Football GM simulation, ideas, thoughts..?

[Epi_862]

So this is whats going on. Im looking for your input on what are you looking for in a single player GM simulation. It's for a pet project of mine. Im not saying i'll ever release ANYTHING. Im not saying i'll ever even push out anything playable. But thats a possibility if it takes form.

Now, what im doing is i'll hook up a list here on what im starting to work on, and ask for input on things.

Stage 1 is pretty clear. The goal is to come up with a match simulation engine that produces realistic stats. This encompasses player and coach rating systems, and some dynamic ratings. It will contain system familiarity, and OC & DC's personal gameplans.

The stage one objective is to be able to produce a dynamic ratings system that can produce almost everything possible in the real world. IF and i really do mean IF i get to that stage 1 goal at some point, then i'll start to actually look at this as something that might become something.

But like i said, im now looking for your input.

A couple of disclaimers.
-you will not get paid.
-you will not be guaranteed to influence the product.

Why in god's name would you take time out of your schedule and use it to disect my idea's and implementations?

I dont know. I myself find it fun. And if the stars align, you might be a part of something that becomes something. Maybe, maybe not.

But now, i'll just lay out my worktable for stage one. So while you certainly can shoot ideas about the future, the most helpful stuff would be things that are currently under development. I'd also like to find a couple of people that would consider being constantly available for feedback & idea-bouncing. But i really do understand that it might be something noone wants to do.

Just quickly before the timeline:

What it WILL be:

A match simulation engine that can produce realistic results with a bunch of variables (players, coordinators, coaches, systems)

What it WILL NOT be, EVER:

A multiplayer or an online game. Also you will never call your plays. It's a match engine that will be built to POSSIBLY be a part of a GM simulation, not a coaching game.

So, the timetable now, subject to changing on a whim and all the time:

001:

Databases with a pretty robust version of the ratings system. This includes a couple of teams, maybe a hundred players, and a few coordinators and head coaches. I want to do this first, it'll help down the line to have the basic setup down.

002:

A single snap simulation with said databases. Looking to getting some sensical snap results out of a bunch of different circumstances. Coordinators, coaches and player ratings will be included in this one. After the results are sensical with different player/coordinator/timing settings, then:

003:

A full game snap-by-snap simulation with timing.

004:

A real database with players and teams.

099:

Labbing, adding, tweaking, kicking about.

100:

A match engine that produces results that im happy with, and onto stage 2. If it doesnt bring anything new to the genre, well, good riddance and onto something else.

[Toddzilla]

Ping Daivd Winter

[Epi_862]

So, for the first set of input im looking from you is about the ratings system.

I have a pretty clear idea of what it'll be, but i'd really like to hear about what do you think about the ratings systems in current games, and what do you think is lacking. I'll see if anyone has interest in this, and then i'll post my current plan, and see what you guys think.

The only thing i'll say: Ratings wise, everything in the NFL will have to be simulated. Of course, it's not possible, but thats the mindset right now. I dont think any current sim gets even close, and thats actually the whole force behind this project.

A couple of things that current sims cannot really replicate, just to tickle your brain:

- Haynesworth
- Mike Williams

Now i cant open up too much on how im approaching them yet, before i put out my ratings system draft, but for any ratings system, those two guys and their careers are a good way to start dissecting.

What i WILL say, is that the system will include static and dynamic ratings, as well as coach ratings, and system and positional ratings. I think my current system is pretty good, but it's bloated as fuck.

So have at it. Or dont. But hopefully some of you wanna lab this with me.

Thanks.

[Mustang]

Originally Posted by Epi_862

A couple of things that current sims cannot really replicate, just to tickle your brain:

- Haynesworth
- Mike Williams

Sorry, for players like Haynesworth, Maximum Football has already replicated player mechanics that make no sense and can bring the game down.

What it hasn't done though is where you don't use it for 2 years and it gets better so, good luck with that one.

[flounder]

I think Football Manager does a pretty good job replicating situations like Haynesworth and Williams. All players have hidden personality attributes that affect their on and off field behavior.

For example, I have two defenders on my team that have almost the same attributes, yet one is by far superior on the field. That's because he's a professional while the other player is a whiny bitch. It's possible I could take a different approach to my interaction with those players and get an entirely different result.

