My Steam Review Based on 1000+ Hours (at least)

[BTBMI]

I'm posting this feedback, with hopes that we can have a reasonable discussion on some of the limitations of this game. I am a very prolific player, and have played over a thousand games manually, as well as endless historical and fictional sims.

There is one huge aspect of the game that frustrates me. It's so realistic that it can be too random for its own good.

I understand this is just one player's feedback, and I'm not here to stomp my feet and to try to burn the game down. I love the game. I just wish there was a way to remove some of the most frustrating issues that it has.

It's not just that historical replays are never truly satisfying to me. Fictional scenarios almost always end up in some sort of a statistical mess, where home runs are way too low for a long period of time, more likely where they're way too high. Not only that but you often get weird things that break, or are so random they are annoying.

I just can't play the game anymore knowing that the time I invest the game will eventually be broken by something so hideously random, or so far away from reality, that it steals the life and fun out of the game

Anyway here's my Steam review.

"I've been playing Out of the Park since version 21. I've simmed over 75 complete historical and fictional scenarios. I've also manually played about 1,000 games. Taking a number of teams from game one to game 162 through a complete season.

At first I absolutely loved the flexibility that Out of the Park provided. Being able to create any world possible was like a dream come true.

This all sounds glorious but there is a downside. The more you play the more you realize that the game isn't very accurate. This might be hard to understand, but there's so much realism factored in that it makes the outcomes unrealistic at times. When I say "at times" I'm being polite.

Yes, I understand baseball Is random. Injuries are random, player development is random. we get all that. Out of the Park does a good job of understanding all this randomness.

Where it utterly fails is allowing you to play historical baseball and create an outcome that looks anything like reality. It's so realistic that it is at its core becomes utterly random, and this is a huge downside when you replay any historical scenario.

For example, my last replay I started in 1960 and I was going to sim up until 1979. I logged on to find out that Nolan Ryan was a mediocre reliever. This epitomizes everything that's wrong with Out of the Park. It's so realistic ai it's understanding of the randomness of baseball, that it turns baseball into a random mess that just isn't fun when you do anything historical.

You will never get Babe Ruth to be Babe Ruth. you'll never get Nolan Ryan to be Nolan Ryan. Why? Because that's the way baseball is. Except when you play historical baseball you want Babe Ruth to be Babe Ruth and you want Nolan Ryan to be Nolan Ryan.

Also, it's almost impossible to play extended historical or fictional simulation without home runs inflating ridiculously above 60. It gets really annoying when you're doing a simulation and all of a sudden players are hitting 79 or 83 home runs and you just wasted 40 hours of your life.

Sure you can go in and modify the math and do all the modifiers, but does it really need to be that complicated? Smart people should be able to create a smarter game.

There might be some pushback against this review, but it's not disingenuous. The main and damaging flaw with Out of the Park is that it's so realistic at its understanding of baseball that it creates unrealistic outcomes unless you're playing a fictional league where you don't expect any specific outcome.

Someone will likely say I don't understand the nature of baseball. I certainly do. I would say to them, they don't understand how fun it is to play a baseball game and actually have Babe Ruth be Babe Ruth."

[pstrickert]

When you set up your leagues, do you choose REPLAY mode or CAREER mode? From season to season, player usage and statistical results track closer to real life in REPLAY mode. CAREER mode allows for greater variance. (like what you're reporting). REPLAY mode works perfectly fine when simulating baseball history or a player's career. I hadn't thought about it before, but users may think choosing CAREER is better when simming more than a single season. Not true. Not when you want individual results to track closer to real life.

[Reed]

For your historical season, single season replay mode is probably best for statistical season accuracy with real lineups/transactions, etc..
Also do you play 1 year, 3 year or 5 year recalc, injuries off, real or neutralized stats, park factors or neutral park factors, development off, RAH and missed seasons on, etc..
Really hard to discuss (agree or disagree) your experience without knowing a lot more about your settings. So many settings will have an impact on your output.

Fictional, I never play fictional but if you don’t like the stat output can’t you just use a year like 1985 or whatever for your LTMs.

Pitchers. I play mostly random and I admit I sometimes scratch my head with who the game uses as a starter or reliever.

