Home/Away Performance needs to be tweaked

[Eugene Church]

The homefield advantage feature still needs some major tweaking...Almost all teams do better at home in real life (they win 54% of their home games in MLB) than on the road...even poor teams. Look at Tampa Bay, KC, Detroit, Seattle, Cincinnati and LA Dodgers...all had better records at home, despite poor overall records.

I think 80% of the teams in MLB have winning records at home.

This past season 24 of the 31 teams had winning records at home.

In my Alltime Allstar Association only 7 of the 16 teams play better at home.

I hope Markus perfects this feature in Version 7.

[gmo]

Glad somebody else feels this way. Including the HFA feature was good, but it still seems to come up pretty short in producing results that match MLB. There can be something said I suppose for not just tuning the game to the MLB ~.540 home team winning percentage, but where it is now I believe should be changed. LGO or anybody else have or seen home/away win/loss data for more than just MLB?

You might remember this thread from going on almost 2 years ago:
http://www.ootpdevelopments.com/boar...ad.php?t=61742

Post #7 cuts to the point. Even though the testing was done with V6.01, I suspect the feature has not been altered since then in as my solo league that has run since V5, the home team winning percentage has not changed noticeably with any iteration of V6 through 6.12. I have observed home teams winning at a rate of about 51%. The smaller-than-MLB home/away difference in results seems to be a result in the game of smaller-than-MLB home/away differences in hitting and no real differences in BB & SO. Thus to get the game in line with MLB, perhaps make hitters just a bit more better at home than away in the contact department (and maybe power, though the change in AVG may drive all the needed change in SLG) and give them the substantial observed differences in BB (eye) & SO (avoid K). Of course some to most of the change in the batters should come not just from altering them but from pitchers being better at home than on the road.

[sporr]

I agree. I think that Markus added the feature to satisfy those earlier complaints that many teams had worse home records than away records. However, I don't believe he gave quite enough benefit to the home players.

[Le Grande Orange]

Originally Posted by gmo

LGO or anybody else have or seen home/away win/loss data for more than just MLB?

MLB info is fairly easy to come by, but minor league info is not, unfortunately. It would make for an interesting comparison to see if minor leagues exhibit the same general level of home field advantage or if there are differences.

[Malleus Dei]

Originally Posted by Eugene Church

Yhis past season 24 of the 31 teams had winning records at home.

In my Alltime Allstar Association only 7 of the 16 teams play better at home.

I hope Markus perfects this feature in Version 7.

Eugene, consider that most real MLB teams try very hard to tailor both their style of play and their team to the properties of their home ballparks. This results in teams that perform better when playing at home than their competition does, because the away team has been designed for another ballpark. OOTP AI GM's and managers do not do this, and so OOTP AI-run teams do not tend to have better home records, home field advantage or not.

[nestajones]

Originally Posted by Malleus Dei

Eugene, consider that most real MLB teams try very hard to tailor both their style of play and their team to the properties of their home ballparks. This results in teams that perform better when playing at home than their competition does, because the away team has been designed for another ballpark. OOTP AI GM's and managers do not do this, and so OOTP AI-run teams do not tend to have better home records, home field advantage or not.

i disagree. i think this hardly happens at all. teams may pay lip service to it, and they may sign the odd player that they think will fit their park, but mostly they just try to get the best players they can. i think the homefield advantage is due to comfort level, familiarity, and batting last. i'd guess that fitting the team to the park accounts for 5% of the HFA at most.

anyway, if Markus does want to get this correct:
homefield winning % in the post-war era is 53.9
since 1960 it's 53.8
since 1980 it's 53.8
the last 10 seasons it's 53.6

it was a little higher before WWII. so yeah, calling it 54% seems fair.

and btw, during the 1970s, the height of the cookie-cutter ballpark era, HFA was 53.7%. Since many of the parks were very similar, if teams built their rosters to fit the park wouldn't we expect HFA to drop when teams fit the visiting parks as well as their opponents do?

I believe i read some study that most of the reason for HFA is that the hometeam walks a little more and strikes out a little less. that seems consistent with what GMO linked above.

