Saturday, November 29, 2008

Seeing The Future: Thoughts on Combat Results in Tactical Games

Stephen B. Patrick presented some detailed thoughts on Combat Results Tables (CRT) in the February 1972 issue of Moves magazine that are interesting not only in their ability to briefly summarize their history but in exploring ongoing issues as today's game designers - both board and computer - continue the quest to best marry playability and realism in a single vehicle. I presented some thoughts on the subject in November 2008 on another site, which I'm presenting here in revised form. ("Professor Professorson" just found a link to the entire archive of back issues, incidentally, which Greg Costikyan uploaded to archive.org - read the details here.)

At the time the article in Moves was written, wargaming at the tactical scale in board games was in its infancy though miniatures rules had been promoted by pioneers in the hobby such as Jack Scruby in the U.S. and Charles Grant in the U.K. for many years. While Avalon Hill's PanzerBlitz contained many innovative concepts compared to the standard fare since board wargames first appeared on the market in 1958, such as isomorphic mapboards and a multiple scenario format, the method in which the game produced combat results remained unremarkable. When SPI began producing tactical games - including Soldiers: Tactical Combat in 1914-15, and Grunt, set in Vietnam, the CRT was similarly - speaking purely from hindsight - uninspired.

James F. Dunnigan defines a CRT in the 3rd Edition of his Wargames Handbook as

A Probability Table that shows the possible results of all combats allowed within a particular game. The greater the ratio of attacker to defender strength, the higher the chance of success. Because so many things can go wrong during the combat itself, a die or other random-number generator is used to determine the actual result. These tables are usually calculated based on what information is available on actual historical losses.

Stephen Patrick noted the trend in games in the late 1960s and into the new decade of the '70s was to simply re-use CRTs from game to game. In his article in February 1972 he identified correctly "the touchstones of authenticity and playability" and how the two concepts inter-related:

One can start with full authenticity and back off far enough to gain playability, or start with a purely playable system and work toward realism by adding the elements of historicity to give the right flavor. There will be a gray area where the playable takes on the flavor of war and where the war becomes playable. Moreover, this point will differ depending on the point of origin.

He then contrasted the Avalon Hill approach to games with the SPI approach; he contended that Avalon Hill's "playability" perspective simply produced games with similar rules for every game while SPI produced more historical games with tailor made rules sets. As an example of Avalon Hill's devotion to playability, he cited their CRT, which was a standard in their line of games to that date:

A - Attacker back 2
D - Defender back 2
Elim - Eliminated


When Strategy & Tactics began to publish tactical games (and it produced tactical titles outside of the "modern" genre on which I focus my attention), they similarly retained a common CRT with simple results in the platoon and company level games of the time:

"No result"
"Elim"
"Dispersed"

Patrick's thesis was that this "tactical Combat Results Table is the most archaic element in S&T's bag of tricks - the most playable/non-realistic element currently in use."

What Does It Mean?

Simply put, the CRT delineates the results of combat, and Patrick suggested that any action in which two opposing forces meet can result one of a limited number of results at the end of a fixed period of time.

* Melee (both forces remain locked in battle)
* Attacker repelled
* Defender repelled in good order
* Defender routed

Patrick noted that there were other possibilities; a pyhrric victory in which the attacker was severely damaged in the battle, for example. Tertiary considerations were fatigue levels, whether an attack was an initial action, a continuation of a previous attack, or the end of a battle. His main question, however, was how to transpose the basic CRT results to a game such that it adequately represented the history being portrayed.

The easiest result to simulate, according to Patrick, was the "no result", and a "dispersed" unit he felt was better described as "shaken" - temporarily unable to fight. He felt "eliminated" was draconian, as

...few battles result in an entire unit being destroyed to a man in a given time period, particularly during the brief period of time portrayed in the tactical games. Thus, there must simultaneously be some way to reflect the decline in strength from being in the thick of it and, at the same time, to get units off the board. After all, pasteboard pieces don't really have morale or take losses, so something must be injected to bring the authentic within the realm of playable. The 1914 solution of stepped units is obviously the best way to reflect casualties short of going the bookkeeping routes. But even (this) is viewed with displeasure by some (and) requires the injection of a whole set of pieces. The object here is to consider the requirements of a Combat Results Table which can be inserted in any game without actually having to totally revamp the rules.

Another solution was to add an increased dispersion rule, whereby a second retreat caused elimination, and a third option discussed by Patrick was to consider two retreat possibilities - retreat and rout - and have routed units equate to eliminated for purposes of the game.

Patrick also talked about using different CRT for different phases of the battle - for example during the initial phase of a battle when morale of an attacker was high, and again when units were tired during the last phase of a battle. In effect, he felt a battle might need three separate CRT to adequately model the distinct phases of a battle. He felt that not only fatigue, but fanaticism and the effects of good training would make themselves felt in the latter phases of a battle and should be reflected in the game mechanics.

Theory and Practice: Soldiers

Tellingly, in Moves issue 4, in August 1972, detailed articles on the development of the game discuss the history of the development of infantry and artillery, show images of the different drafts of the map, talk about rules development and scenario drafts, but have no discussion of the CRT. It's not known if a non-standard CRT was ever even contemplated.