[LansdowneSt]

The Nolan Ryan "issue" is settings. You have settings that reflect him as he is for that time period. In Career mode, he may never become a starter depending on when you load him and if it is in the minors or majors. In a Default mode, his stamina and ratings will conform to the settings you choose (I'm a three-year stamina and recalc guy) as the years go by. Replay reflects defaults much stricter than any of that to stay closer to history. I doubt any out-of-the-box settings are perfect for everyone and to meet every expected assumption or preferred way to play. It's a sandbox, take the time to review the settings in the game and ask questions, as needed. Then you can use or change your settings as your personal preferences determine.

[uruguru]

If you want historical accuracy, you need to use recalc.

But if you do that then you, as the player, have a MASSIVE advantage over the AI because it does not know how players are going to develop. That means you need to enjoy wiping out the AI with your perfectly timed All-Star rosters, or implement major restrictions to trading and drafting.

1000 hours played is rookie level!

[NoOne]

Fictional mode - statistical mess is avoidable. I've been making ficitonal leagues that remain consistent for hundreds of years since 2012 or 2013. so, it's definitely possible.

There's af ew things that can help here. If you start with MLB player,s you will always have some transition over 25-30 years to the new distribution of fictionally created players for your draft. They will represent a different curve of talent distribution and it's far better one, for that matter.

Even if you start with fictional players, those "seed" players are not quite the same dsitribution as what will result from your drafts. Also, note that if you fiddle with number of players created (think default for 30t is 28rounds and 36 rounds of players created?) this too may impact overall talent levels in leagues.

Once there is an equilibrium with your draft created players filling the league, statistical results will remain fairly consistent. There will be ebb and flow just from random chance as well as rise and fall of talent in your league over time. It's greater for some stats than others. e.g. SB attempts might vary 15-20% but BA might be a +/- 5-10% ... i'm not sure how it plays out in the more recent versions of ootp.

What i suggest is to zoom out 30 years before you start playing. this makes the league and MiL filled with players created for your draft. Erase all league history. Purge the database. Now you are starting as if it is Day 0 of your sim. (pre-date your league, but set 'historical year' to 2024 or later to start, so you don't auto-import setting changes.

Now, as long as number of teams doesn't change or creation modifiers aren't edited or number of rounds/created players for draft, the talent will ebb and flow around a specific median.

At this point an "auto-calc" with appropriate League Totals in Stats and AI settings will result in a fairly consitent league average.

I take this a step further. I zoom out 5-10 years at a time, initially, to set fielding errors and the stuff that doesn't cause shifts to other stats - e.g. SH/SF/errors and such don't really impact the resulting averages of other things, but having htem where i want is important to the first long-term simulation givingme better information.

Once some of the simple stuff is well-callibrated, i run a sim 50-100 years. I'll start tweaking SO/BB to get them in line and Batting average where it belings. After that i work on SB and ratio of the hits... You add to 3b, it takes away from hr/2b/hits etc... you reduce hits, it'll proportionally reduce 1b/2b/3b/hr, if nothing else is changed. If you do it in the right order, you play less whack-a-mole with odd results.

I have a spreadhseet that does the hard work. It automatically uses the data from the exported league_stats (in almanace can make yearly "dumps" of batting, fielding and pitching stats). I adjust work table to fit how many years there are, but otherwise automated process. I can see a 50-year average all the various stats that directly tie to the League Total Modifiers.

This makes any ebb and flow that occurs 100% caused by ebb and flow of talent turnover inthe league as opposed to changing the modifiers each year. Seems more logical to me.

Anyway, you don't have to be that anal. A simple auto-calc and some tweaking to avoid 'extremes' can get you to roughly the same point.

If you see HR climb too high, reduce the HR modifier a bit.

I'm not a fan of yearly auto-calc, but that will make the statistical results very consistent if you use them every year. Probably too consistent.

"Nolan Ryan was a mediocre reliever. This epitomizes everything that's wrong with Out of the Park" -- this is wrong. First, you have a choice how this plays out. If you want nolan ryan to be similar to what he did in real life, you need to use the approriate settings. You can't have this without taking away randomness of development for the entire league in this regard.

so, the fact he ended up that way was a matter of your settings chocies. again, you don't get it both ways. You want a model and simulation, or a reply of history with limited randomness except for how the PA play out? you'll still have different results but the player talent will more closely tie in with historical talent from real life.

