My 3D ballpark models

[BigRed75]

You'll need to supply the distance and wall height for left/right field lines, straightaway left, left-center, straightaway center, right-center and right field. You'll also need month-by-month temperature and precipitation data (easy to get from Wikipedia) as well as altitude (also a Wikipedia thing).

Plug it all in, and it will calculate the park factors as you go. It's interesting to watch them change with each new piece of data you put in.

OOTP Park Generator

[drksd4848]

If only it could generate 3D ballpark .obj files...

[LouGerig_IronHorse]

Originally Posted by drksd4848

If only it could generate 3D ballpark .obj files...

Nah - Silvam is the King of the 3D ballpark .obj files! No computer automated computer program can match his skill.

[LouGerig_IronHorse]

Originally Posted by BigRed75

You'll need to supply the distance and wall height for left/right field lines, straightaway left, left-center, straightaway center, right-center and right field. You'll also need month-by-month temperature and precipitation data (easy to get from Wikipedia) as well as altitude (also a Wikipedia thing).

Plug it all in, and it will calculate the park factors as you go. It's interesting to watch them change with each new piece of data you put in.

OOTP Park Generator

Month by month average temperature is just the average of high/low for the month, correct?

Precipitation data - Wikipedia does provide how many "precipitation days" a given city has on average per month (I think averaged out over a 30-year span of time, which is what the National Weather Service uses for "normal" or "average" for a given date. But this doesn't appear to line up with what is input into the game.

How are the numbers derived that the game shows? How are we arriving for a given location that the percentage chance of precipitation for, say, New York, is - March = 3 .. April = 4 -- May = 3 - June = 3 - ... and so on? What data is being used to derive this and how is it calculated? Even on the Jaffe Ballpark Creator - those numbers are plugs into the ballpark creator - so I'm at a loss as to how those specific percentages are derived.

Wikipedia shows number of rainy days for (I think June?) as something like 10.2. So in a 30-day month, that's roughly a 33% chance that the month will see a rainy day. How are we then getting a plug number of 3 for a month? Are we taking that number then dividing by 24? (hours in a day) ... what's the math involved to get this number?

This part isn't intuitive - and there's absolutely nothing in the online manual that details this.

[BigRed75]

Originally Posted by LouGerig_IronHorse

Month by month average temperature is just the average of high/low for the month, correct?

Precipitation data - Wikipedia does provide how many "precipitation days" a given city has on average per month (I think averaged out over a 30-year span of time, which is what the National Weather Service uses for "normal" or "average" for a given date. But this doesn't appear to line up with what is input into the game.

How are the numbers derived that the game shows? How are we arriving for a given location that the percentage chance of precipitation for, say, New York, is - March = 3 .. April = 4 -- May = 3 - June = 3 - ... and so on? What data is being used to derive this and how is it calculated? Even on the Jaffe Ballpark Creator - those numbers are plugs into the ballpark creator - so I'm at a loss as to how those specific percentages are derived.

Wikipedia shows number of rainy days for (I think June?) as something like 10.2. So in a 30-day month, that's roughly a 33% chance that the month will see a rainy day. How are we then getting a plug number of 3 for a month? Are we taking that number then dividing by 24? (hours in a day) ... what's the math involved to get this number?

This part isn't intuitive - and there's absolutely nothing in the online manual that details this.

I use the average of the average high and low for the month, yes.

I'll be honest with you, I thought that the precipitation numbers were average inches if precip per month on Jaffe's, and didn't think it was precip days until I saw the in-game editor myself. What does Wikipedia say is the average amount of rainfall for NYC in June? Is it 3 inches, or round up to that amount? Does it line up for the other months?

You're right, some guidance on the precip numbers would be helpful...

[LouGerig_IronHorse]

Originally Posted by BigRed75

I use the average of the average high and low for the month, yes.

I'll be honest with you, I thought that the precipitation numbers were average inches if precip per month on Jaffe's, and didn't think it was precip days until I saw the in-game editor myself. What does Wikipedia say is the average amount of rainfall for NYC in June? Is it 3 inches, or round up to that amount? Does it line up for the other months?

