.

I don't have any data for sharing at this point, but I'm hoping that someone has done a comparative analysis to see how OOTP stacks up against reality. And, again, I'm not talking about injuries overall. I am interested ONLY in injuries as they are divided between regulars vs. depth players, and injuries as they compare to real life injuries and player usage.

[rpriske]

Aren't starters more likely to be injured?

You know... since they play more?


But in your actual post you seem to say players are more likely to be injured if they played less... I'm not sure which side you are arguing.

[Charlie Hough]

Rpriske, you misunderstood my post.

First of all, what do the real life data say? Conjecture will get us nowhere. Does anyone have data on real life injury rates for starters vs. others on the roster? Sure, it stands to reason that starters should face a greater injury risk, but is that really the case? How many injuries are occurring on the field as opposed to during batting practice, fielding practice, etc.?

More importantly, I stated that injuries seem weighted heavily toward starters. In other words, unless real life data suggest otherwise, it seems like the injuries may be disproportionately weighted too heavily toward starters, even assuming a higher rate for regular players in real life.

Also, when I discussed injuries happening to those with 350 to 450 at bats, I was talking about players who had 350 to 450 at bats in real life. In OOTP, however, since trades and free agent signings can occur, those players can end up in situations where they are going to get 500+ at bats in OOTP.

But it seems like the game repeatedly punishes them when they're used that often, and it injures them again and again. Maybe it's just coincidence, but it seems to me that the game may be assigning these players hidden injury ratings based on their real life usage. So, if they weren't used for 500+ at bats in real life and only got 400, but they're used as every day starters in OOTP, they will be injured repeatedly.

However, their lower real life usage is not necessarily because they were injured. It could be due to platooning, or who was on the real life roster, or defensive considerations, etc. So if the game is assuming that they were injured and that's why they played less, this is the wrong way to go. It should make no assumptions and should use injury data like DL time wherever possible.

Here's a great example and the most recent one I can cite: in the final 30 days of my current season, there were 5 injuries to starting players on my team, rendering the top team in the league unable to field a proper lineup. A free agent third baseman had to be signed on Sept. 22 just to avoid having a massive defensive liability. There are now only 3 starters from the regular season who are healthy and able to play in the NLCS.

Someone please show me where anything like this has ever happened in the history of baseball. Show me a case of any MLB team that lost 5 of its 8 starters at the field positions to mult-week and multi-month injuries in just 30 days. If it's happened and can be documented, then I'd like to know how often this has occurred. Because exceptions prove the general rule, but, in OOTP, it seems like the general rule is that things like this can happen quite frequently. And unless I've missed something in the annals of baseball history, there is no way it has happened that often in real life.

In every historical sim I run, these rashes of injuries occur off and on all season with the OOTP injury setting at normal. No, they're not usually as bad as 5 starters out at the same time and heading into the NLCS, but it's commonplace for 2 or 3 starters to suffer major injuries within days of each other. And it seems like the injuries particularly target starting players whose real life usage was 100 to 150 AB lower than the other starters. So starters who saw 500+ at bats don't seem to get injured as often, and those with 350 to 450 at bats seem to suffer more injuries.

Moreover, depth starters and bench players seem like they are injured far less frequently. Some of that may make sense, but are they injured too little? Is the game imbalanced and injuring starters far more than it should? Is it distributing injuries at a ratio that doesn't match the real life split between starters and backups?

These are questions I'd like to address if anyone had real life vs. OOTP data.

[rpriske]

I did misunderstand you about the 450 plate appearance situation.

I DO think it is more likely that starting players are injured than bench players... but I don't have the data. Do you have counter data?


As for your question, has any team lost five starting players to injury in a 30-day period.

Well, probably, but I don't have that data either. I do note that the New York Yankees are currently without their starting 1B, 3B, SS and CF (or whichever outfield position Granderson would have been playing this season...)

[Charlie Hough]

Further to this, I'm taking a look at individual players. I have a strong suspicion that OOTP players are suffering more long-term injuries than individual players do in real life. In other words, players get hurt in real life and often multiple times, but I don't think OOTP is distributing injuries on an individual basis. The randomness factor is leading some players to get injured so many times that it defies their injury proneness ratings and skews their performances dramatically compared to real life.

