Home/Away Performance needs to be tweaked

[Le Grande Orange]

Originally Posted by RonCo

As I remember, one point that has been statistically proven to _not_ affect performance in any way is the overall distance of travel. A New York team visiting LA doesn't appear to play any differently agasint them than they would against a team closer to NY. This would lead me to think that jet lag/sleep/whatever isn't much of a factor.

But bear in mind the east to west travel is not that bad, and scheduling rules require a day off in most cases when an eastern team travels from the west back to the east.

Road trips are also limited in length to a maximum of about 14 games these days, with most averaging around 10 games or so. In earlier days, road trips 24-28 games long were not uncommon.

One would have to run the numbers to see if there was any correlation between typical length of road trip and the relative success of a team's road record.

[RonCo]

Don't mistake my sloppy posting for the total truth. The numbers were run. I don't have them at my fingertips, but know the numbers showed miles travelled had no impact--hence my interpretation that jetlag/hotel fatigue is either non-existent or negligible. I'll admit I don't remember if it took into account 20-game+ road trips of the olden days, but I think you're graspng at straws to say a 10-14 day trip would show zero impact, but a 20 day trip would be a major influence--especially since we have 4 hard stats that we can show small but significant influences in (ks, bb, triples, defense) with regard to even modern road trips.

That said, however, the home field advantage has been shown to be dropping over the years--so perhaps the long road trips did have an effect, and perhaps they showed up in striking out even a little more, walking even a little less, and playing even poorer defense while on the road than happens today.

Regardless, though, I think people are talking past each other here. If you tell me my team is going to strike out 1-3% more often, walk 1-3% less often, hit a few less triples, and play a little worse defense on the road, that seems like it would be enough to account for an awful lot of the 3-6 win difference most teams see as a home field advantage/road disadvantage over 81 games. The ability of the other team to tailor its team to its park seems like a pretty reasonable answer to fill in the unexplaied parts. We have lots of examples of teams trying to do just that, even in the olden days.

Still, the great fun in the discussion is that no one involved has yet been able to prove beyond mathematical doubt what the full cause of the home field advantage is...so if you want to believe jet lag is the cause of some of it, be my guest. I'll stick with the things we've learned to date. We can both be happy.

[RonCo]

Dola,

And my NY --> LA example was just off the cuff. Didn't mean to infer anything about east-west directional differences, as I don't think the study confirmed there was any difference.

[jarmenia]

Originally Posted by Malleus Dei

Nonsense, and I showed you why. If the cause was anything BUT tailoring to any significant degree, then the home teams would have better stats. They don't, but they still win more games. The only way they can do that is to be more efficient - i.e, be better tailored.

Ok, I've got to ask this since I've seen you state this several times in this thread but can you please elaborate on what you mean by taloring their teams to their stadiums? MD, you have said yourself that the only significant stastical difference between home and away is strikeouts and walks. How would these two stats be affected by taloring a team to the park.

In my mind, taloring a team to the park would show a difference in hits, home runs, errors, and ect... but would not expect to see a difference between walks and strikeouts.

Now, perhaps its not the tayloring to the ballpark but being used to the ballpark. Taking batting practice with the same batters eye every day, with the sun at the same angles. That to me would mean the batters could see the ball better, thus being able to determine a strike vs a ball.

This could be taken even further. Perhaps add a familiarity rating for each ballpark. You would see teams playing in new stadiums strugle for the first part of their first season but their familarity with the stadium would increase faster than that of their opponents. In turn, if a player was traded, he would have to become more familer with his park before he starts to benifit. On the flip side, if he visited his old team's park perhaps he sees the ball better there until his familarity decreases because he doesn't play there as often.

Just my

John

[Treadle]

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.

Huh? Home teams win 54% of the time. And home field advantage is statistically significant in the minors too. If the home team doesn't have an advantage, then there's a flaw in the simulation.

There's a lot of reasons for the home field advantage, from strategical advantages to batting last, to tailoring to home park, to the fact that players just perform better at home. Whether it's familiarity, comfort or stress it doesn't matter since it's part of the game.

Just because there is a difference doesn't mean it's "forced" if a simulation game simulates it. It means its accurate. For example, batting averages are higher with men on base than with bases empty. This is just another fact of baseball. It has something to do with holding the runner on, put perhaps there's also pressure on the pitcher at play. If the game doesn't refelct this then its missing something in real baseball.