You're right, some guidance on the precip numbers would be helpful...

According to the online OOTP manual - the number in the precip column is the % chance of precipitation for a given month. So if they're saying June for NYC = 3 - sounds like they're saying that for any given time throughout the entire month of June, there's a 3% chance of precipitation. OK ... all well and good. But how is that percentage derived? Not a word about how to get that number in the manual.

A great mystery indeed! We must unravel this mystery!

[Scott Vibert]

The Rainfall percentage is (or started with) the number of rainy days / days in the month / 6 or 8

Essentially the percentage of days with rain.. then scaled down to the fraction of the day that the game actually takes place.

The setup was originally derived out in OOTP4 or 5 (the first one with rain delays) when a straight # of rainy days / days in the month yielded too many interruptions.

[LouGerig_IronHorse]

Originally Posted by Scott Vibert

The Rainfall percentage is (or started with) the number of rainy days / days in the month / 6 or 8

Essentially the percentage of days with rain.. then scaled down to the fraction of the day that the game actually takes place.

The setup was originally derived out in OOTP4 or 5 (the first one with rain delays) when a straight # of rainy days / days in the month yielded too many interruptions.

Hmmm... very interesting! So then are we saying that if for NYC the month of June has 10 days with rain, which out of a 30 day month is 33% of the days - we then figure if a baseball game takes roughly 3 hours - and it's safe to say there's roughly a 12-hour window (noon to midnight) in which a baseball game would be played - therefore - 3 is 25% of those hours... so we'd be looking at 25% of the 33% - or, putting it in real numbers - if June has 10 rainy days - then 25% of 10 is roughly 2.5 - round up to 3. Hence - June = 3.

Something like that?

I'd still like to know the actual calculation that is used to derive this to know I'm getting an accurate number. What I presented above is kinda a stab in the dark based on what was posted (above).

[zappa1]

Travel Weather Averages (Weatherbase)

Here is a site for just about any weather related stats you want. Browse the country, pick a state, select a city or town from the list and you can spend lots of time going over all the data available. It's pretty cool. And there are also other sites to browse. This isn't the only one. Check it out. Have fun.

[LouGerig_IronHorse]

Originally Posted by zappa1

Travel Weather Averages (Weatherbase)

Here is a site for just about any weather related stats you want. Browse the country, pick a state, select a city or town from the list and you can spend lots of time going over all the data available. It's pretty cool. And there are also other sites to browse. This isn't the only one. Check it out. Have fun.

Thanks for linking this site! It has a BOATLOAD of weather data - this is far and away one of the best weather data sites I've ever seen.

An interesting feature is the daily averages breakout - two pie charts with a % breakout of every kind of weather condition possible, including differentiation in precipitation (rain, thunderstorms, drizzle, showers, etc.) as well as sky conditions (clear, scattered clouds, cloudy, overcast, etc.) - and wind direction breakout from every possible direction on the compass.

It would be a TON of work to compile this - but I believe using this data it would be possible to derive something pretty accurate. As I am semi-retired due to health issues - I might have to see if I can work up a test spreadsheet to see what the data looks like (and to see how long it would take to compile it). It would all be manual input - but pretty simple to do in Excel. Taking a sample month (say August) and sample location (the Bronx) - perhaps I can compile and calculate an average number to plug into the game that would accurately reflect the actual weather trends. According to the site, this section is based on 21 years of data (the National Weather Service typically uses a 30-year block of time to calculate 'norms'/'averages') - a 21-year average should be sufficient to derive a reliable average.

We'll see what it looks like. The overall 'monthly' number would simply be an average of the compiled daily numbers.

Given that it looks possible to differentiate between weather types, particularly precipitation types, I would love to see the developers work something like this into the game. Right now "rain" is simply a single category of "rain" ... it would be great to see the game differentiate between long-duration rain events vs. short-duration events (showers vs thunderstorms)... it would add a lot of realism to the game.

