Weather in Solecismic sims

[gstelmack]

Has anyone done a study to verify the frequency of rain/storms/snow in Solecismic games vs. actual football games? This is something I first noticed in TCY with my South Florida career, and is continuing over to my current FOF2k4 career. It just seems like 1/3 of all my games are played in bad weather (worse in TCY with a South Florida career, where it seems like it's raining on every home game I play sometimes), but that in real football weather affects far fewer games.

Sure, Tampa has lots of rainy days every year, but those tend to be short 30 minute thunderstorms that dry out pretty quickly afterward. It doesn't rain enough to affect football games that much. How many Bucs' games this year were played in the rain?

And I'd expect 1 or 2 snow games in New England in December, but not 3-4 rain games during the fall.

I think the weakness is that the games use "rain days" and "snow days" stats that are based on almanac data, not actual "how many games were affected by rain or snow?" data. And it's definitely possible to have rain or snow on a day that have zero or negligible impact on a football game.

This bugs me because in the game rain is the great equalizer. Gameplans get screwed up (runners slip on turf) and random events happen more often (fumbles notably), reducing the effects of roster and gameplan decisions and increasing the impact of random number generation.

Anyone else seen this and agree / disagree? Has anyone compared weather stats to see how often teams actually play in bad weather?

[CraigSca]

I haven't, but your theory sounds entirely plausible.

[BC Bob]

I agree with you completely. I've been a Tampa Bay Bucs season ticket holder for about 8 years. In all that time, I've had no more than three games (including pre-season) in which there was more than a brief trickle of rain. but in FOF, I've had about 33% of the Tampa home games with "rain" and 17% are stormy. This contrasts with 15% "rain" and 0% stormy at away games. Sample size for each was about 70 games. The weather issue is one of the few things that has bothered me about FOF.

[Ben E Lou]

Originally Posted by gstelmack

I think the weakness is that the games use "rain days" and "snow days" stats that are based on almanac data, not actual "how many games were affected by rain or snow?" data. And it's definitely possible to have rain or snow on a day that have zero or negligible impact on a football game.

I've never really thought about this or examined weather in FOF/TCY closely, but that theory would make a lot of sense as to an explanation if it is happening as often as you say. I wonder if that's what's going on, that the almanac says that there are an average of 10 rain days in September in Tampa Bay, and therefore 1/3 of the games get rain, when in actuality it may only rain during the 1pm-4pm time frame 2 or 3 times a month. For example, I'm sure that a good number of the September/October rain days in Charlotte/Tampa Bay/Jacksonville/Nashville are heat-related thundershowers, too--the kind that show up at around 5pm and therefore have no impact at all on a football game.

[Dutch]

I've been a Tampa Bay Bucs season ticket holder for about 8 years.

I wish...

[VPI97]

Originally Posted by gstelmack

Has anyone compared weather stats to see how often teams actually play in bad weather?

I have nearly all the weather conditions from the games in IHOF (pre-season and post-season, too), so I can post those numbers, but I wouldn't know where to begin when it comes to analyzing it.