I'm in the MOgul beta test and I just simmed 100 seasons. The home team won 53%. The AI tailoring to home park is in Mogul, but it also seems to get the other stuff right too in this regard.

[Treadle]

Originally Posted by RonCo

Regardless, though, I think people are talking past each other here. If you tell me my team is going to strike out 1-3% more often, walk 1-3% less often, hit a few less triples, and play a little worse defense on the road, that seems like it would be enough to account for an awful lot of the 3-6 win difference most teams see as a home field advantage/road disadvantage over 81 games.

You're totally right RonCo. Perhaps all of it in fact. I think we're mostly on the same page. Its the folks who think teh game shouldn't reflect the facts you stated above that seem to be missing it.

[Le Grande Orange]

Originally Posted by RonCo

Regardless, though, I think people are talking past each other here. If you tell me my team is going to strike out 1-3% more often, walk 1-3% less often, hit a few less triples, and play a little worse defense on the road, that seems like it would be enough to account for an awful lot of the 3-6 win difference most teams see as a home field advantage/road disadvantage over 81 games.

Don''t forget though that 54% figure is a broad average. Sometimes individual teams will post home records that are significantly better, even lousy teams.

A few cases in point:

1909 Boston (NL)
Overall: 45-108 .294 ; Home: 27-47 .365 ; Away: 18-61 .228 ; +.137

1935 Boston (NL)
Overall: 38-115 .248 ; Home: 25-50 .333 ; Away: 13-65 .167 ; +.166

1963 New York (NL)
Overall: 51-111 .315 ; Home: 34-47 .420 ; Away: 17-64 .210 ; +.210

1973 Texas (AL)
Overall: 57-105 .352 ; Home: 35-46 .432 ; Away: 22-59 .272 ; +.160

1977 Atlanta (NL)
Overall: 61-101 .377 ; Home: 40-41 .494 ; Away: 21-60 .259 ; +.235

1978 Toronto (AL)
Overall: 59-102 .366 ; Home: 37-44 .457 : Away: 22-58 .275 ; +.182

1985 Pittsburgh (NL)
Overall: 57-104 .354 ; Home: 35-45 .438 : Away: 22-59 .272 ; +.166

And here's a really staggering example of the home field benefitting a team that had a poor record overall:

1945 Philadephia (AL)
Overall: 52-98 .347 ; Home: 39-35 .527 ; Away: 13-63 .171 ; +.356

[sporr]

The other thing to keep in mind is that, correct me if I'm wrong, MLB homefield advantage is significantly less than that in other sports. Since MLB is the only sport in which you can tailor your team to your park, we would expect just the opposite to be true if that were the #1 difference in record. Of course, something else could be at work here.

[Le Grande Orange]

Originally Posted by sporr

The other thing to keep in mind is that, correct me if I'm wrong, MLB homefield advantage is significantly less than that in other sports.

I took a little look at that going back a roughly equal amount of games in each of the big four pro sports in North America, working backwards from 2003.

The numbers looked like this:

    |       |       OVERALL       |         HOME        |         AWAY        |
    |   GP  |    W    L   T   PCT |    W    L   T   PCT |    W    L   T   PCT |    W%   DIFF
NBA | 4756  | 2378 2378   - .5000 | 1450  928   - .6098 |  928 1450   - .3902 | 21.95  .2195
NFL | 4816  | 2405 2405   6 .5000 | 1410  995   3 .5862 |  995 1410   3 .4138 | 17.23  .1723
NHL | 4920  | 2155 2155 610 .5000 | 1203  952 305 .5510 |  952 1203 305 .4490 | 10.20  .1020
MLB | 4850  | 2425 2425   - .5000 | 1313 1112   - .5414 | 1112 1313   - .4586 |  8.29  .0829

As you can see, the NBA came out on top, with the NFL coming in second, the NHL third, and MLB last. It's interesting to note how fairly similar the NHL and MLB figures are.

[f.montoya]

One more thing of note is that there have been teams that have tailored their ballpark to fit a certain player or players and had great success at home as a result.

Whereas NBA tends to officiate in favor of home teams(a marketing strategy that they'll never admit to), MLB, while not without its controversial calls, has never shown such blatant disregard for the game of baseball. The NHL having much more lax rules regarding contact than the NBA, leaves the NBA as the sport with the most oportunities for swayed officiating.