For example, in my current game, which began in 1978 and just finished the 1980 season, Jack Clark was injured 8 times in 3 seasons.

He had 2 day-to-day injuries and 6 injuries that lasted weeks at a time, including 3 that lasted 5 weeks or more.

3 of his long-term injuries happened during a 90-day period toward the end of 1980. That seems a bit extreme.

Yet the editor claims he has an injury proneness of just 70 on a 200 scale. How is a player with 70 injury proneness suffering 8 injuries in 3 seasons with 6 of them knocking him out for long durations? Does OOTP pay any attention to the lengths and numbers of injuries per player to make sure that there is a proper distribution compared to real life tendencies?


Another player, Fred Lynn, has been injured 8 times in 3 seasons, including 5 injuries in 1980 alone. He has suffered from major injuries that include a partially torn labrum (6-7 weeks); a separated shoulder (2-3 weeks); and now a torn labrum that has sidelined him for 10 months.

Yet his overall injury proneness is 96 out of 200 and his arm injury proneness is the same 96 out of 200. So how is a player with average injury proneness suffering through all of this? And why do I see many players having the same problems despite having average or below average injury proneness?


Bobby Grich, a player with 69 injury proneness, has also been injured 8 times in 3 seasons. His injuries, however, have tended to be shorter, so half were day-to-day and the other half were longer term.

Interestingly, while his leg injury proneness is only 33 out of 200, 5 of his 8 injuries have been to his legs.


Obviously we're looking at limited sample sizes here, but I'm trying to understand why OOTP injuries occur, what effect the injury ratings have on them, and why some players seem to be suffering more injuries than others. Their injury proneness doesn't seem to provide an explanation in many cases.

Looking at various historical player profiles, I don't see a clear correlation between injury proneness and real life stats anyway. It looks like proneness is assigned randomly but that the distribution of injury events is also highly random.

It's looking like starter vs. bench status and possibly other factors are coming into play much more than injury proneness itself, which I think should be the biggest factor. Yes, being on the field should increase injury chances, but should it be a huge multiplier that renders injury proneness almost meaningless?

I'm seeing plenty of bench players with injury proneness over 100, but they aren't getting injured very often, and they don't seem to sustain some of the repeated, long-term injuries that I see from starters. Some of that should be expected from lower usage, but it looks suspiciously like usage is trumping proneness by a long shot.

Maybe I'm going down the wrong rabbit hole here, but constantly dealing with these rashes of injuries and seeing historical player deviate dramatically from their real life usage is frustrating and ruins a lot of the fun in OOTP. Yes, I could turn injuries to low or very low, but then that can skew the stats the other way.

To me, it seems like something is out of balance, and the way to fix it is not to push it out of balance in the other direction.

[thehef]

Originally Posted by Charlie Hough

Someone please show me where anything like this has ever happened in the history of baseball. Show me a case of any MLB team that lost 5 of its 8 starters at the field positions to mult-week and multi-month injuries in just 30 days.

A different sport - and I'm certainly not challenging your premise (I find it interesting) - but in the NBA the Lakers finished their very-abbreviated playoff "run" with their starting guards (Nash & Bryant), top two backup guards (Blake and Meeks), and starting small forward (Artest) all out due to injury...

On a pro basketball-sim forum some years ago, when gamers and the developer were wondering how to implement an injury rating using only common stats, I had suggested that some formula could be implemented that would look at games played and minutes played to determine that a player who played, say, 50 games and averaged 7 minutes per game was unlikely to have been affected by injury; rather, he was likely and end-of-bench scrub. A guy who played 50 games and averaged 32 minutes, however, was much more-likely to have been affected by injury (although holdout, suspension, etc., are certainly possibilities). From there, the idea was that most players would be assigned an injury rating of normal, whereas those who played (say) 78-82 a games were durable, and those who missed a lot of games but surpassed a certain threshold for minutes per game would be designated as injury-prone...

