The Wrong Left Turn and the Uncanny Valley

[Yardstick]

Originally posted by acrashb:

</font><blockquote>quote:</font><hr />Originally posted by Angryson:

I love the level of realism.

Like a squad partly-milling about in the street when they should be stacked at a corner?

Like a MG firing at the grid corner of a house instead of walking fire across the face?

Like shooting through walls?

Like a clump of T-72's just sitting around waiting to be plinked by M1's?

Read MD's original post - it is so well constructed one could call it erudite. </font>

[dalem]

Originally posted by Blackhorse:

Nice topic.

As someone else stated, command span and the potential for command span overload are definite pitfalls that challenge any wargame designer.

The Army believes a Command Span of 2-5 units is optimal.

This creates problems for designers. Unless the game is truly multi-player, then this span of control issue becomes problematic and has the potential to overload players. Adding RT into the mix further complicates the design process.

A game system that seems to have somewhat effectively addressed this issue is the Airborne Assault series. There, players can in fact give orders to just a few units, and those units will in turn transmit orders to their subordinates.

The game system is apples to oranges compared to CMSF however, so a comparison is not enntirely fair.

However, if CMSF were able to do something like that, would it change the nature of the issues people are having with it?

I think that if the level of abstraction shown matched the level of abstraction used, then the issue we're discussing here would go away. As it stands the game output tells me things that contradict each other.

-dale

[acrashb]

Originally posted by Angryson:

</font><blockquote>quote:</font><hr />Originally posted by acrashb:

</font><blockquote>quote:</font><hr />Originally posted by Angryson:

I love the level of realism.

Like a squad partly-milling about in the street when they should be stacked at a corner?

Like a MG firing at the grid corner of a house instead of walking fire across the face?

Like shooting through walls?

Like a clump of T-72's just sitting around waiting to be plinked by M1's?

Read MD's original post - it is so well constructed one could call it erudite. </font>

[GreenAsJade]

I'm a bit late to this excellent thread, and I agree with the first post, and where it leads.

I have just one thing to add:

I don't wish to appear overly smug, but isn't that the problem? Having a game that looks like it is 1:1, but isn't? I think at some point, there is a desire to have one or the other.

I think that there are *two* problems

1) Having a game that looks like 1:1 but isn't.

2) Whether a 1:1 game is even interesting and fun.

I think that #1 is clearly a problem for all the reasons give in the last three pages.

I think that #2 is also a big problem.

What does 1:1 add to the tactical or strategic play options or experience?

When I order 1-PLT A-Co to cover some barbed wire blocking a forest approach, I really don't care if Private Jones is picking his nose.

I don't want to give orders for individuals, because it is the behaviour of the squad I am planning. So why do I want to see the behaviour of individuals? What I want is a clear representation of what the squad is doing in response to my orders combined with the circumstances they are in.

And guess what. The 3-figure representation perfectly does that. If someone in the squad throws a grenade, I see a grenade getting thrown. I don't need to know which one of 10 men threw it.

I really do think that 1:1 introduces a fundamental mismatch in the abstraction and the gaming thinking.

GaJ

PS: I think the title of this thread is unfortunate, and would benefit from being changed. As has been observed, no-one thinks Steve is a whiner, so why imply it?

This thread is really about "The Wrong Left Turn and the Uncanny Valley".

[Michael Dorosh]

Originally posted by acrashb:

Firstly, faith is not a plan.

Second, these aren't bugs, they're design issues - especially when you read MD's original post, which you admit you haven't because it caused your head to explode.

Finally, you said you loved the realism, I pointed out some highly non-realistic elements, you respond with faith. Some people find it difficult to say "I'm wrong".

Well stated, acrashb. CM:SF has been stated to be the "cup of tea" of many people. I think that's great. I enjoy working in the editor, and I do enjoy playing the game. I also enjoy discussing design issues, which I've been saying for months now, and so now that the opportunity is here, I'm glad that there are well spoken people to do that with. Those not capable of participating in such a discussion are free to - well, go enjoy the game or something. We don't need to hear from them, because frankly, it is possible to enjoy the game at the same time as question some of its underlying assumptions.

And edit to say - well stated GaJ, and changed as per your suggestion.

[Blackhorse]

How do we reconcile the popularity of the Close Combat series games to this topic?

Weren't they the same, conceptually, in terms of 1:1 representation, generally the same number of units the player had to control, RT, etc.

What was different was MUCH Smaller maps and an overhead view allowing the player to see everything as well as a bar along the bottom that reflected the status of every unit, thus helping to deal with some of the span of control overload.

[Michael Dorosh]

Originally posted by Blackhorse:

How do we reconcile the popularity of the Close Combat series games to this topic?

Weren't they the same, conceptually, in terms of 1:1 representation, generally the same number of units the player had to control, RT, etc.

What was different was MUCH Smaller maps and an overhead view allowing the player to see everything as well as a bar along the bottom that reflected the status of every unit, thus helping to deal with some of the span of control overload.

No it wasn't. There were no doors and windows, IIRC, and the fidelity was easier to achieve. The vehicle pathing was still poor and heavily criticized. As you point out, the scale was smaller - units and maps both. And the map elevations were so hard to see, that LOS was just kind of assumed, at least by me, to be hit or miss at the best of times and part of the challenge!

You did have goofy stuff like chasing down single routed men and having to kill them at point blank range. That all said, I liked it because there was nothing better at the time out there. It was groundbreaking with respect to 1:1 modelling, but individual morale was actually tracked, and individuals split from the squad if they had seen too much. That too is not shown in CM:SF despite the 1:1. I understand the reason for the abstraction, but that just argues against 1:1 at the company level.