Predicting the Future

Patrick ended his article with the following:

Returning to the real life situation...if the research is good (the result) should be a Combat Results Table which can complement the accuracy of the rest of the rules in evoking the period in question. The obvious point, though, is that the Combat Results Table now becomes an integral part of game design, rather than a handy plug-in section, such as the initial description of the pieces and the game map, and it is as important to make the Combat Results Table valid as it is to calculate the Attack Strength of a crossbow.

Logical Outgrowths

CRT development stagnated in tactical games at the squad and platoon level; two years later,Tank! still had simple odd-ratios driving the results of the CRT, though there were now panic results. Game development was focusing on whether play should be simultaneous movement or sequential and the CRT was still being viewed, perhaps, as simply a given.

When man-to-man games like Sniper! and Patrol came along, however, their very nature caused further development of the CRT as there were a greater number of weapons systems in play.

Which brings us to John Hill. He viewed the possible combat results in Squad Leader as still a fairly simple proposition. Despite the fact he was contemplating what would be an enormously ambitious and complicated game system in which multiple weapons systems would interact, he argued that it didn't matter if a squad of ten men were machine-gunned in the open, shelled moving through woods, pelted in their foxholes with grenades or burned out of a bunker with a flamethrower, the results would be the same - they would be killed, they would suffer some form of morale loss, or there would be no appreciable impact at all. "Design for Effect" became the mantra for Squad Leader's development, and was used to explain away inconsistencies in the design, wherein European streets became 80 metre wide boulevards, and physically fit men could only move 160 metres in two minutes. He simply "factored in" grenades as part of "close combat" and "point blank fire" and did away with the need for special rules or counters for them altogether.

PC Games

The first Combat Mission titles remained partially faithful to the notion that players wanted to see CRT results; while there were no visible dice rolls or interventions of fate, there was lip service made to such things as "fanaticism", and moreover, firepower stats were presented in unit information screens, and was available in the game via mouseclick during the orders phase, as was cover stats for infantry units, morale and fatigue levels (though not necessarily the explicit effects of same), as well as detailed armour value for AFVs, general penetration capabilities of weapons, blast values for artillery, etc.



Firepower and cover stats were available right in the main game space of Combat Mission: Barbarossa to Berlin.
While there was no longer a CRT for the player to refer to, enterprising players could recreate specific battlefield phenomena in the editor to determine probabilities - if he was curious how often a specific tank type would bog in a particular type of terrain and weather conditions, he could create a sample scenario and run it repeatedly until he had a sample to estimate from. Websites and collections of forum postings with links to just "research" have taken the place of the CRT in some cases. A "Player's Guide" was released for Combat Mission: Barbarossa to Berlin with tables of unit stats, which players could look up. The handbook was not billed as a "scenario designer's guide" - though it did contain interviews with scenario authors as an appendix and the data was likely aimed as much at them as players.


Hidden Outcomes and the Deletion of the CRT:

Given everything that was said in 1972 about the importance of having a valid CRT, there seems to be a trend in video games to keeping game routines hidden from the gamers who play them. The obvious desire is for "realism" and the common argument is that real life commanders "don't count firepower factors." Nonetheless, the player has to have a way to relate to the game, which has to use mathematical equations and logarithms to simulate results. Having access to the data increases understanding of how the data works - and the more "realistic" the simulation, the less likely the player is to have access to the data. In Panzer Command, for example (based on the Panzer War miniature rules), players have the ability to modify unit data if it doesn't fit their perceptions of reality, though the data isn't easily accessible inside the game (altering it is done via editable "xml" files - text documents which are loaded into the "back end" before the game is started). In the second generation Combat Mission game engine, where small arms and tank fire is ostensibly tracked by a real world physics engine, there is very little way for players to anticipate "hit chances" or probabilities beyond very general assumptions regarding terrain and situation - which is exactly what the developers intended. They would argue "CRTs" are a wargame construct, and that their wargame should be devoid of them!

My questions to you:
Patrick starts his proposition with "if the research is good". How much is actually "knowable" about what goes on at the tactical level, that would justify things like firepower factors or combat results to begin with? And must the designer choose between accurately modeling the proceedings (tracking every exchange of gunfire with precision) or the outcomes (10-25% killed in every average engagement, 25-50% wounded)?


This screenshot of Tigers Unleashed was unveiled on another gaming site. Do wargames really have to have hexes and counters in order for players to be able to reasonably access detailed data about their own troops' capabilities?

Addendum: in the comments to my original article, James Lowry noted that Anzio's CRT permitted step-reduction results as early as 1969. Advanced Squad Leader added a form of step-loss results to the CRT as well, an extension of a game function started in the original SL game series.

2 comments:



  1. Originally posted at gamesquad.com. The comments there read:


    Scott Tortorice - 29 Nov 08 18:00
    As usual, this is another thought-provoking blog entry from you. Personally, I am a big fan of the Keep it Simple principle. I think John Hill has the right idea because, no matter how you slice it, you cannot really ever have a hyper-accurate CRT that is able to incorporate every conceivable factor during combat - at least, not without making the game absolutely unplayable. Given that inherent limitation, I believe the best solution is one that produces not incredible results with a minimum of fuss.