I still say that "if" anyone were to put this to the test, i.e. creating 3 leagues (One with varying park effects coupled with adjusted player ratings with 4-6 players somewhat suited to their homefield. Another league with varying park effects and totally random, A.I. drafted rosters. And another with park effects set at 100 and random, A.I. drafted rosters.) , running 20-30 replayed seasons each, I'd be willing to bet money that the first league will produce results that could kill this thread.

Just thinking out loud.

[RonCo]

Originally Posted by Le Grande Orange

Don''t forget though that 54% figure is a broad average. Sometimes individual teams will post home records that are significantly better, even lousy teams.

A few cases in point:

1909 Boston (NL)
Overall: 45-108 .294 ; Home: 27-47 .365 ; Away: 18-61 .228 ; +.137

1935 Boston (NL)
Overall: 38-115 .248 ; Home: 25-50 .333 ; Away: 13-65 .167 ; +.166

1963 New York (NL)
Overall: 51-111 .315 ; Home: 34-47 .420 ; Away: 17-64 .210 ; +.210

1973 Texas (AL)
Overall: 57-105 .352 ; Home: 35-46 .432 ; Away: 22-59 .272 ; +.160

1977 Atlanta (NL)
Overall: 61-101 .377 ; Home: 40-41 .494 ; Away: 21-60 .259 ; +.235

1978 Toronto (AL)
Overall: 59-102 .366 ; Home: 37-44 .457 : Away: 22-58 .275 ; +.182

1985 Pittsburgh (NL)
Overall: 57-104 .354 ; Home: 35-45 .438 : Away: 22-59 .272 ; +.166

And here's a really staggering example of the home field benefitting a team that had a poor record overall:

1945 Philadephia (AL)
Overall: 52-98 .347 ; Home: 39-35 .527 ; Away: 13-63 .171 ; +.356

I was using your numbers form the first page of this thread to indicate that the home field advantage works out to between 3-6 games difference (a .500 team that wins .54% at home wins 43.7 games at home rahter than 40.5, for example. That there is statistical variation in the numbers is expected.

Regardless, the numbers are the numbers.

I'm fine with Markus coding a home field advantage if it entails small changes in k, bb, triples, and defense. I'm against it if it entails coding an across-the-board decrease in capability. I take this position because the numbers we can measure support it.

[Le Grande Orange]

Just to add more fuel to the fire, since Earl Weaver was mentioned earlier in the thread in reference to the value of tailoring a team to its home park, I decided to take a look at the results posted by the clubs he managed.

From 1969-1982, Weaver managed the same club (Baltimore Orioles) in the same ballpark (Memorial Stadium) for the full season. Here's how the team did over those 14 years:

 Year    W   L   PCT   HW  HL   PCT   AW  AL   PCT  DIFF     W%
 1969  109  53  .673   60  21  .741   49  32  .605  .136  10.09
 1970  108  54  .667   59  22  .728   49  32  .605  .123   9.26
 1971  101  57  .639   53  24  .688   48  33  .593  .096   7.68
 1972   80  74  .519   38  39  .494   42  35  .545 -.052  -5.00
 1973   97  65  .599   50  31  .617   47  34  .580  .037   3.09
 1974   91  71  .562   46  35  .568   45  36  .556  .012   1.10
 1975   90  69  .566   44  33  .571   46  36  .561  .010   0.95
 1976   88  74  .543   42  39  .519   46  35  .568 -.049  -4.55
 1977   97  64  .602   54  27  .667   43  37  .538  .129  10.65
 1978   90  71  .559   51  30  .630   39  41  .488  .142  12.63
 1979  102  57  .642   55  24  .696   47  33  .588  .109   8.53
 1980  100  62  .617   50  31  .617   50  31  .617  .000   0.00
 1981   59  46  .562   33  22  .600   26  24  .520  .080   6.78
 1982   94  68  .580   53  28  .654   41  40  .506  .148  12.77
---------------------------------------------------------------
TOTAL 1306 885  .596  688 406  .629  618 479  .563  .066   5.50

Interesting, the differential between home winning percentage and away winning percentage for those full 14 seasons is actually lower than the historical MLB average! It comes in at .066 when the historical differential is .080-.090.

If Weaver is a strong proponent of tailoring his team to the particulars of his park, it's curious why his tenure ended up posting a differential lower than the historical average. If anything, it should have been equal or higher.

[Malleus Dei]

I hope that you all had a Merry Christmas.