Anaheim
42 - Clear
22 - Rain
 
Ann Arbor (Retractable roof)
70 - Clear
 
Asheville (Dome)
66 - Clear
 
Bordeaux 
29 - Clear
36 - Rain
2 - Storm
 
Boston 
47 - Clear
18 - Rain
 
Boulder
51 - Clear
10 - Rain
4 - Snow
1 - Storm
 
Brooklyn
52 - Clear
14 - Rain
1 - Storm
 
Chesapeake
59 - Clear
13 - Rain
 
Columbus, In (Dome)
67 - Clear
 
Conyers, Ga
44 - Clear
19 - Rain
8 - Storm
 
Fairbanks (Dome)
77 - Clear
 
Fort Worth (Retractable roof)
66 - Clear
 
Frederick, Md
46 - Clear
24 - Rain
1 - Storm
 
Georgia 
46 - Clear
17 - Rain
1 - Storm
 
Gothenburg
37 - Clear
29 - Rain
 
Maassluis 
34 - Clear
34 - Rain
 
Miami (Leesburg, Va)
39 - Clear
27 - Rain
1 - Snow
1 - Storm
 
Minnesota (6 seasons dome, 1/3 season outdoors)
63 - Clear
2 - Rain
 
North Plainfield, NJ
59 - Clear
16 - Rain
1 - Snow
1 - Storm
 
Orlando
42 - Clear
17 - Rain
9 - Storm
 
Outer Banks, NC
43 - Clear
23 - Rain
1 - Snow
2 - Storm
 
Paris
40 - Clear
25 - Rain
 
RayCo (Nashville, TN)
44 - Clear
18 - Rain
2 - Storm
 
Rochester
40 - Clear
25 - Rain
1 - Snow
 
San Antonio (Dome)
71 - Clear
 
San Luis Obispo
58 - Clear
6 - Rain
 
Springfield (Dome)
65 - Clear
 
Syracuse 
42 - Clear
24 - Rain
1 - Snow
 
Telluride (Stillwater, OK)
49 - Clear
15 - Rain
3 - Storm
 
Tucker, Ga
36 - Clear
24 - Rain
5 - Storm
 
Vicksburg
42 - Clear
14 - Rain
8 - Storm
 
Williamsburg (Astoria, NY)
51 - Clear
12 - Rain
1 - Storm

[Ben E Lou]

VPI:

Just a quick combining of Tucker/Conyers/Georgia gives us 126 clear, 60 rain, 14 stormy.

I watched the Falcons virtually every Sunday from around age 6 to 22. There's no way it rained during a third of their home games. Methinks gstelmack is on to something.

[gstelmack]

Originally Posted by VPI97

Boston
47 - Clear
18 - Rain

This is just a quick spot-check, I'll have to build up a more detailed database (or I wish I could just find one), but this past season, including preseason and the home playoff game against Jacksonville (so 19 games), the Patriots had:

18 - Clear
1 - Light Snow (home against the Jets)

[dawgfan]

Weather has actually been a minor annoyance of mine with TCY. Now, it's no shocker to anyone that it rains in Seattle, and quite a bit in the Winter.

However, I've been going to Washington home games since 1978, and most home games in September have had fantastic weather - temps in the 70's, sunny. Very rarely is it not a nice day weather-wise for the September home games.

Not in TCY - more often than not, it's listed as bad weather, usually "Stormy".

Not a big deal to me, but then again I haven't run tests to see how much the weather affected my gameplan settings, and granting that gameplanning has gotten a lot more in-depth with the recent FOF games since TCY.

[gstelmack]

Originally Posted by dawgfan

Not a big deal to me, but then again I haven't run tests to see how much the weather affected my gameplan settings, and granting that gameplanning has gotten a lot more in-depth with the recent FOF games since TCY.

They key to this is that bad weather is "the great equalizer". Random numbers take on a bigger importance, with random events having a bigger impact on the game. More fumbles, more drops, and the like. What I've found (admittedly a bit anecdotally) is that more upsets appear to happen when the weather is bad. Or, maybe a better way to put it is that if the game is raining, I might as well flip a coin to see who is going to win, with the coin flip simulating who is going to get hit with more turnovers, drops, missed FGs, and sloppy play which can have a huge impact on the outcome.

[jbmagic]

and bad weather can cause injuries increase too.

[CraigSca]

It's odd, because I never really thought there was an issue in TCY with the frequency of upsets (as opposed to other college games). Considering that weather is the great equalizer, you'd think this would occur far more often than it does.

[Solecismic]

Very odd.

While the comment about the misuse of the almanac numbers is certainly warranted, the additional comments about injuries and upsets are not. I hope Greg undertakes that study, just because he sounds so convinced himself.