Whether something like this exists or not in OOTP (and I would be curious if it does) I don't know that it would be practical for baseball because days off for rest are common, especially if you are trying to nurse an aging start through a complete season.

[RchW]

Does OOTP claim accurate historical injury rates? The only thing I've seen and admittedly not looked further for is a claim that OOTP injury rates are based on modern day or current MLB injury rates. Even then I believe that OOTP classic setting is less than RL.

Is the injury history for each player in historical leagues based on historical fact or randomly assigned? If randomly assigned then you have no expectation that injury rates and length will be close.

Injury rates expressed in DL terms in historical leagues with the reserve clause are likely to be very low vs modern day baseball because players in those eras played through many injuries that in current days would have a player sit out.

I don't know the answer but thought these questions might help find it.

[SteveP]

I wish I could give you definitive answers, but until someone with better knowledge comes along (and there are a few of those), here's my

I believe injury proneness is assigned randomly, even for historical players, just as personality ratings are randomly assigned.

I have seen lots of threads on the subject of injury frequency, and it seems pretty well established that frequencies in OOTP match up well with known historical patterns, if you use the available settings in accord with the historical period you are playing.

I believe that when injuries occur in OOTP is a random event the probability of which, during any given play, is governed by the chosen frequency rates. Therefore, players who are playing more often are more likely to be injured.

I do not know the extent to which being injured increases the proneness to future injury. Probably somewhat, but Markus tends to set these kinds of things to very modest levels.

Injuries that occur during games always result in the injured player being removed immediately from the game, even when the injury is minor. This is unrealistic, since a substantial number of players play thru their injuries and then are unable to play the next day. Along with being unrealistic, this causes problems because the AI isn't very good about figuring how best to replace the injured player, causing bad defensive assignments, etc. This is why I continue to use a lower setting on injury frequency than would otherwise be desirable for realism.

[Charlie Hough]

Originally Posted by SteveP

I believe injury proneness is assigned randomly, even for historical players, just as personality ratings are randomly assigned.

I have seen lots of threads on the subject of injury frequency, and it seems pretty well established that frequencies in OOTP match up well with known historical patterns, if you use the available settings in accord with the historical period you are playing.

I believe that when injuries occur in OOTP is a random event the probability of which, during any given play, is governed by the chosen frequency rates. Therefore, players who are playing more often are more likely to be injured.

I agree with you on all counts and believe the same things. I think the critical issue is whether the injury frequencies and severity are realistically distributed per individual player.

I don't mean whether they match a player's real life injuries or player usage. I mean whether the total numbers of injuries per month, per season, and per career, along with their durations, match the real life frequencies and durations of injuries that typically happen for one player in MLB.

In other words, what is the average number of times that a player suffers an injury during an MLB career, adjusted for number of years in the leagues? What is the average length of those injuries? What is the mean of injury occurrences and durations? What are the highs and lows? What are the outliers and exceptions?

This is an important distinction. Just as OOTP uses league totals as a basis for game engine results and statistical output, it probably uses injury data in the same way. But is it focused enough on keeping individual injury output regulated in the same way that the game is tweaked and coded to keep individual statistical output in line with real life?

To me, when a player with an injury proneness of 70 out of 200 suffers 8 injuries, including 5 major injuries, in just 3 seasons, this is the equivalent of a player hitting 200 home runs in a 3-year period or a player with a 10 out of 20 power rating hitting 130 home runs in a 3-year period. It doesn't seem realistic or consistent with what we typically see from players, and while we might expect an occasional saved game with a player who hits 200 home runs in a 3-year period or a player with average power who goes wild for a few years and racks up big home run totals, we sure as hell would never expect 10 or 20 or 30 players to do these things. Yet I have numerous players in all of my saved games that pile up injuries like this.

So, I want to reiterate that I don't have the real life data to know whether I'm right or wrong in my suspicions. But I suspect that something is amiss and that individual players are suffering injuries too frequently and with too much severity compared to real life. In the end, the total number of injuries and the total length of injuries might be okay for league wide totals, but I suspect that the individual distribution of those totals is too often skewed compared to real life. And that means that some players are probably not being injured often enough or severely enough as well. I suspect that there is a significant imbalance in the game.

