Hey Mike,

Training is exactly what you have described: combat aptitude. There are three levels of combat aptitude: Unskilled (0xp), Skilled (3,000xp), Highly Trained (6,000xp). They each have attributes (related skill earn rate, related skill bonus rate, and hitpoint earn rate).

In addition there are three level of arcane aptitude that mesh with combat aptitude, these are arcane aptitude... generally they are not germane to the conversation. Suffice it to say that they are mooshed in there with combat aptitude.

You also described classes... but let me refine it a bit: classes are a collection of skills that reasonably describes the basic knowledge of a profession. They are internally balanced, although one might not say precisely calibrated. Balance comes in the form of Experience Point Value equality: all Single classed professions are built using 10,000 XP (including the XPV of the level of combat aptitude); while Dual classed professions are built using 15,000 XP (including the XPV of the level of combat aptitude). Single and dual gain levels faster.

The last question you asked was what are types: Types are sets of skills and abilities that are applied to Racial templates (and in rare cases to individual characters). These sets of skills and abilities describe the minimal commonality of a group of like creatures. Example -- Type: Humanoid [Any Two Background Skills (1,000 XP) + Susceptible: Influence Humanoids (-1,000 XP)] 0 XP. This means that all Humanoids share these traits. The collection of their common traits, in this case, does not add any XPV to the character when building a Humanoid. Humanoid is implied when building playable races.

For the most part, the following are true statements about systems that accurately reflect a real world game setting: They don't have Levels of Experience, instead they rely heavily upon skills (usually purchased using skill points or in rare cases experience points); They do not have Types or Races as they are described here (essentially everything is a skill, and the skill tree is a very well populated one, otherwise, there is little difference between characters); They are heavily statistics based and most often use percentile dice for rolls against abilities (often skills and abilities are intertwined such than there are skills of affinity that add extra % points to other skill rolls).

Since the engine you are looking at designing should be calibrated to reflect the activity of athletes being athletic, you should focus on necessary attributes followed bu a set of derived attributes followed by a complex tree of skills. You should focus on 101 point percentile system where 0% means the skill is absent (can only be successful if external forces increase the percentage score above 0%) and 100% which means that use of the the skill is always successful when unmodified. You should focus on modifying those percentage points.

Further you should consider the accuracy of an action based on the percentage of the skill or score roll that you rolled. My gut feeling is that if you roll within the top 5% or bottom 5% of your modified score or skill roll you execute the maneuver such that it impacts the rolls of others and may mean corrective actions on your part. A successful roll that is within 2.5% of failure, means that the maneuver is executed dangerously and the player is in the DANGER ZONE. This may mean that you would have to spend effort to not lose the ball or not hurt yourself, or whatever the possible negative occurrence is for the play.

If a skill such as Juggling were rolled where the juggler had a 50% skill roll and there were no modifiers, then the attempt is successful if 50% or less is rolled; however, if that roll is within 2.5% (2%) of 50% then the juggling is inexpert requires another action to stabilize the juggling balls. However, if the roll is in the bottom 2% of the range, ie. very far away from failure, then the crowd goes wild, opposing player's reaction to THAT maneuver is penalized by 5% per percent difference between FLARE and 0%, while the agreeing player gains an equivalent bonus in reaction to THAT manuever (to be exemplified later), and the player earns a "Flare" point. Flare can be exchanged for things later... maybe a re-roll, maybe it can be cashed in on crowd appeal, maybe +5% on a future maneuver. Who knows... there are many possibilities.

For example, Bob the quarterback has a Short Forward Pass skill of 80% and an Acuity Secondary Attribute of 44%. Jeff the receiver has a Jink skill (which allows him make space between himself and opponents, thus increasing his chance to catch the ball) of 54%, and a Tight Catch of 60%. Roy on defense has an Obstruct skill of 55%, Diving Tackle skill of 58%, and a Reflex Secondary Attribute of 30%.