I posted on this issue at length years ago (including the raw numbers) - my study of weather effects on games indicates weather (temperature and precipitation) has a very small effect on rushing and passing numbers. Any quarterback who can't deal with a little rain probably lacks the concentration to make it past high school ball.

Remember, turf conditions affect both offensive and defensive players in a similar manner. I suppose there might be wide receivers who, like horsies, are good "mudders," but I haven't tried to study that at all to date. Maybe by FOF version 25.0. Any time weather affects the outcome of a play, it's mentioned in the expanded game log.

Wind speed has a much greater effect on a game, and that is not modeled in FOF, but would be in future versions.

[gstelmack]

Originally Posted by Solecismic

While the comment about the misuse of the almanac numbers is certainly warranted, the additional comments about injuries and upsets are not. I hope Greg undertakes that study, just because he sounds so convinced himself.

I posted on this issue at length years ago (including the raw numbers) - my study of weather effects on games indicates weather (temperature and precipitation) has a very small effect on rushing and passing numbers. Any quarterback who can't deal with a little rain probably lacks the concentration to make it past high school ball.

Remember, turf conditions affect both offensive and defensive players in a similar manner. I suppose there might be wide receivers who, like horsies, are good "mudders," but I haven't tried to study that at all to date. Maybe by FOF version 25.0. Any time weather affects the outcome of a play, it's mentioned in the expanded game log.

Wind speed has a much greater effect on a game, and that is not modeled in FOF, but would be in future versions.

Okay, maybe I will dig in. This could just be me, it just seems like I see more fumbles and dropped balls in games with rain vs. those without. I did not realise that if it doesn't say "slipped on the wet grass" or "the ball is hard to hang on to" (or whatever the quote is for a fumble due to wet ball), then weather had zero impact on the play. I certainly only see one or two of those during the course of a rain game, but it SEEMS like my running game tends to suffer and other oddities occur more frequently in rain games.

But since I've got LogFileProcessor, maybe I'll just have to do some analysis of drops / YPC / fumbles and the like for rain vs. fair vs. snow games. See if it is all in my head.

[gstelmack]

The statisticians out there may be able to make more headway out of this, but what I've got:

Game	Conditions	Fumbles	Drops	INTs	YPC
1986	fair Result	2.01	4.35	2.05	4.11
545	rain Result	2.24	4.19	1.88	3.77
16	snow Result	2.69	4.06	1.13	3.29
58	stormy Result	2.14	4.28	2.43	3.1
2605	Grand Total	2.06	4.32	2.02	4.01

I'm not sure if I have enough snow / stormy games for a comparison, although rushing YPC sure seems convincing with those.

For rain, we definitely have more fumbles (1 extra every 5 games), which is expected but not necessarily significant. We're not seeing something like 2 more fumbles every game. Interestingly, there are fewer drops and INTs, but again just 1 every handful of games or so. I could also add number of rushes and number of passes (maybe drops/attempt, ints/attempt, fumbles/rush would be handy to know?) if that would help.

We do lose .4 YPC, which sure seems potentially significant to the running game when in bad weather.

Full spreadsheet for the stat gurus that want to do further analysis to see if there are any real correlations is at http://www.gstelmack.com/summary.ods. I could also grab log files from other leagues to have even more games if desired (my league runs 0 - 2605, so leagues with game numbers higher than that are easier for me to deal with, or I can just combine spreadsheets after the fact).

[gstelmack]

Okay, here are the extra numbers added in (and spreadsheet linked above updated with them):

Game	Temp	Conditions	Fumbles	Drops	INTs	rushes	rushYds	passes	passYds	YPC	YPA	F/R	D/P	I/P
1986		fair Result	2.01	4.35	2.05	56.12	231.06	68.5	469.52	4.1127	6.8742	0.0361	0.0636	0.0299
545		rain Result	2.24	4.19	1.88	56.26	212.27	68.91	474.82	3.7691	6.9207	0.0401	0.0610	0.0274
16		snow Result	2.69	4.06	1.13	56.5	186.31	71.13	400.13	3.2952	5.6302	0.0479	0.0568	0.0164
58		stormy Result	2.14	4.28	2.43	55.76	173.21	68.17	446.93	3.0983	6.5777	0.0389	0.0628	0.0360
2605		Grand Total	2.06	4.32	2.02	56.14	225.57	68.59	469.7	4.0132	6.8697	0.0371	0.0630	0.0294