So for Haynesworth, the game could simulate an interaction with one coach where he is happy and performs well. Then, when he switches teams, he has poor interactions with his new coach and his on field stats suffer accordingly.

I think FM does this sort of thing really well, but they have a large team to work on things like that. I don't know if it's feasible to do the same sorts of things with just one person.

[albionmoonlight]

It might also be worth noting that spending a lot of time and resources trying to replicate the rare instances of a Mike Williams might take away from simming the vast majority of players who follow a "normal" development curve.

You could argue that Ryan Leaf, Akili Smith, and Tim Couch all busted for different reasons. But that a sim can do a pretty good job of capturing all three of those by simply having a percentage of guys drafted at the top lose ratings.

I understand and admire what you are trying to do. I just wonder if by focusing too much on the exceptions, you are going to short-change the rule.

[RedKingGold]

When you win, the game should give you free porn.

[JonInMiddleGA]

Originally Posted by RedKingGold

When you win, the game should give you free porn in a beer tent

Fixed that for you

[NorvTurnerOverdrive]

if it were me i'd focus on a true gm sim. an accurate business simulation for football. hire personnel, find players, set philosophies, manage the cap etc. but hands off when it comes to the on-field stuff.

the snap by snap match engine just seems like an impossible amount of work that will never be satisfactory to everyone. i'd be happy with a game recap and/or box score that doesn't go into a great amount of detail.

[Epi_862]

Originally Posted by NorvTurnerOverdrive

if it were me i'd focus on a true gm sim. an accurate business simulation for football. hire personnel, find players, set philosophies, manage the cap etc. but hands off when it comes to the on-field stuff.

the snap by snap match engine just seems like an impossible amount of work that will never be satisfactory to everyone. i'd be happy with a game recap and/or box score that doesn't go into a great amount of detail.

That was actually my initial idea too.

But when i fiddled with it a few months back, it just turned out that for a game as complex as football, there will have to be snap-by-snap simulation behind the scenes at least. It would remain a pretty simplistic game otherwise. Again, this is after a lot of thought on the subject, and im fairly certain that the things i want to accomplish even in this first stage, need a snap-by-snap simulation.

I can already see time-management code kicking me in the nuts.

[Epi_862]

And also, im sorry for not making this clear in the first post: TEXT SIM ONLY. No graphical representation of the action, ever. I know my limits.

[JediKooter]

Make it compatible with the Mac OS.

[RedKingGold]

Originally Posted by JonInMiddleGA

Fixed that for you

Let's up the ante to a porn tent with free beer.

That sounds fantastic....I might invent my own game now.

[NorvTurnerOverdrive]

well, good luck man. there's a reason fof is still the gold standard.



edit: i was responding to epi, not you rkg. your game sounds like the revolution the text sim world needs.

[CU Tiger]

Just a couple passing thoughts.
Football doesn't have matches, it has games.
If it has no intent of online multiplayer, I'm not interested...

[Noop]

I hope this works out for you. The football sim market is definite need of something new and good.

[Epi_862]

I have made a couple of prototypes of GAME engines now, and have a feel for what i need from my player/team/other databases, and how they interact with each other in regards to the GAME engine.

I even again fiddled with the idea of not doing snap-by-snap sim, but it will never produce good enough results. There are too many variables in the game of football for that.

Now i'm in the process of outlinging my databases, and trying to decide how will i store and access the data in-game. The problem is, i have a hard time leaving any of my ideas out, so the databases will become pretty large. I thought about storing data in propertiary files ala FOF, but im guessing in the long run i should just man up, learn to deal with a real relational database, and i will benefit from it from the long run.

So before we get to 0002, im taking my time with 0001 and learning a new skillset while im at it. But based on the tests i've done, the end result will be that much better.

[aran]

I've given a lot of thought to football sim design and have some partially completed play-by-play sim design documents. If you want to discuss this or maybe want someone to collaborate with, I'd be interested. We can talk through PM here, AIM (evizaer) or MSN ([email protected]).

[fantom1979]

Originally Posted by JediKooter

Make it compatible with the Mac OS.

Seems like alot of work to sell 2 or 3 more copies.

[sabotai]

BURN!!!!

[Epi_862]

Whats up guys, just a heads up.

I've been working on this thing for a few months, while managing a newborn & school, but i've continued working on it.