This discussion is, frankly, appalling.

I'm LOLing at LGO's inappropriate use of statistics yet again. Earl Weaver complained about being hosed by home field advantage due to teams like the Royals taking significant advantage of tailoring their team to their home stadiums in WEAVER ON STRATEGY, but of course you had to have read it to know that (it's available for really cheap on Amazon; if you play OOTP, go buy a copy).

Bill James has not only discussed this issue on numerous occasions (for example, in his collected work THIS TIME LET'S NOT EAT THE BONES he mentions that the Yankees' winning years correspond precisely with the years when they tailor their team to their park) but his knowledge of how to tailor a team to a particular park is attributed to be one of the reasons for the recent Red Sox resurgence.

The Dodgers and Mets have publicly admitted to tailoring theirs team to their parks, the Mets doing it based on park factors. There is little doubt that other teams do this as well, but don't talk about it much.

None of this, of course, will mean anything to anyone if they hasn't bothered read the seminal/definitive works on baseball analysis and strategy and if their mind is already cast in concrete.

Which is why I am leaving this discussion with a message: read more.

LGO, you make really fine schedules.

I hope that eveyone has a nice New Year.

[mlyons]

Originally Posted by Malleus Dei

You like to look at data - check the Yankee's records when they had left-handed power versus when they did not.

The most obvious example of this in recent memory was 1990, when they were pretty awful and only had two left-handed bats in their regular lineup, with only 17 homeruns between the two of them. The thing is, though, that season they went 37-44 at home and only 30-51 on the road -- a more extreme home/road split than the historical one. They certainly weren't tailored to their ballpark that season, but the home/road split was even more pronounced. Why? (For the record, the lack of left-handed power combined with a larger than usual home/road split was a feature of each of their losing seasons from 1989-1992, if you're worried about sample size issues.)

[mlyons]

On the opposite end of the spectrum, you have the 1927 Yankees, a great team loaded with left-handed power playing in a park specifically built to accomodate their best player. They finished 57-19 at home and 53-25 at home -- a smaller than normal home/road differential.

[Dark Horse]

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.

If anything changes have been made for the next version, or any subsequent version, i'm guessing it will remain optional. No sense taking out something that's already in.

The homefield advantage shouldn't be to home team players anyway. i don't think anybody wants to make Adam Everett into Miguel Tejada just because he's playing at home. But my theory would be that visiting team players are less effective because they are on the road. And, that doesn't mean Miguel Tejada turns into Adam Everett when he plays on the road.

Originally Posted by statfreak

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.

Again, it's not as much about making either the home or visiting team better, as it is about making the visiting team less effective. There are ways to simulate the homefield advantage.

1. Players tailored to parks. - MD is right about many of us not taking full advantage of picking players that fit our parts. And i doubt that the AI takes advantage of it at all. IMHO how OOTP does stadiums needs to be tweaked. But as the game is now, you can do this to some degree. If you have a stadium where RHs get a "boost" in BA & HR then your team should have more RH hitters. And when a visiting with more LH hitters will struggle. Also, LH pitchers will struggle more as well. If we did this it would help HFA without artificially boosting or dropping players.

2. Home team more familiar with parks. - There's no statistical way of tracking this. But that doesn't mean it doesn't exist. And again it doesn't have to be a big really obvious thing. And it is something that can be tweaked during development or even a customizable iteam. My idea, visiting defensive players get a "familiarity check" maybe there's a 1% or 1.5% or 2% chance that an outfielder will lose a fly ball in the lights or misplay a ball off the wall or an infielder will be surprised by the bounce of a grounder or a fielder misjudges the dimensions of the field and let a catchable fall because he didn't know where a wall was. That small chance will result in a team being involved in such a play maybe two or three times a year. Again this isn't changing the players. It's not even guaranteeing that the home team will win. But it is a HFA.

3. Visitors are road weary, more tired. - Slightly increase the amount of fatigue a player experiences while on the road. And slightly decrease the amount of rest a player gets on the road when they don't play.

4. Strikeout/Walk differences that have been mentioned in this thread. - It looks like these stats are available and easily translatable into the game.

Originally Posted by statfreak

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.

Part of what makes a game a reasonable simulation is if it makes me react to situations the way we should IRL. If we didn't have lefty/righty splits how many people would think it was a reasonable game; even if all the batters and pitchers overall stats came out "accurate"? In the game, i actually feel dread sometimes when i go into a big series or playoff series when i have home field advantage. When i played solo leagues, i wondered if it were better for me hold back at the end of the season so i wouldn't get home field advantage.