You can find all these records on baseball-reference. when i run them, it also shows the pythagorean winning % for home teams, and these are consistently less than the actual winning % (like 51-52%). i would guess this is because of home teams winning more games by 1 run than visiting teams do, due to the game ending when they get a lead in the bottom of the 9th. home teams score about 3% more runs than the visitors.

[QuestGAV]

Originally Posted by nestajones

i think the homefield advantage is due to comfort level, familiarity, and batting last. i'd guess that fitting the team to the park accounts for 5% of the HFA at most.

I'm pretty sure the home team bats last in all my OOTP leagues so I'm pretty sure whatever advantage that provides is accounted for by OOTP

I think the Red Sox are a unique example of a team that does pay attention to their home park when putting their team together. Fenway is probably the most exploitable park in the majors but they've also got the will and knowledge to be able to break down and quantify exactly how playing in Fenway would benefit/harm a player they're considering rather than making the vague lip-service comments that most teams are willing to make. But like I said, they're a unique case.

[nestajones]

Originally Posted by QuestGAV

I'm pretty sure the home team bats last in all my OOTP leagues so I'm pretty sure whatever advantage that provides is accounted for by OOTP

well, yeah. i didn't claim it wasn't. and apparently there is a small HFA in OOTP.

I think the Red Sox are a unique example of a team that does pay attention to their home park when putting their team together. Fenway is probably the most exploitable park in the majors but they've also got the will and knowledge to be able to break down and quantify exactly how playing in Fenway would benefit/harm a player they're considering rather than making the vague lip-service comments that most teams are willing to make. But like I said, they're a unique case.

i'll buy that, you may be right. that is an extremely unique park though. the supposed poster-boy in recent years for getting a guy who fits your park would be Bill Mueller in 2003. still, the last two seasons he's gone back to the level of production one would expect out of him. so is he really a good fit for Fenway, or was 2003 just a fluke, or has he been slowed by injuries? i don't know...

[RchW]

Originally Posted by Malleus Dei

Eugene, consider that most real MLB teams try very hard to tailor both their style of play and their team to the properties of their home ballparks. This results in teams that perform better when playing at home than their competition does, because the away team has been designed for another ballpark. OOTP AI GM's and managers do not do this, and so OOTP AI-run teams do not tend to have better home records, home field advantage or not.

I'd add no nerves, no family probs (my kids always get sick when I go on the road).

Like the closer debate, I have accepted that my OOTP universe is different than real life and cannot always match MLB. Maybe it's me but I just look at the overall numbers. I find it more of a problem that in the last playoff series I lost, a 5-8 6.45 ERA pitcher beat my ace 20-7 3.65 ERA twice.

I'm still playing

[StyxNCa]

I'm not particular where I win, as long as I win.

[Malleus Dei]

Originally Posted by nestajones

i disagree. i think this hardly happens at all.

Bull. Go read WEAVER ON STRATEGY about this, especially the section about the Orioles at Kansas City versus at home.

[Le Grande Orange]

Originally Posted by Malleus Dei

Bull. Go read WEAVER ON STRATEGY about this, especially the section about the Orioles at Kansas City versus at home.

From 1901-2003, just under 84% of all major league teams finished with a better home record than road record. That's close to five out of six teams over 103 seasons. Sometimes the record was only marginally better, sometimes it was a lot better, but it was better nonetheless. And that includes teams which finished with great records overall as well as lousy records overall.

There's no way the "tailoring a team to its home park" is going to explain that 84% figure over that many seasons.

It may be a small part of the reason, but in no way is it the main reason.

[statfreak]

I basically disagree with homefield advantage and hope any such artificial adjustment to the home team players remains optional within OOTP.

While I don't disagree with the facts and real life results, I don't think arbitrarily boosting a home team player's ratings and decreasing a visiting team's player ratings is the answer. I want my players' ratings to be the same whether they are playing at home or on the road. I think too many users get too hung up in trying to replicate real life when the purpose of OOTP is to provide a reasonable simulation of the game of baseball.

There are just too many unknown variables and reasons why things happen in real life and trying to artifically model or account for that within OOTP seems not only pointless, but actually moves OOTP away from providing an accurate, organic, and reasonable simulation to a program that forces certain results to occur. My enjoyment of the game will be drastically reduced if OOTP does in fact begin having the home team win ~54% of the time since that would most likely mean the game results are forced and fixed and nothing I do as a user affects the outcome. At that point, OOTP ceases to be a game and is simply a program that spits out pre-determind numbers to satisfy the masses.