[Questdog]

Here's my cent and a half:

I think the whole concept of injury proneness as a rating is just wrong. Every player should have the same chance at getting injured, period. The only case where a rating would be desirable is when trying to recreate actual results; i.e. historical replays, yet we do not utilize the ratings for this......

I don't think there are players who are injury prone in real life (except for a very few cases, like Dave Dravecky trying to pitch when he had cancer or any idiot who uses a head-first slide). I think players seem prone sometimes since the mind likes to find patterns in events and when a player suffers two or three injuries, he is marked in the mind as prone. But in reality, it is probably just random chance being unkind to some (which we would still see, if all players were given the same rating in OOTP).

[VanillaGorilla]

From the v 14 Manual:

"Player Injuries Generally

Injuries occur randomly in OOTP, both during games and between games. Players in OOTP have a hidden injury proneness rating that affects how likely they are to get injured. Injuries in OOTP can be illnesses, or they can be related to one of three different areas of the body: arms, legs, and back. Players who experience injuries to certain areas of the body may become more prone to future injuries in that area. For example, a player who suffers a major knee injury will be much more likely to experience future knee or leg problems than a player who has never been injured before. Having a good team trainer can be a huge advantage in dealing with injuries. Commissioners can also edit the injury proneness of players using the Player Editor."

To the OP's point re a 70/200 injury prone player receiving multiple injuries over a three year span:

Once the 70 player is injured, he is more likely to be injured again than a 70 player who has never been injured is to be injured in the first place.

What you are observing is happening, but it is not thorugh fluke randomness which would give the 10/20 power guy 200 HRs over those three seasons.

*All 70 injury rated players may be created equal, and their chances for initial injury may be equal, but after that they are no longer equal. Their chances of being injured subsequently are different due to differences in trainer ratings. The 70/200 is just one part of the equation.*

Whether or not this is an accurate representation of reality, or even an adequate approximation for a simulation is an argument I am not making one way or the other.

also from v14 Manual:

"Day-to-day injuries don't prevent players from missing games, but they reduce a player's effectiveness by a certain percentage. Additionally, a player who plays with a day-to-day injury is at greater risk for additional injury, particularly in that area of the body. For example, a player playing on a sprained ankle is more likely to suffer a foot injury. Day-to-day injuries are represented on player lists with a yellow cross icon and by listing the player's name in orange."

Again, another factor that plays into the end result of a player being injured.

Players of identical injury rating have a dissimilar probability of being injured because injury rating is not the sole determinant of the probability of injury.

[VanillaGorilla]

Originally Posted by Charlie Hough

It's strongly weighted heavily toward starters, and when it comes to historical play, toward players who might have had only had 350 to 450 at bats but weren't necessarily injured in real life. I wonder whether OOTP is calculating injury ratings based on at bats and usage and assuming that lower usage should mean a higher injury rating.

Doing a quick and dirty sampling gave me an AB/Season median of approx 300.*

If injury prone scores are distributed randomly, the fact that you see, through casual observation, more players with < 450 RL ABs get injured, is expected. There are simply more of these players than players who had > 450 ABs in a given season.

[Charlie Hough]

Originally Posted by VanillaGorilla

Once the 70 player is injured, he is more likely to be injured again than a 70 player who has never been injured is to be injured in the first place.

If this is the case, where is OOTP storing this post-injury information? I don't think the injury proneness ratings are changing after injuries, but I will have to do some testing and check.

I wouldn't expect a player who has been injured 8 times with 5 major injuries in 3 seasons to have a proneness rating of 70 after the injuries. I suppose it's possible that he started with a very low rating and it increased dramatically to 70, but that still doesn't seem to match what you'd expect for a player who has suffered that many injuries. Wouldn't you think a player who suffered that many injuries in a short span of time should have a rating well over 100? The 70 after all those injuries implies that he is well below average in proneness, and that makes no sense.

I suspect that OOTP is storing additional injury information and calculating the increased likelihood of injury. Are there further injury ratings that are hidden and that change once injuries have occurred?