Somewhere in the play this happens:

Jeff runs past the line of scrimmage. He breaks away from the pack but is closely followed by Roy, Bob throws the ball to where he knows Jeff should end up. The rolls look like this:

Roy Obstructs (88% / 55% - FAIL)
Jeff Jinks (52% / 54% - DANGER ZONE)

In this case the impact of DANGER ZONE means that Roy gets a Reflex roll to interfere with the ball and Bob has to make an Acuity roll to adequately track the players' movement in order to stick the placement of the ball.

Bob's Acuity roll (12% / 44% - SUCCESS)
Bob Short Forward Passes (2% / 80% - FLARE 2%)

The difference between FLARE (the highest number in the range of numbers that make up the bottom 5% of the skill or score being rolled) and the actual roll result is represented after the word FLARE. In the example above, FLARE is 4% and the roll is 2%, so the difference is 2%. This percentage is multiplied by 5 and applied as a penalty to opposing players actions that are immediately affected by or in response to that roll and friendly player's immediately affected actions.

Roy has to choose to interfere with the ball (requiring a Reflex Roll penalized by Bob's FLARE or to attempt a Diving Tackle after Jeff catches the ball. Since the Reflex roll is low and further penalized, and Roy chooses to tackle Jeff. Roy's tackle is not directly affected by Bob's throw, so there is no FLARE penalty.

Jeff Tight Catches (62% / 70% [60% + 10% FLARE] SUCCESS)
Roy Tackles (58% / 58% - DANGER ZONE)

In this case the impact of DANGER ZONE means that Roy hits early. A flag is thrown for pass interference. The play ends, and the new play starts with the affects of the penalty in place.

===

From the expose above, it is obvious to me that the most important and distinguishing factor in play would be Actions. The term Play would essentially be equal to a combat round. From the time the ball lifts off the field to when the ball comes to rest or the play is called.

The term Action would be the number of times per Play a given player can act. This number should either be a derivative stat or should be a primary stat. There should not be too much difference in this number from layer to player. If a player has no remaining Actions, he can do no more. Think: Hero Games.

You may also need an attribute called Energy, how much energy a player has. Energy is recouped after breaks in play. Energy is required to commit Actions. If a player has no remaining Energy, he can commit no further Actions. Think: Mana from Magic the Gathering.

Another important stat may be Toughness. What happens when A player is injured? What % penalties does he suffer? How injured is too injured to play?

===

Some other skills that may be interesting to explore are:

Sundering Blow - where the attempt is to force a fumble. Additional FLARE effects: Target uses actions to prevent injury. DANGER ZONE effects: Attacker uses actions to prevent injury + Flag on play unnecessary roughness.

Hand Jive - when the ball handler moves the ball from side to side, to protect and distract. What are its FLARE and DANGER ZONE effects?

Good thoughts, I appreciate the feedback.

Indeed one of the things I will (eventually) consider is player health, injury status. Probably not out of the gate, but it is a consideration; could work well, i.e. in conjunction with a HP style simulation.

Another is player motivation. Plays could be designed with different AI, what to do in game situations. For instance, if a team is down 10 points and needs to make up for lost time, players might be motivated to gain more yardage on offense, but are then susceptible to turning the ball over. Or on defense, players can be motivated to go for the turnover, but are then susceptible to giving up the big play, or missing the tackle.

How this will be interesting is building on the underlying game engine, with game-oriented AI, player motivation, and so on. i.e. roll for success on a field goal kick, the degree to which the kick is successful, it's really successful. That could be reflected by some animation, for instance. Or degree to which he misses, could be really wide, left or right, depending on what the pass/fail/contention criteria are.

Thinking in terms of modeling the skills themselves...

In general, totally agree, it seems that "skills" are akin to "play instructions", the skills enable players to participate, and/or at a better chance for success.

The simpler variety are a simple modifier, to a play action, usually contested; i.e. "DB Bob tackles RB Jim; Jim breaks tackle, or Jim evades tackle".