-----------

the historical stuff is unavoidable. there are options to make it more rrandom, which makes strategy and devleopment more important, or there are options to make it more siilar to history, which means you can simply look uip when players had great years andmake sure you hav ethem on your team for that year or range of years, lol.. kinda takes the guesswork out of it, but the good thing is that you have choices on this matter.

The problems here are ineherent with what is being done and unavoidable. Pick your poison -- a new alternative reality where a player may become a drunken bum and never realize the potential they did in real life, or the game ensures they take the same path.. you don't get it both ways.

I do agree that with historical leaguegs that enter new eras inherently have some issues. That statistical stuff is more about distribution if you use your auto-calc and other settings properly. You may see league average home runs wher you want them, but distirbution may not be the same. Do you care more about individual results or the leauge averages? Again, this is one of those either-or choices to some extent, but less drastic than historical player talent settings.

----------

some other stuff i think it is a lack of udnerstanding Law of Independent results and other things you'd learn in a more advanced college-level statistics class.

Also that playing games and outcomes don't cause a player to get better. The improvement or loss of ability are the result of the things they do off the field and their genetics.

Results in games are not causal to anything.. they are merely outcomes. they are the effect, not the cause.

---------------

I do wish they let the player talent would dictate the ebb and flow of stats more. The game relies on the modifiers too much to create differnt eras. The ratings of the players produced should create different eras (back to the cause and effect thing). But, that's not how it works, so era transitions are messy. However, due to various ways in which the RLleague evolved, it's also the only way to do it.

One thing to consider why this doens't match well with history as good as it sounds... Sometimes massive shifts were due to equipment changing and not so much talent distribution. you can't model that without doing it the way they do it right now.

All of a suddent he mound is 6" shorter and the ball is tighter one year. The same "power" ratings all of a sudden equates to drastically more HR. And, it should for these exact reasons.

Sometimes the historical shifts in statistics are due to an influx of talent, but more often it's due to equipment changes. Equipment changes are not represented by the player ratings.

So, it's one of those damned if you do, damned if you don't. This is why i stick to fictional leagues and i don't create "new" eras like that. I do get my leauge average somehwere near modern "normal" but i focus more on individual results.

e.g. i want 1-2 guys reaching 60+ hr per century... if it's 2-3 or 3-4, i won't worry too much about that, either. I don't want 50 guys getting to that level, though. This is how i choose my hr/year average and appriate modifier to make it happen. (this is usually +/- 10% if no other settings are changed and only player telent shifts -- also for a 30t 162g league.. size of league will impact volatility seen... i don't use numbers from 15 team subleague, i use the full 30 teams because everyone has a DH in my leagues)

I want only the very best stealers to get 100+ SB. Only a guy with >MAX SB and speed can do it. This is how i tweak my SB att and SB % sucess modifiers.

so, this usually results iia league with slightly higher offense. This helps distribution of that stats too based on talent. I see my higher end guys have greater value over median than when the offense is lower... and that's for pitchers too. there's greater stratification among the various levels of talent then with smaller numbers in a less offensive league.
BA .......... OBP........Slg..........ERA........Whip.......Fld pct
.2539.......3168........4087.....4.115......1.293. ......98210
.2632.......3274.......4226.....4.421......1.359.. ......98360

Unchaging modifiers or totals. This will remain consitent in future, too because of how i set it up. Avoiding seed players, avoiding RL players. RL player talent distribution is so biased it's incredubly top-heavy talent distribution curve. too much personal bias impacts the ratings of existing players one way or another. Talent curve should be skinny at the top, so it's jus ta bunch of lying to appease "homers" -- which is also why stat distribution in historic leagues is tough to nail down... the talent curve doesn't match the probabilities properly. Too many have higher ratings than they deserve.

One note on that -- they've made huge strides in talent curve for the fictionally created players. Love it. Much easier to keep leauge totals in line while still having the individual results i want. This means the rate at which hr are accrued relative to rating is well-callibrated. It also helps the elite guys take a larger portion of the home runs than having a tone of overrated power guys sharing the counting numbers. Average is utter trash (it is below median due to shape of curve). Median isn't even an everyday player. People don't realize just how terrible an average-to-median player is.. more than half the leauge is horrible, and that reflects RL better.