[SunDevil]

Originally Posted by Charlie Hough

If this is the case, where is OOTP storing this post-injury information? I don't think the injury proneness ratings are changing after injuries, but I will have to do some testing and check.

I wouldn't expect a player who has been injured 8 times with 5 major injuries in 3 seasons to have a proneness rating of 70 after the injuries. I suppose it's possible that he started with a very low rating and it increased dramatically to 70, but that still doesn't seem to match what you'd expect for a player who has suffered that many injuries. Wouldn't you think a player who suffered that many injuries in a short span of time should have a rating well over 100? The 70 after all those injuries implies that he is well below average in proneness, and that makes no sense.

I suspect that OOTP is storing additional injury information and calculating the increased likelihood of injury. Are there further injury ratings that are hidden and that change once injuries have occurred?

Based on the information in this thread and in the manual, the answer would have to be yes.

[rpriske]

Originally Posted by Questdog

Here's my cent and a half:

I think the whole concept of injury proneness as a rating is just wrong. Every player should have the same chance at getting injured, period. The only case where a rating would be desirable is when trying to recreate actual results; i.e. historical replays, yet we do not utilize the ratings for this......

I don't think there are players who are injury prone in real life

I don't agree with this. There ARE players who are more injury prone than others (Nick Johnson springs to mind).

That doesn't mean the term is used properly, however. I remember a number of years back reading a pre-season article that referred to Roy Halladay as being injury prone because he missed time in each of the previous two seasons. One of those DL stints was due to appendicitis. The other was a line-drive off the shin. Neither of these two things should have ANY bearing on his being injury prone.

[VanillaGorilla]

Originally Posted by Charlie Hough

If this is the case, where is OOTP storing this post-injury information? I don't think the injury proneness ratings are changing after injuries, but I will have to do some testing and check.

I wouldn't expect a player who has been injured 8 times with 5 major injuries in 3 seasons to have a proneness rating of 70 after the injuries. I suppose it's possible that he started with a very low rating and it increased dramatically to 70, but that still doesn't seem to match what you'd expect for a player who has suffered that many injuries. Wouldn't you think a player who suffered that many injuries in a short span of time should have a rating well over 100? The 70 after all those injuries implies that he is well below average in proneness, and that makes no sense.

I suspect that OOTP is storing additional injury information and calculating the increased likelihood of injury. Are there further injury ratings that are hidden and that change once injuries have occurred?

Take the example of a player with a day-to-day injury who has his output reduced by 14% for 6 days. We do not see his contact/power/eye ratings reduced by 14%, but his performance is reduced. The basis of his output (pwr/con/eye) remains the same, but an additional factor that reduces the chance of him having a successful PA is introduced into the equation.

In the same way, a player with a D-T-D injury is more likely to be injured if he plays, but his injury proneness number remains the same. A 70 player with a D-T-D injury is not the same as a 70 player without a D-T-D injury, yet the 70 player with the D-T-D rating had a 70 before his D-T-D injury occurred and stays a 70 rating during the 6 days of his injury and is a 70 player once the D-T-D injury is over.

Con/pwr/eye is the basis for a hitters probability of success. However, what he brings to the plate (sorry about that) is more than simply those three ratings.

The proneness to injury rating is the basis used for a player to receive an injury, but there are other factors involved. Based on the manual's documentation, we expect to see players who get injured accumulate more injuries than players who avoid them while the injury ratings for each group of players (previously injured vs previously uninjured) remains the same.

[Charlie Hough]

If OOTP is storing additional, hidden injury ratings and using them to calculate future injuries to players, that is a HUGE problem for historical games and particularly for users who might want to use OOTP's 3-year recalc or 1-year replay mode.

If a single injury in OOTP can dramatically increase a player's future chances of being injured, and that rating is not controlled or adjusted by historical recalculation, then that's a showstopper for replay-style simmers.

That means historical players will often deviate significantly from their real life career arcs or performances based on injuries alone. And while I'm okay with random injuries on occasion, if a player like Jack Clark is doing to end up with HUNDREDS fewer at bats during a 3-year plan because of this injury effect, then we've got a problem. Over the course of an OOTP career, that adds up to potentially hundreds more missed at bats.