Or even more "knowledge" based skills: "QB Rob reads coverage, QB Rob finds open receiver" or "LB Jim reads QB, or LB Jim reads RB".

How they're actually determined, contested, etc, haven't quite gotten that far, but I've got some notion what numbers are involved, at least quantitatively, if not also qualitatively.

Shifting gears to more complex variety of skills, the 250, 500, and even 1000 XP skills seem straightforward enough in their application. For instance, Tackling is a background skill, probably 500 XP. Whereas, Sure Tackling, is more like a weapon specialization, 1000 XP. Also, the simple Quarterback Throw, versus all the specialized kinds of throws that can happen.

However, skills such as Brawling, Wrestling, or even Martial Arts; in terms of gauging quality from XP cost, jumping into the 2-4k, sometimes 5-6k, varieties. Such as in the blocking game, aggressively knocking a guy into the turf during a run block, versus grappling (but not holding) a guy during a pass block. The overall effect seems simple enough, but I'd like to decompose them a bit; in other words, they seem like more of a composite of several simpler skills, if you will, or at least get a better sense as to the decomposition. For instance, Wrestling seems like at least 2x Street Fighting, plus perhaps a Weapon Specialization, bundled into its 2000 XP cost, plus perhaps 500-1000 XP value-added in the package.

Interesting thoughts re: penalties assessment. Could (no, probably should) be attached to negative skills outcomes. Rolling the proverbial 1 on a d20, for instance; at least potentially.

Sincerely...

Would you mind also explaining "internally balanced, ..." etc, etc?

Because how ever I slice it, I can't balance at least the XP. At least on the surface, this is what I might expect:

Training: Untrained? Trained? Highly Trained? 0, 3000, 6000?
Type: Single? Dual? 3000, 6000?

From there, subtracting out the skill XP costs?

Mechanically, basically, just assign a value to training, type, etc, and spend on skills, abilities, advantages, disadvantages, etc.

Thank you...

I really think that you are barking up the wrong tree with trying to model the kind of system that you discuss on any system like Dawn of Eana. It can be a fast fun system, but it bogs down in combat, and it is unnecessarily complicated. Experience points and skills purchased with XP are a mistake for the kind of system that I think you want to create. A better model is Top Secret, or even better Call of Cthulu from the old days... not the new system.

Think in terms of a limited set of primary attributes, a larger set of derivative attributes, and a large set of skills. The skills should be based on derivative attributes and derivative attributes based on primary attributes.

The way in which skills are acquired can be based upon training (ie. failed attempts) or upon use (ie. successful attempts). I will present two very good ways of increasing skills, one fore each method.

First some stipulations:
1. Skills are derived from a derivative attribute.
2. Skill scores are percentile based.
3. Every player has every skill in a range of 0% to 100% based upon whether or not points were initially distributed to the skill.
4. I don't know how many skills there are so I don't know what the initial distribution is.

Method 1: Failed Attempts
Bob's Forward Pass skill is currently 76%. During the course of a game, Bob fails a Forward Pass. Bob gets to mark that skill as used (this would happen automatically in a largely calculated game environment, like online games). At the end of the game, Bob is awarded a small number of XP, each point can be spent at a rate of 1 XP to increase a skill by 1%. XP can be spent only upon skills that are marked as used. If points go unused, they can be spent at a later time at a rate of 5 XP per 1% for any skill, used or not, or a rate of 1 XP x the current percentage of a Primary Attribute to increase that attribute by 1%, ergo, increasing a 20% Strength to a 21% Strength requires 20 XP.

Method 2: Successful Attempts
Bob's Forward Pass skill is currently 76%. During the course of a game, Bob Succeeds 21 Forward Passes. Bob gets to mark that skill as used 21 times (this would happen automatically in a largely calculated game environment, like online games). At the end of the game, Bob is awarded a small number of XP. At the end of any game where Bob's accumulated successes with Forward Pass exceed his Forward Pass score, all existing Forward Passes are removed and Bob's Forward Pass score increases by 1%. XP may be spent at a rate of 5 XP per 1% for any skill, used or not, or a rate of 1 XP x the current percentage of a Primary Attribute to increase that attribute by 1%, ergo, increasing a 20% Strength to a 21% Strength requires 20 XP.