[Le Grande Orange]

Here's a little table I think I've posted before, but I'll post it again anyway. It shows the records of each position of finish in the AL and NL from 1901-1960, sixty years in which both leagues were constant in size at eight clubs in each.

NATIONAL LEAGUE, 1901-1960

          W     L    Pct     HW    HL   HPct     AW    AL   APct     W%  PctDiff
 1st   5813  3317  .6367   3063  1528  .6672   2750  1789  .6059   4.79   .0613
 2nd   5386  3740  .5902   2869  1702  .6277   2517  2038  .5526   6.35   .0751
 3rd   5100  4003  .5603   2746  1830  .6001   2354  2173  .5200   7.11   .0801
 4th   4766  4348  .5229   2606  1964  .5702   2160  2384  .4754   9.05   .0949
 5th   4418  4671  .4861   2419  2124  .5325   1999  2547  .4397   9.54   .0927
 6th   4030  5076  .4426   2209  2341  .4855   1821  2735  .3997   9.70   .0858
 7th   3687  5390  .4062   2021  2457  .4513   1666  2933  .3623  11.11   .0891
 8th   3222  5877  .3541   1806  2737  .3975   1416  3140  .3108  12.26   .0867
-------------------------------------------------------------------------------
TOTAL 36422 36422  .5000  19739 16683  .5420  16683 19739  .4580   8.39   .0839

AMERICAN LEAGUE, 1901-1960

          W     L    Pct     HW    HL   HPct     AW    AL   APct     W%  PctDiff
 1st   5814  3275  .6397   3150  1420  .6893   2664  1855  .5895   7.75   .0998
 2nd   5378  3714  .5915   2914  1640  .6399   2464  2074  .5430   8.18   .0969
 3rd   5078  4032  .5574   2778  1793  .6077   2300  2239  .5067   9.03   .1010
 4th   4755  4326  .5236   2580  1978  .5660   2175  2348  .4809   8.10   .0852
 5th   4451  4622  .4906   2424  2109  .5347   2027  2513  .4465   9.00   .0883
 6th   4086  4991  .4501   2263  2259  .5004   1823  2732  .4002  11.17   .1002
 7th   3689  5396  .4061   2074  2452  .4582   1615  2944  .3542  12.85   .1040
 8th   3104  5999  .3410   1731  2790  .3829   1373  3209  .2997  12.29   .0832
-------------------------------------------------------------------------------
TOTAL 36355 36355  .5000  19914 16441  .5478  16441 19914  .4522   9.55   .0955

Chart legend:

W is the aggregate total of wins posted by clubs finishing in that position.
L is the aggregate total of losses posted by clubs finishing in that position.
Pct is the overall winning percentage for each W-L record.
HW is the aggregate total of home wins posted by clubs finishing in that position.
HL is the aggregate total of home losses posted by clubs finishing in that position.
HPct is the winning percentage for each HW-HL record.
AW is the aggregate total of away wins posted by clubs finishing in that position.
AL is the aggregate total of away losses posted by clubs finishing in that position.
APct is the winning percentage for each AW-AL record.
W% is the actual number of wins at home divided by the number of predicted number of wins at home as calculated by the overall winning percentage. The result indicates the percentage of wins at home above what would be predicted based on the overall record. For example, a W% of 5.0 indicates that a team won 5% more games at home than it's overall winning percentage suggests it should have won.
PctDiff is the difference between the HPct and the APct.

We can see from the chart that over the sixty year period teams finishing in first had on average a .6367 winning percentage in the NL and a .6397 winning percentage in the AL.

The figures to note are the PctDiff figures for each position of finish. The NL shows something of a bell curve while the AL is more varied. Note how last place teams derive the biggest benefit in terms of W% from home field advantage due to them have fewer wins overall so each additional win gained at home is naturally a greater percentage of the whole.