Based on the discussion here, I would expect Clark to continue to sustain major injuries throughout the rest of his OOTP historical career, even if I was using 1-year recalc of ratings. So, in all likelihood, Clark's performance will be far removed from his real life performance. The same can be applied to the other players mentioned here.

That is not a desirable outcome for those who want to play with development turned off and recalc turned on. If you're looking for a historical gameplay experience with players who generally follow their real life career arcs with occasional and reasonable deviations, allowing injuries to run wild is not the way to do it.

[Lukas Berger]

Originally Posted by rpriske

I don't agree with this. There ARE players who are more injury prone than others (Nick Johnson springs to mind).

Also, one of the players whose in game injuries are objected to above, Fred Lynn.

[VanillaGorilla]

Originally Posted by Charlie Hough

If OOTP is storing additional, hidden injury ratings and using them to calculate future injuries to players, that is a HUGE problem for historical games and particularly for users who might want to use OOTP's 3-year recalc or 1-year replay mode.

If a single injury in OOTP can dramatically increase a player's future chances of being injured, and that rating is not controlled or adjusted by historical recalculation, then that's a showstopper for replay-style simmers.

That means historical players will often deviate significantly from their real life career arcs or performances based on injuries alone. And while I'm okay with random injuries on occasion, if a player like Jack Clark is doing to end up with HUNDREDS fewer at bats during a 3-year plan because of this injury effect, then we've got a problem. Over the course of an OOTP career, that adds up to potentially hundreds more missed at bats.

Based on the discussion here, I would expect Clark to continue to sustain major injuries throughout the rest of his OOTP historical career, even if I was using 1-year recalc of ratings. So, in all likelihood, Clark's performance will be far removed from his real life performance. The same can be applied to the other players mentioned here.

That is not a desirable outcome for those who want to play with development turned off and recalc turned on. If you're looking for a historical gameplay experience with players who generally follow their real life career arcs with occasional and reasonable deviations, allowing injuries to run wild is not the way to do it.

If injury ratings, like personality ratings, are generated randomly, as SteveP offered, then I think that is where your deviation occurs. Jack Clark could be created as an Iron Man and hit 700 HRs, or he could be brittle and hit 70.

In my current random debut league, I have injuries set to "Very High" and positional fatigue set to "High." I am in August and there are 206 players on the league list for a 30 team league which began in 2004. The career HR leader is still active and has 670 HRs, having entered the league in 2008 and the modifiers from 2013 onward being those for 2013.

In my two complete run throughs testing and tweeking my HOF algorithm in v 13, I had players that never had 450 ABs in a single season reach the HOF. Bill Delancey reached the HOF in both run throughs (yeah, I had to look him up, too).

In both complete run throughs, I had about 50 RL HOFers reach the HOF each time (some RL HOF players were generated as older players at the start of the league and some were generated at the very end of the run, so they had no chance of producing HOF numbers). These HOF entrants are based on my dynamic entry standards, not OOTP defaults.

I had Joe Morgan play 27 years in a "high" injury environment. He had a LOT of injuries, and did not reach 3000 hits, but his season to season % stats and career % stats were very good.

I know that my league set up is different from what you are wanting to do with your's, but I am simply offering these bits to the discussion.

How the injury model should be is a different discussion from what it is. I am not arguing for the status quo. I am not arguing for a change.

[SteveP]

Originally Posted by Charlie Hough

with occasional and reasonable deviations

This could be the heart of the problem. It's clear that among the people who play historicals, there is quite a wide divergent of opinion about what constitutes a satisfactory game, especially how much in the way of "deviations" they can tolerate (and which deviations they are willing to tolerate). I recall having to lobby pretty hard for adjustments that would allow historical SPs to get realistic CG rates, and had to do battle with people who could not tolerate the notion of SPs getting more IPouts than they should. Markus ended up splitting the baby. But that was a case where the statistical data and analysis were pretty easy to assess.

Injuries have been argued over and over again, with various adjustments made from version to version. Doesn't mean that Markus would refuse to do more, but it's worth knowing that it's already heavily trod territory.