I have built some snap engines with really different influences, but i have not yet found a good mix of my ratings system. Im still hell-bent on trying to replicate almost every player type there is. When you take into account the fact that im doing it with systems & coordinators too, it really is a challenge. My current version has 21 physical & mental attributes (like agility, speed, intelligence, courage, learning), which dont chance much ever. Then there are about 20 football skills (like run blocking, catching, arm strenght), which are something that might fluctuate, or at least grow with experience. There's also about 10 traits that influence certain aspects (like Run over people, afraid of contact, ect.). Also there's positional experience, and system experience (if you need a west coast qb to plug in mid-season, you need one that has both qb experience, and experience in a west coast system).

Now these things are used to calculate things per snap. Before the snap, based on certain things, i'll calculate Morale, Confidence and Condition. There's a lot of things that influence these. Then the snap is "played" based on a lot of ratings. Just an example on how things work:

- first, based on the coordinator & depth chart, formation, personnel and play are selected, f.ex. 2 TE, 1RB, 2WR, long pass.
- second, ratings are calculated for the play. There are temporary ratings for this. In this example, there's a Long pass route running, long pass catching, run after catch, break tackles, avoid tackles. All of these are derived from the base ratings and mental state.

Why do all this work?

Because like i stated, my one and only goal is a match engine that can produce anything that real life could. Only if i achieve that, will i actually try to build a game around this. I want a player with low confidence to crumble after a fumble or a dropped pass. I want a role player with great determination and stamina and familiarity with the system to just show up in the 4 qtr. I want jay cutler and aaron rodgers and mike vick to all demand different pieces around them to make them successful.

So, what next?

I dont know. Im getting pretty weary of bumbing in to one of two issues:

a) Making a snap engine that is too simple
b) Making a snap engine that is way too complicated

Option A just runs against my whole reason for this project. Option B means the code becomes unmanageable. While working solo, there's always the tunnel effect - not seeing an obvious solution or a workaround.

But the point of this post is to say, the project is alive, and just giving you an update on why it's taking so damn long to get anything out.

[Marc Vaughan]

Originally Posted by flounder

I think Football Manager does a pretty good job replicating situations like Haynesworth and Williams. All players have hidden personality attributes that affect their on and off field behavior.

I think FM does this sort of thing really well, but they have a large team to work on things like that. I don't know if it's feasible to do the same sorts of things with just one person.

FMH has the same style of approach - basically model a few semi-hidden personality variable and then have a variable morale which is affected by interactions or in-game circumstances

The 'trick' imho is to try and think about players like real-life people and then model them accordingly - SO its not as simple as saying 'So and So is unhappy with his coaches comments so make him play badly' ... make sure you consider the exceptions in psychology - the lazy player who will play up when near the end of his contract, that sort of thing.

PS - If you ever want to PM me regarding this sort of thing feel free ...

[Epi_862]

Just a little info on this project:

It's not dead. I've been doing some prototyping, and coming up with a full design document. And for the foreseeable future, thats where it's gonna stay. There are a couple of factors:

- To simulate the game in the depth i want to do it, doing a 2d/3d/some other graphical representation would be the only way to go. It's also not something that can be done in a few months or so. There's already FOF with its simple (but very effective) single number vs single number system, where nothing happens in space. It's good for what it's trying to do, but it does leave something to be desired in general football logic. There's also PFS, which is doing a simplified simulation of the game pretty well, so there's no reason for me to try and do that either. What i want to do is a simulation that encapsulates many of the aspects that make football a great game, and i've touched on those earlier.

Now, i know i can do it. The problem is, it's really not a sensible thing to do financially right now. BUT: The design document is getting to it's finished state soon, and then i'll leave this project alone and work on getting my digital media company off the ground. I'll definetly come back to this project when it'll be more viable. There's a lot of things that the current sims dont do, that i already have answers for. There's also a some football-related stuff, that would propably draw in a general audience as well - much in the way football manager series has done for soccer.

But, all in all, since the market has FOF and PFS, both of which do a good job of a simple simulation, there's no reason to aim for that segment. And when the aim was set higher, and features mapped out, it became clear that as of right now, theres no way to get it done... But it will be done one day. Hopefully.

[Marc Vaughan]

Aim for something solid, realistic (within sense) and playable initially - then evolve it from there - thats what I always do with my sims, if you try to do everything at once you'll generally just end up with a mess imho.

(although the more development experience you have the easier it is to write something which can evolve ..)