Sign in to follow this  
zanner

Sport Simulation How To's

Recommended Posts

I'm looking into the "science" behind sport simulation software. I've played tons such as OOTP, Bowl Bound, Draft Day, and even the whole EA Sports genre across a number of sports. I'm interested in getting a basic overview in what is happening in the background to take players' ratings in several different areas and grouping them against another group to come out with an outcome that is ultimately affected by these ratings. Is this more related to the statistics that drive the program or what? A general overview on what's going on technically in the background would do for me at this point. If there are any books or other resources out there that goes into more detail that would be GREATLY appreciated!

For example

Player A at 5'8", 150 lbs: 95 speed, 90 running ability, 60 break block, 70 agility

Offensive line players (each player for simplicity) - 70 run blocking, 66 strength

Against

Defensive line (each player) - 90 run stopping, 100 strength


A simplified overview of what's going on technically and mathematically in the background is what I'm looking for. Using the ratings and to a degree size which could be a rating of itself (and for simplicity sake, text based so don't take into consideration any graphics).

Thanks in Advance

Share this post


Link to post
Share on other sites
As you might have guessed from the lack of response this is an extremely complicated topic! Essentially the games attempt to boil down the interactions of several athletes who have trained for years for each role, into something that can be modeled relatively simply.

Stats are the obvious way of doing that but the clever part is how those stats are used. Normally they would feed into the AI - turn rate, acceleration, top speed, observation, etc. Then there's the animation system to figure as well. The AI is limited to what it can do at any moment by the animation it is currently playing, or the players will no longer look realistic.

Then there's a lot of tuning that goes on; and probably a lot of hacking to cover edge cases!

Share this post


Link to post
Share on other sites
Thanks Noggs. I'm looking at more of a text based system as I think I've got the right assumption in that the basis of a good sim feeds the more complicated things such as animations, etc. A graphical representation I'm not worried about so much as what drives the performance, stats, etc in a team game.

I search, but there is very little I can find from the searches I do on this. My thought is a very basic and simplistic thought process regarding this, and I'm just looking for the methodology or more specific technical aspects that go into a decent simulation.

My thought process is as follows (and note these are more questions I'm trying to clarify rather than known facts) :
Statistics are a huge part of the framework or world that these players compete in? For example, in baseball I'm sure there is a statistical average for the majority players in the league. Now the better hitters we know are considered great by hitting anywhere from .330 -- .450 where .450 may only be done once in 100 years. Now let's say for the simplicity of this discussion only a Batting Rating of 10 exists and there is a player with a 10 rating...What is exactly being used to make a 10 rating perform (and on random) within this range more times than not? And then take in affect Pitcher ratings that he faces. Is it more statistics or some other mathematical focus (if it's something outside of statistics what topic or specialization within the mathematics field I should be trying to learn more about) that is doing the work or are these developers coming up with complex formulations of their own to "make the system work" for a season and even more.

Share this post


Link to post
Share on other sites
Quick thoughts.

For baseball you may want to look at player skills which are in opposition. For example, what skills does the pitcher have that the batter can overcome or accentuate. A batter has a home run skill and a pitcher has a home runs surrendered skill. Compared to the average, where do they place? A good home run hitter might be somewhat tamed by a pitcher who doesn't give up a lot of home runs. Change that pitcher with a pitcher who may as well pitch in the homerun derby and that same batter looks like superman. If you use statistics to generate your results then maybe the following works;

ResultPercent = average + (batterSkill - average) + (pitcherSkill - average)

Plug in batting average/opposition batting average. Below will be a very good batter against a better than average pitcher. The batter still hits well in the situation but the pitcher still has some effect.

Result = .280 + (.330 - .280) + (.270 - .280).

Contrast this with a poor pitcher;

Result = .280 + (.330 - .280) + (.360 - .280).

If you aren't using statistics in your formulas then maybe a ratings system could be used. A rating can be compared to a die roll internally. A rating of 10 means the checked for result will happen if a roll of 3 six-sided die is under 10. You'd need to analyze what kind of results are generated by the number and type of die to ensure a change in rating doesn't produce ridiculus results. Assume a rating of 5 in home runs means a batter will, on average, hit 15 home runs over a 162 game season. You'd want to make sure that a rating of 6 (an increase of only 1) doesn't produce 80 home runs a season. Your highest allowable rating should be reserved for the very best players and only produce results that these very best players might actually produce. Your average rating should produce numbers an average player might produce.

To make a good baseball simulation, you'll need to know the sport, the rules and the statistics it produces. What are the odds that a base runner on second base scores on a base hit? Bottom of the ninth, one out, tie game, base runner on third, deep fly ball down the right field line. The right fielder can catch it in foul territory. Should he? What era are you trying to reproduce? Baseball of the 1920's is different than today. Heck, baseball from 30 years ago is different.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this