Again, the statisticians can tell me what's actually relevant here, but the interesting tidbit is that it does appear to back up another feeling I've had about rain games: the passing game works a bit better and the rushing game a bit worse. More fumbles and a lower YPC, while fewer drops / INTs.

[theclassic]

gstelmack,

I'm curious of what you mean when you say you have a log file processor. Is this a program I can download somewhere myself or purchase? Or is this something you have coded yourself?

[sovereignstar]

Originally Posted by theclassic

gstelmack,

I'm curious of what you mean when you say you have a log file processor. Is this a program I can download somewhere myself or purchase? Or is this something you have coded yourself?

I think it's in the FOF Extractor, which can be found in his sig.

[Solecismic]

Originally Posted by gstelmack

Or, maybe a better way to put it is that if the game is raining, I might as well flip a coin to see who is going to win, with the coin flip simulating who is going to get hit with more turnovers, drops, missed FGs, and sloppy play which can have a huge impact on the outcome.

So, still feel this way?

[gstelmack]

Originally Posted by Solecismic

So, still feel this way?

Nope. It's clear that my "it's just a coin flip" is vastly overblown. But, there are still some differences. Updated numbers (with points, differential, and FG numbers included):

Game	Temp	Conditions	Fumbles	Drops	INTs	rushes	rushYds	passes	passYds	FGM	FGA	Pts	Diff	YPC	YPA	F/T	D/P	I/P	FG%
1986		fair Result	2.01	4.35	2.05	56.12	231.06	68.50	469.52	3.02	3.85	40.65	9.73	4.11	6.87	0.0162	0.0636	0.0299	0.79
545		rain Result	2.24	4.19	1.88	56.26	212.27	68.91	474.82	3.15	4.10	39.28	9.20	3.77	6.92	0.0179	0.0610	0.0274	0.77
16		snow Result	2.69	4.06	1.13	56.50	186.31	71.13	400.13	3.56	4.69	31.38	8.88	3.30	5.63	0.0211	0.0568	0.0164	0.75
58		stormy Result	2.14	4.28	2.43	55.76	173.21	68.17	446.93	3.36	4.16	35.84	8.43	3.11	6.58	0.0173	0.0628	0.0360	0.81
2605		Grand Total	2.06	4.32	2.02	56.14	225.57	68.59	469.70	3.05	3.91	40.20	9.58	4.01	6.87	0.0166	0.0630	0.0294	0.79

Fumbles, drops, INTs all seem to change, but with minor differences only (one extra / fewer every 5-10 games). The most significant difference seems to be in rushing yards, with 1/3 yard less per carry. This means to me that rushing teams, or those that depend on a balanced attack (using the run to help improve the passing game), are at a bit of a disadvantage when playing in the rain. Games also seem to be a bit closer, although with still over a 9-point margin of victory on average I'm not sure how significant the difference is. Good passing teams may be making up the point differential.

I don't think any of this is unrealistic. Without the more detailed numbers you've got, my gut might tell me to expect even more randomness (hence my original conclusion). I'm not at all claiming that the sim of these weather games has issues.

My main problem is simply that there are too many of them. From these numbers, 1 in 5 games is a rainy game, and those 5 games include all those teams playing in domes, meaning that more than 1 in 5 games played outdoors is in bad weather.

Of course, if I really feel this way, all of this appears fixable. I can simply edit the city list to reduce the number of bad weather days expected for a city and get the results I want.

[Galaril]

How is the weather effects in FOF 2K7? Is it similar results to FOF 2K4.