    Michael Dorosh - 29 Nov 08 18:07
    I don't want to harp on the new Combat Mission engine - we do a lot of that in the CM forum - but it kind of baffles me, given your comment here, Scott, why they then feel the need to track individual bullets in a squad-based game. I mean - there are man-to-man games that don't do that! Unless it is simply an empty claim.

    The CRT definitely became more layered; I had some pictures I didn't post of other games - Trenchfoot, which was an incredibly simplistic man-to-man game of trench warfare, had split out different kinds of combat from each other, for example, and while Squad Leader had the very unique Infantry Fire Table to handle grenade, HE and small arms combat all in one CRT, they still had separate tables for tank combat, and for hand-to-hand, they still resorted back to odds calculations based not on morale, oddly, but on firepower!


    scrub - 29 Nov 08 19:40
    Give some wargamers a CRT and you'll get gamey tactics up the wazoo. Squeezing just one more FP factor from somewhere just to make an arbitrary number irks me to no end.


    Michael Dorosh - 29 Nov 08 19:43
    That's a great point I wasn't thinking about, but you're obviously right. It comes down to the realism/game playing divide again.

    Scott Tortorice - 29 Nov 08 20:51
    I don't want to harp on the new Combat Mission engine - we do a lot of that in the CM forum - but it kind of baffles me, given your comment here, Scott, why they then feel the need to track individual bullets in a squad-based game. I mean - there are man-to-man games that don't do that! Unless it is simply an empty claim.

    I hate to say it, but I think that is how wargame companies are compensating for a lack of modern graphics and design elements. In other words, PC wargame companies can't wow their customers with eye-candy or robust multiplayer components, so they do the one thing that they are able to do: use the number-crunching capabilities of the modern PC to the utmost.


    Egbert - 29 Nov 08 21:26
    But what is wrong with using the number crunching to the utmost. It's what computers are good for. You can take a very complex CRT and make it invisible to the player, making it more playable.

    The more invisible number crunching we have for CRTs, the closer we can step to reality while keeping ti playable.

    ReplyDelete
  2. gamesquad comments continued:

    Michael Dorosh - 29 Nov 08 21:35
    The more invisible number crunching we have for CRTs, the closer we can step to reality while keeping it playable.

    But that's what I was getting at, and I think is what scrub is getting at what his comment. Are they the same thing (playability/realism)?

    The article in 1972 drew a very heavy dividing line and said you can have one or the other.

    scrub is saying you can either count firepower factors, or make realistic decisions.

    I think that invisible number crunching as in the new CM would tend to take a game more into sim territory - "realism" - but that doesn't necessarily mean it would be more "fun" or make it a "better" game insofar as there would be a lot of frustration at the user input level as far as knowing what the difference between a BAR and a Bren Gun would be in determining possible outcomes on the enemy, or why you shouldn't run really fast before shooting at someone - or if, indeed, doing such a thing would have any appreciable effect in the game at all. Real world experience would tell you it should, but for all you know, fatigue isn't modeled in any given game unless it is included in the user feedback somehow via a chart or screen or icon.

    Egbert - 29 Nov 08 22:46
    Your point is good if counting each and every bullet is actually closer to realism. It is, only if it is relevant (to a sniper it would be) and the rest of the environment is up to the challenge. Otherwise it's merely an advancement in programming.

    Using that number crunching and administrative tracking to do something else, such as improved morale adjustments and keeping the player from having to do all the modifyer tracking (such as ASL) and becoming a master of the rules (ASL again), while allowing them to actually play, then you've moved back to the playability realm.


    Rindis - 30 Nov 08 11:59
    I think that invisible number crunching as in the new CM would tend to take a game more into sim territory - "realism" - but that doesn't necessarily mean it would be more "fun" or make it a "better" game insofar as there would be a lot of frustration at the user input level as far as knowing what the difference between a BAR and a Bren Gun would be in determining possible outcomes on the enemy, or why you shouldn't run really fast before shooting at someone - or if, indeed, doing such a thing would have any appreciable effect in the game at all. Real world experience would tell you it should, but for all you know, fatigue isn't modeled in any given game unless it is included in the user feedback somehow via a chart or screen or icon.

    And this is precisely the point I've harped on on occasion. The trend towards throwing more and more numbers at a game isn't, in itself, bad. But the fact that the user cannot easily know what is and isn't included in the model is no end of frustration, for me at least. Egbert's point that 'more reality' isn't the same as 'more realism' is one I wish more computer game designers would realize. I think we're still in the stage of them including things because they can, not because they should.

    And on a gaming history note: AH started playing around with the CRT in 1969. The original edition of Anzio featured step reduction (which made the game somewhat unpopular as you had to replace counters with lower-strength versions of themselves). So while it was a d6 odd-based CRT, it featured separate results in each cell for weather one side had to retreat, and how many losses they took. And yes, it was possible for both sides to take losses.

    ReplyDelete