It's also interesting to see a difference between the two leagues. The NL teams on average posted a home winning percentage .0839 better than the away winning percentage, which works out to almost 8.4% more wins at home. The AL showed a .0955 better home winning percentage, or almost 9.6% more wins at home than the overall winning percentage would suggest it should have. In the early decades of the AL the difference was quite pronounced, but as the years go on the spread narrows.

Given all this, it's obvious that there are a lot more factors at work than just a club tailoring its players to the particulars of its home parks.

[Le Grande Orange]

Originally Posted by statfreak

My enjoyment of the game will be drastically reduced if OOTP does in fact begin having the home team win ~54% of the time since that would most likely mean the game results are forced and fixed and nothing I do as a user affects the outcome.

No, that's only the overall average which includes all teams. A better way of looking at it is the spread between home winning percentage and away winning percentage, as the chart I posted above shows. A last place team in the NL finished on average with a .3975 home winning percentage and a .3108 away winning percentage, for example.

The fact remains that if OOTP shows results over a period of many decades where only 50% or so of teams are finishing the season with a better home record than away record, then those results are NOT realistic. It's strange to me why many folks would demand exacting accuracy when it comes to individual player statistics and yet allow an anomaly that large in team results to remain unchanged.

[f.montoya]

Interesting topic here. Malleus brings up a good point in his first post and it made me take a good look at my online league. We use real historical stadiums and the GM's often consider their parks when drafting, trading, etc. From the appearance of our league standings, most of the 20 teams are above .500 at home and five or six are at or above .600 at home. Maybe the AI needs tweaking to consider it's ballpark factors when making transactions.

Here's our league standings: http://www.league.montoyahome.com/

[f.montoya]

But I do believe there are other off-field factors that affect performance on the road vs at home. I just don't know if that's worth writing into the code. I mean it's variable. There are some players that can absolutely dominate wherever they play. There are players who suck at home and not on the road. To just blanket the entire universe of teams with a code that says "all players must suck just a little" for road games just may not be the answer.

[RchW]

Originally Posted by statfreak

I basically disagree with homefield advantage and hope any such artificial adjustment to the home team players remains optional within OOTP.

While I don't disagree with the facts and real life results, I don't think arbitrarily boosting a home team player's ratings and decreasing a visiting team's player ratings is the answer. I want my players' ratings to be the same whether they are playing at home or on the road. I think too many users get too hung up in trying to replicate real life when the purpose of OOTP is to provide a reasonable simulation of the game of baseball.

There are just too many unknown variables and reasons why things happen in real life and trying to artifically model or account for that within OOTP seems not only pointless, but actually moves OOTP away from providing an accurate, organic, and reasonable simulation to a program that forces certain results to occur. My enjoyment of the game will be drastically reduced if OOTP does in fact begin having the home team win ~54% of the time since that would most likely mean the game results are forced and fixed and nothing I do as a user affects the outcome. At that point, OOTP ceases to be a game and is simply a program that spits out pre-determind numbers to satisfy the masses.

You said it much better than I could. Thanks

[QuestGAV]

Originally Posted by statfreak

At that point, OOTP... is simply a program that spits out random numbers to provide a reasonable simulation of the game of baseball.

I wholeheartedly agree with you that it should remain an option. There's no reason for it not to be. But the distinction you're trying to make in your 2nd and 3rd paragraph is every bit as artificial as what you're trying to argue against. In the game of baseball, at least at the highest professional level, players play better at home when they do on the road and that difference shows up as a significant change in winning percentage. Short of adding a random number of runs to the home team after a game is simulated, I'm not sure how you simulate the actuality of home field advantage other than relative adjustments in the ability of players on the field.

[Eugene Church]

Originally Posted by RchW

Like the closer debate, I have accepted that my OOTP universe is different than real life and cannot always match MLB. Maybe it's me but I just look at the overall numbers. I find it more of a problem that in the last playoff series I lost, a 5-8 6.45 ERA pitcher beat my ace 20-7 3.65 ERA twice.

I'm still playing

I agree with you that it is a great game. I spend almost as much time on this game as I do in real life...thanks to retirement and an understanding and very wonderful wife.

In the overall realm of what is important, this is minor. But I also think, that when there is a chance to make the game like real life, then it would be good to do so, if it is possible to do so within the time limitations of the developer or developers and doesn't take them away from major features.