Originally Posted by statfreak

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.

It's funny, you're saying "To keep the game accurate, let's keep a part of the game inaccurate."

The game can be tweaked during development until it "feels" right. It doesn't have to come out that way before that. To that end, HFA doesn't necessarily have to be changed immediately. SI/Markus could do some research to see the effects of HFA, how often HFA comes into play, etc. and then bring it to the gam.

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. At that point, OOTP ceases to be a game and is simply a program that spits out pre-determind numbers to satisfy the masses.

i'd be bored too, if every home team won 54% of their games. That would mean that every team played .500 ball... every season. LGO pointed out the more important stat. That 84% of the teams won more games at home than the they did on the road. Everybody didn't win the same amount of games, but they won more at home.

[RonCo]

Malleus Dei is, of course, right when he says that better teams tailor their rosters to their ballparks.

He's also right that parks with equal dimensions can still be tailored to. Anyone who watched the Cubs of the late 80s trying to paly on astroturf fields should be able to immediately understand that. Some clubs even used to be notorious for tailoring their fields to their personnel (watering the infield, or not, depending on the speediness of thier players is an obvious example). In addition, the fact that the dinemsions are the same all around is not the requirement for tailorability...it's the distance to the fences that matters there--bring all the fences in and you want to tailor for power, push all the fences back and you want to tailor for speed...as an example. Foul ground is another tailorable characteristic, as is hitting background. Neither of those has to do with fence distance at all.

Look at the St. Loius Cardinals of that time, and you'll find a ballclub about perfectly tailored to a park--consistently in the upper regions of the league in OBP, doubles, triples, and SB, and at the bottom in HR...while winning a world series and a couple NL pennants. They played in a park that accentated speed and defense...and that's what they built their team around. Back then, if you couldn't run and get on base, you didn't play much for the Cardinals.

The numbers are still the numbers, though. I've yet to see any sabermetrician's study show definitively what causes the home field advantage. Perhaps it's been done, but I know I haven't seen it. If it exists, please point me to it.

I have seen data, however, that indicates travel distance (hence travel fatigue/jet lag) is not the cause of the homefield advantage.

[mlyons]

Originally Posted by RonCo

Look at the St. Loius Cardinals of that time, and you'll find a ballclub about perfectly tailored to a park--consistently in the upper regions of the league in OBP, doubles, triples, and SB, and at the bottom in HR...while winning a world series and a couple NL pennants. They played in a park that accentated speed and defense...and that's what they built their team around. Back then, if you couldn't run and get on base, you didn't play much for the Cardinals.

The 1987 Cardinals, who I think are the team you're referring to, finished 49-32 at home and 46-35 on the road, a much smaller home/road split than we've seen historically. Contrast that to the 1998 Cardinals (built around power without much emphasis on speed or defense) who had an extreme home/road split of 48-34/35-45.

Anyone have an example of a team tailored to its home park that actually exhibits a more extreme home/road split than normal?

[RonCo]

I was looking at St. Louis throughout the 1980s, a span where they consistently were in the top of the league in OBP, doubls, triples, and stolen bases. They also had several seasons where they showed larger than norm home/away splits...as well as some where they did not.

Statistics are like that, though. Random variance requires larger sample sizes than a handful of years.

Still, take the Cardinal's numbers during that span along with Whitey Herzog's very public acknowledgement that they were acquiring players to match their park (and a few others) "because it's hard to hit home runs in these big parks" (as I remember him being quoted, and I'll stand by my selection of the Cardinals as a team prudently tailored to their park.

P.S. I'm quite confused as to why you singled out 1987, since I never isolated that year in my post, and specifically gave a generic "late 80s" tag to my comments.

Additional Edit: The Cardinals won more than 54% of thier home games 7 of 9 seasons from 1981-1989. 5 of those 7 they won substantially more than 54%--specifically including 1987, when they won 60.5% of their home games. Their "problem" with proving home field advantage based only on that season is that they were a danged good team all-around, and won 56.7% of their away games.

[RonCo]

Dola -- the 1998 Cardinals are a great example of this, too. They played in a park that was configured very differently from the park the Cardinlas of the 1980s played in. So it's not surprising that they build to a different criteria.