Slow down there, Adam, take a step back, my friend.

I'm not completely done with my background reading, but for the moment, I just need to know the "balance" mechanics. At least as DoE (and/or Talislanta) are concerned, it would be helpful to understand the classes (single/dual), training, and how that balances a class/role, etc.

I am aiming to keep things simpler (i.e. fewer tiers of abilities, derived abilities, etc), and possibly even marginally "heroic". Where combat mechanics are concerned, there's the simultaneous nature of combat (i.e. play) resolution, but that's true for any system. I'm not sure players exactly need to "bid" for initiative, but that's exactly what that is, a bid and/or contest as deteremined during resolution, bearing in mind movement is all occurring simultaneously on or off screen.

It's easy to enumerate the skills (and/or abilities) and their potential contests. Also easy to handle a derived calculation tree. It will be more challenging to join up the physics with the actual game mechanics I think, but that's minutiae I am unconcerned with right at the moment.

Scenario in mind, receivers, defensive backs, will tend to be faster than running backs, tight ends, and especially so from (most) linebackers and linemen. That should be reflected by their Speed and potentially also initiative rolls. Movement itself should be ultra precise; i.e. movement precisely represented by Speed, for instance, possibly through the filter of player role, i.e. a Receiver's Speed 16 will not be the same quality as that of a Linemen Speed 16, type thing, a reflection of player "size" (height/weight/gear/etc).

The thing I couldn't get from, say, RuneQuest 6, was, in my opinion, a realistic model of "Speed", "Initiative", "Strike Rank"; being, Average(INT,CHA). INT and CHA? We're talking about trained combatants (i.e. athletes), who hone there physical prowess to the point of reaction speeds, not thought processes. Anyhow...

Of primary concern at the moment, let's assume DoE (or even Talislanta) are the system of choice. I am working through what constitutes skill "template" or "model". The simple ones appear to be asserted/contested. Simple enough. The more complex ones appear to be more far reaching, possibly composite (i.e. one more more asserted/contested), in nature.

I'm not as concerned about the actual numbers, at the moment, although I should have a sense of their quality, i.e. 250 versus 500 versus 1000 versus 2000+. Ultimately that issue will be tucked away and configured at the moment they are enumerated by the engine.

Thank ye...

When I take a step back, I'm not sure I can easily identify how classes are balanced, taking into consideration, training, type (single or dual), and so on, internally, much less inter-class. That's the first thing.

Some of this will need to be a judgment call on my part, I realize that. I'd like to at least be basing that on dependable, if not repeatable, numbers, because that sort of logical out-working can best be modeled through a simulation. I can then at least see the examples working out through the illustrated examples.

I may be over complicating it like you suggested.

Let me know when you have a chance.

Thank ye...

Taking a look at Top Secret.

Also taking a gander at the White Wolf, Storytelling System; I recall an early rendition of that back in the day, Werewolf, Vampire, etc. Which actually, reading some of the background, if it's accurate, is the system behind EVE Online back in the Naughts.

Actually, this all has me thinking of ways to further decouple the game engine behind the logistical parts of the simulation. Identifying a player, QB Bob, for instance, is one thing. Actually having the mechanics of his game work itself out in simulation is something else entirely (or should be).

Or at least on paper...

In Hero System, besides the plethora of ways to model things...

Scenario, play action fake pass: considered an Offensive Mental Combat maneuver? It's not a physical attack, per se, i.e. OCV versus DCV, but "mentally" the QB is trying to fake out defensive player(s).

Another scenario, QB "hits" WR on the run. Believe it or not, QB is actually trying to hit his moving target, so it's QB OCV versus WR DCV.

HS might be the way to go, but it's so much, will make a serious effort to keep it simple for simulation purposes.