Something else - knowing the % of time skies are either cloudy or overcast (according to the weather data) vs the % of time there is precipitation could drive how often precip falls when the game reports 'cloudy' sky conditions. Right now it appears that 100% of the time in the game, if the sky conditions start off 'cloudy' - at some point in the game it transitions to 'rain'. That doesn't happen in real life. Many times games are played under a heavy overcast without a drop of rain falling. Or generally cloudy skies (high clouds for example) - with no rain.

This could be an interesting project.

[zappa1]

Lou Gehrig, Iron Horse:

Thanks. I found this site last year just doing a search. I was starting a fantasy season using two counties from my home state. Not all of the city and towns were listed the OOTPark's game, so I needed to find them out. I then added the data to the weather file in the game. I then saved a copy of it in a different folder. When the new OOTP17 came out, I replaced the weather file with the saved one. You can add many cities or towns to the weather file. Just keep a copy of the original file around in case you might need it. I've spent a lot of time checking out that site. It is loaded as you say with data. Very interesting stuff. Hope more people just give it a check. It is cool. Hope your plan works out for you.

[LouGerig_IronHorse]

Originally Posted by zappa1

Lou Gehrig, Iron Horse:

Thanks. I found this site last year just doing a search. I was starting a fantasy season using two counties from my home state. Not all of the city and towns were listed the OOTPark's game, so I needed to find them out. I then added the data to the weather file in the game. I then saved a copy of it in a different folder. When the new OOTP17 came out, I replaced the weather file with the saved one. You can add many cities or towns to the weather file. Just keep a copy of the original file around in case you might need it. I've spent a lot of time checking out that site. It is loaded as you say with data. Very interesting stuff. Hope more people just give it a check. It is cool. Hope your plan works out for you.

Thanks - me too!

I've started a spreadsheet using August as a sample month for location of the Bronx. (for Yankee Stadium, obviously). What's interesting, on the pie charts, the breakouts display percentages - but hovering over each pie shows 'records' - for each weather type - my guess is that if at that local observation station a given sky or precip condition was observed, it was recorded as a single 'record' ... so if it was clear, scattered clouds, cloudy, rain, thunderstorm, then clear - then 'clear' would have been recorded twice and the others once. The total of the records for the day comes out to 521 - obviously there aren't 521 years of records for a particular day in August. So viewing this from a percentage standpoint likely is the most accurate way of looking at it.

In order to get the cumulative average for the month, I think it makes the most sense to build the spreadsheet based on number of records for each category for each day - sum them - and get an overall percentage of that weather type for the month.

How that will translate to giving us a more accurate percent chance of precip in the game is yet to be seen. We shall see.

[silvam14]

Learning a lot about weather here haha


Sent from my iPhone using Tapatalk

[LouGerig_IronHorse]

Originally Posted by silvam14

Learning a lot about weather here haha


Sent from my iPhone using Tapatalk

I should have been a meteorologist. I studied (independently) meteorology since the age of five (I'm turning 54 in a few weeks) - and I wanted to get a masters degree (my undergrad was in communications) - but my father developed cancer at the time and I couldn't not be there for my folks.

Either way - the weather part of the game is good - but I think it can be a lot better. That's why I'm willing to try this experimental project to see if it makes any sense, impact, etc.

Honestly, I would like to be able to present to the higher ups at OOTP, using good, solid data, a strong idea for how to expand the weather portion of our favorite sports sim. Again - it's good - but it could be a LOT better.

[LouGerig_IronHorse]

Originally Posted by LouGerig_IronHorse

I should have been a meteorologist. I studied (independently) meteorology since the age of five (I'm turning 54 in a few weeks) - and I wanted to get a masters degree (my undergrad was in communications) - but my father developed cancer at the time and I couldn't not be there for my folks.

Either way - the weather part of the game is good - but I think it can be a lot better. That's why I'm willing to try this experimental project to see if it makes any sense, impact, etc.

Honestly, I would like to be able to present to the higher ups at OOTP, using good, solid data, a strong idea for how to expand the weather portion of our favorite sports sim. Again - it's good - but it could be a LOT better.