There's my min/max values of a 22 year sim i used to verify i had things wher ei wanted. This is 2138-2159. There's been no major shift in my stats after i got rid of seed players by 2050, give or take. The only changes are what i did to my modifiers to tweak things.

They've done soemethig with error distribution this year, i can tell. MY fld pct needs to be higher. but i can't get the SS's and such to have a typical number of 'max' errors. oh well... i'll suck it up and move on.

The point is you do your best. These aren't so much limitations as the fact that it's a complicated thing to model for all the various forms a league can come in. One-size-fits-all doesn't work.

I think most of what you don't like you can fix.. and a few things you'll just have eto accept as good enough, and more cases than not the deviation from RL is mostly arbitrary and in the eye of the beholder rather than backed by any evidence.

[mytreds]

Great review. Thanks for your honesty.

[Garlon]

There are no issues with historical players performing as they should if using recalc and retire according to history. If you turn off those options, then you are basically playing a fictional universe with real player names where the players will not necessarily develop as you want.

[Brad K]

Originally Posted by Garlon

There are no issues with historical players performing as they should if using recalc and retire according to history. If you turn off those options, then you are basically playing a fictional universe with real player names where the players will not necessarily develop as you want.

Add to that TCR set to none.

But then you have the randomness which in my observation might be in the range of +/- 30 or 40 points on BA.

But my question is when we have players playing in a non historical environment why do we expect performance within a few percent of actual? The standard should be that a player have a plausible career if in OOTP he plays in a hitters park instead of his historical pitchers park.

[Brad K]

Originally Posted by uruguru

If you want historical accuracy, you need to use recalc.

But if you do that then you, as the player, have a MASSIVE advantage over the AI because it does not know how players are going to develop. That means you need to enjoy wiping out the AI with your perfectly timed All-Star rosters, or implement major restrictions to trading and drafting.

1000 hours played is rookie level!

TCR is the equalizer. I've had it drive Ryan and Bunning down to mop up RP quality in years where they historically had WAR of 5 or greater. And once TCR gets a grip on a player like this, it doesn't give up.

I also had a situation where in two consecutive years during the season TCR drove a LFers rating from 10 (20 scale) to 1 by the end of the season. First off season day it popped back up to 10. Third year it stuck.

Then there's the situation where players underperform their ratings year after year. Yes, it happens.

Also playing with historical retirement off brings a lot of variability into the game. What will become of Ken Hubbs and Sandy Koufax and J R Richard?

I agree knowing RL is an advantage for a human GM but with the right settings its not quite as extensive as might be thought initially. However if there's a draft the human GM's advantage becomes much greater.

[phenom]

Originally Posted by uruguru

If you want historical accuracy, you need to use recalc.

But if you do that then you, as the player, have a MASSIVE advantage over the AI because it does not know how players are going to develop. That means you need to enjoy wiping out the AI with your perfectly timed All-Star rosters, or implement major restrictions to trading and drafting.

1000 hours played is rookie level!

This is why when I want to take part in historical I do it as a manager and not a GM.

[Brad K]

Originally Posted by uruguru

If you want historical accuracy, you need to use recalc.

But if you do that then you, as the player, have a MASSIVE advantage over the AI because it does not know how players are going to develop. That means you need to enjoy wiping out the AI with your perfectly timed All-Star rosters, or implement major restrictions to trading and drafting.

1000 hours played is rookie level!

Drafting is where the big advantage comes in. I've created interesting and challenging saves by assigning players to historic teams and by voluntary restrictions on payroll. The Pirates Moneyball save was not about wiping out the computer teams.

[Brad K]

Originally Posted by uruguru

If you want historical accuracy, you need to use recalc.

I don't want historical accuracy. I want historical plausibility.

[David Watts]

Originally Posted by Brad K

I don't want historical accuracy. I want historical plausibility.

I'm with Brad. I have little interest in turning off all the features that make the game enjoyable, so that I can get perfect stat rates. I've stopped using recalc until random debut is fixed. I'm 3 seasons into a historical minors league starting in 1961 using only the OOTP development engine. I have TCR set to 1. I'm absolutely loving it and I have my doubts as to whether I will ever return to recalc. I was ready to toss this game aside due to how mad I was about random debut, but the cool thing is, this game has so many ways to play, one doesn't have to toss the baby with the bathwater. My next game will be a random debut without recalc. But, I plan on playing this historical minors game for awhile.