And, in truth, this is just as much about putting statistical analysis to good use - only using a different kind of data set - weather/climate data - as opposed to sports statistics - for the sake of improving the game.

Honestly, I believe it's possible to beef up the weather to really make it realistic. It's good now - it's close. But not all the way there yet.

[adion]

Since there are historical rainfall records for most, if not all, US cities and if they are daily records, can't the calculation of probability work? Using the
Average Number of Days With Precipitation numbers for June from Weatherbase.com for Cleveland & L.A. I figure the following: For instance, in Cleveland the average for June was 11.1 days of rain. Therefore the probably of rain was 37%, 11.1 divided by 30. For L.A. 1 day in June the probably chance of rain for a single random day would be 1 divided by 30 or 3.3%. Those percentages are higher than what is in games weather details. I don't know what impact they would have on games. Also, if you could find the actually number of days of rain for each city by year, you could refine it more. I have search for that detailed data here on the interweb but have not yet found it.
Travel Weather Averages (Weatherbase)
How to Calculate Probability (with Cheat Sheets) - wikiHow

[LouGerig_IronHorse]

Originally Posted by adion

Since there are historical rainfall records for most, if not all, US cities and if they are daily records, can't the calculation of probability work? Using the
Average Number of Days With Precipitation numbers for June from Weatherbase.com for Cleveland & L.A. I figure the following: For instance, in Cleveland the average for June was 11.1 days of rain. Therefore the probably of rain was 37%, 11.1 divided by 30. For L.A. 1 day in June the probably chance of rain for a single random day would be 1 divided by 30 or 3.3%. Those percentages are higher than what is in games weather details. I don't know what impact they would have on games. Also, if you could find the actually number of days of rain for each city by year, you could refine it more. I have search for that detailed data here on the interweb but have not yet found it.
Travel Weather Averages (Weatherbase)
How to Calculate Probability (with Cheat Sheets) - wikiHow

Yes - it could work, though given that the numbers are so much higher than what's in the game, it's likely that we'd see a LOT more weather-related interruptions, something one of the guys back up in the thread mentioned was a problem with earlier editions of OOTP when rain delays were first introduced.

My thought in looking at the data of actual logged events is that it might give us a VERY accurate figure for percent chance it might rain. What's more - I'm interested in seeing the different kinds of weather events - culling out the percentage of each kind of event (showers, drizzle, thunderstorms, etc.) - because I would love to present the developers with some data backing up a suggestion of "let's really beef up the weather and break it out beyond what is in the game now. Here's how we do it from a data perspective..." Something like that.

The weather site where I'm getting the event data also has a complete wind/direction event breakout. One of the things I find to be TOO simplistic in the current setup of weather in OOTP is that is appears to rely on one prevailing wind pattern when setting up a ballpark (not sure how the game engine then randomizes based on the "prevailing" wind pattern) ... but I would love to see how this might shake out. Wrigley Field, for example, is oriented that home plate out to dead center faces directly NE. Knowing how often the wind blows from what direction relative to the stadium - and breaking that out by month (the winds blow differently in April and May at Wrigley than they do in June and July) ...

No idea if this project will bear any fruit or result in any additions to the game, but I think it's worth trying.

[adion]

Where are you getting your weather info. I will have a look at the weather events and fool around with the probability thing and see what I come up with if you don't mind.

PS: You are just a kid. I will be 76 in November!

[zappa1]

Originally Posted by adion

Where are you getting your weather info. I will have a look at the weather events and fool around with the probability thing and see what I come up with if you don't mind.

PS: You are just a kid. I will be 76 in November!

Wow! That's great! To be your age and computer literate as you are with the backgrounds of Silvam's ballparks is awesome.

[adion]

I built my first computer back in the late 70's/early 80's and learned to program. I was an electronics tech in the USMC and after I got out and until I moved up the mgmt ladder. I have also had an interest in that kind of technology. Most of the nitty gritty I don't understand completely any more as I was trained and worked at the component level, e.g. resistors, capacitors, actual wires, etc. Now it is all stamped out of mythology!

Thanks for your nice comments.