[Brad K]

Originally Posted by phenom

This is why when I want to take part in historical I do it as a manager and not a GM.

Playing with recalc need not be a massive advantage.

Use development and TCR, kill the draft and assign rookies to historical teams or keep the draft and let the assistant GM make the choices, and limit your payroll to under half the MLB median.

If you don't use retire according to history then you have a risk signing a 30 year old who RL retired at age 33 to a huge 8 year contract because you can't count on him retiring. But you also can't count on him being bad after age 33 either.

The game defaults to recalc, dev on, TCR on, and retire according to history off. Obviously those who designed the game aren't sold on Garlon's recommended settings as being right for most.

[uruguru]

Originally Posted by Brad K

TCR is the equalizer. I've had it drive Ryan and Bunning down to mop up RP quality in years where they historically had WAR of 5 or greater. And once TCR gets a grip on a player like this, it doesn't give up.

TCR is not really a good equalizer because the recalc will wipe out any TCR changes that occurred in the previous season.

[uruguru]

Originally Posted by David Watts

I'm with Brad. I have little interest in turning off all the features that make the game enjoyable, so that I can get perfect stat rates. I've stopped using recalc until random debut is fixed. I'm 3 seasons into a historical minors league starting in 1961 using only the OOTP development engine. I have TCR set to 1. I'm absolutely loving it and I have my doubts as to whether I will ever return to recalc. I was ready to toss this game aside due to how mad I was about random debut, but the cool thing is, this game has so many ways to play, one doesn't have to toss the baby with the bathwater. My next game will be a random debut without recalc. But, I plan on playing this historical minors game for awhile.

TCR = 1 is probably the best stab at historical accuracy without using recalc. There are some problems with it, though. For example, there's no amount of TCR that will turn relievers into starters. So if Phil Niekro starts as a reliever in the minors, then he stays a reliever. A very good reliever, but not a 300-win guy who pitches into his late 40s.


Another issue is the mediocre-to-average player who had one spike season in his 30s. There are lots of examples. That spike becomes his potential and suddenly he reaches it at age 23-24 and has a long productive career because of that one oddball season.

[Brad K]

Originally Posted by uruguru

TCR is not really a good equalizer because the recalc will wipe out any TCR changes that occurred in the previous season.

My experience with Ryan and Bunning is that it bounces mostly back the day after the season ends then TCR starts driving them down again. You won't have a staff ace by the time spring training camp breaks. There's a thread around here somewhere about my experience with Ryan.

Anyway, apparently you agree the others are decent equalizers. I stopped playing development only when Tim Foli won the batting championship several times and ended up in the Hall of Fame.

[Brad K]

Originally Posted by uruguru

TCR = 1 is probably the best stab at historical accuracy without using recalc.

Does TCR do anything without recalc?

[uruguru]

Originally Posted by Brad K

Does TCR do anything without recalc?

TCR controls the rate of random changes to current and potential ability. It is completely seperate from the recalc. A talent change, for example, can suddenly raise the potential and actual ability of a hitter to draw walks. A pitcher might suddenly gain better control or possibly even learn a new pitch (the jackpot). Talent changes are like the opposite of injuries.


TCR at its lowest setting, 1, means that there will be very very few of these unexpected changes to potential. Players will more predictably develop to their potential. A high TCR (max is 200) means that you will a lot of these types of unexpected changes across the league, over time. A mediocre prospect might suddenly become good -- maybe Tim Foli develops a power stroke and starts hitting 20-30 home runs every year.


If you have recalc off, then TCR changes will remain with players over their careers and they could go on to have vastly different careers than in the real world. If recalc is on, then any random TCR changes that may have hit your player will be wiped clean when the next offseason begins. This keeps their OOTP career on track with history.


Now, you may ask, what is the point of having TCR and recalc at the same time? Well, if you are mixing fictional players with historical players, then the fictional players will not be affected by recalc. Let's say you want to start a league in 1941 and then have ALL future players that enter the league be fictional. Recalc will ensure that Joe Dimaggio and Ted Williams continue to be who they were, but a high TCR will allow all of your fictional players to develop in unpredictable ways.