CC was a platoon level game at best in any event.

It's 3D version, GI Combat, was panned and went quickly to the discount bin. It's follow up, EYSA was more successful, in its second incarnation. I haven't followed it's current popularity - perhaps Eric can weigh in here.

[ August 11, 2007, 06:30 PM: Message edited by: Michael Dorosh ]

[thewood]

EYSA wasn't much better than SA. They just managed to sucker a few more people into buying it. TOW is really the successor to EYSA, for what that's worth.

[Yardstick]

Originally posted by acrashb:

</font><blockquote>quote:</font><hr />Originally posted by Angryson:

</font><blockquote>quote:</font><hr />Originally posted by acrashb:

</font><blockquote>quote:</font><hr />Originally posted by Angryson:

I love the level of realism.

Like a squad partly-milling about in the street when they should be stacked at a corner?

Like a MG firing at the grid corner of a house instead of walking fire across the face?

Like shooting through walls?

Like a clump of T-72's just sitting around waiting to be plinked by M1's?

Read MD's original post - it is so well constructed one could call it erudite. </font>

[Michael Dorosh]

</font><blockquote>quote:</font><hr />Originally posted by Angryson:

Bite me.

That said, I'm not going to cry because I cannot see tracers walking back and forth over the front of a building, as long as the ENY inside are suppressed, I give two ****s where the tracers go. I havn't had any problems with guys stacking or using cover, maybe I'm crazy but I have yet to have an issue. I've got nothing to say about the walls, BFC will correct it, thats what they do.

What part of that didn't you read? </font>

[Yardstick]

Sorry, I didn't realize I'm clearly not qualified to post in your thread. In the future I will ensure my posts are at least 500 words long and are free of slang.

Or maybe I'll just go play CM:SF and leave you guys to your groovy discussion about the history of wargames. How exciting.

[thewood]

Micael, didn't know you spoke American.

[JasonC]

Blackhorse - personally I couldn't stand Airborne Assault, just as a game. One because I don't like real time, but also for the reason you mention. The major control variables on the outcome are simply taken out of the player's hands and put under the computational hood, so to speak. As a result, whenever I played it I felt like I was watching a (bad, flat) movie that someone else was directing, and that the game designer was sitting in my chair.

Strategy gaming is not simply about simulation. Simulation is actually a mere means, and not valuable in its own right. Go works as a strategy game, without being immersive of course, but it works. Because the outcome depends on the matched wits of the players, and they are competing in a combinatorially huge space of tree-d out moves and counters to those moves.

Consider instead the real time 1 to 1 simulation of the battle of Guadalcanal that I have available in my thought-experiment back pocket. You get to spend 6 months playing the campaign scenario. It isn't pausible ("that's not realistic"), and you play the last 2/3rds shaking with malarial fever and soaked to the skin. As a concession to the entertainment crowd, we promise that you will not be dismembered by 14 inch shells or sawed in half by a Nambu. But your socks will really stink and you will be in mud and blood and chiggers up to your ears continually. None of your decisions will have the slightest effect on anything, other than perhaps what remains of your self respect.

There are some things we simply do not want to simulate.

Yes the army thinks the right command span is 2-5, which is a good example of the power of the KISS principle. But alas, 2-5 is too small for strategy gaming. To be blunt, the number of real decisions a typical officer makes in the course of a battle is not interesting. When you instead include the fact that 2 out of the 3 are made based on poor information and 1 of those is flubbed by somebody else before it has the slightest effect either way, we are left with a single probabilistically selected decision with impact for the whole action. And we'd rather have root canal on a sunny Saturday afternoon.

Tactics works as a strategy game context by multiplying the command spans under one mind's control, beyond a realistic level, but still within a managable one. That means something like 3 separate army levels of hierarchy - occasionally a 4th simulated but not controlled. That results in scores of items to articulate, and a number of sensible actions for each, at each stage, of the same order as a typical army command span.

Orchestrating that much is more control than all but the most finely tuned cooperative organizations typically achieve, when everything is going right. But it is enough to both challenge player's computational ability to predict things and to think through moves and counters, and to effectively put the player's decisions in charge of the outcome. Sure there is still chance, and sometimes it matters. But there are enough decisions that matter and are executed, that the overall outcome mostly turns on the player's matched wits (just like go or chess).

And that, not simulation, is the attraction and intellectual interest of the things.

It is also worth pointing out that some settings are more suited to strategy game treatment than others. When units are undifferentiated bags of combat power, only very broad patterns of fighting count, and the level of combinatorial richness is seriously reduced. The "paper scissors rock" relationships that combined arms give rise to, on the other hand, are "strategy rich". Every move has a counter, doctrine is meaty and meaningful, etc.

For modern or WW II combat, this means entities below brigade level, pretty much. For earlier periods it can be larger ("grand tactics" in Napoleonic warfare e.g. - the whole fight on one field of battle, but not the campaign of marches etc).

[Michael Dorosh]

Is that 2-5 limit tied to specific time frames, though?

With a turn based system (I realize TCP had optional time limits), you have as much time as you want, so the constraints become one of how perfect the information you are assessing is, rather than how fast you are transferring data within the system - no?

[konstantine]

Interesting stuff.

Curious after reading the thread; what's the suggested compromise? There are a lot of good points raised so far, but when it's all said and done, if BFC re-abstracts squads I'll feel like a five year old who just had his birthday cake stolen by a gang of evil clowns.

[DavidFields]

Well, an excellent initial post by Michael Dorash.

I must say I am surprised, however, at so much electronic ink being spilled over the 1:1 representation.

I would have thought that one of the main issues to discuss with CMSF was the issue of how Battlefront tried to cure "Borg Spotting". Given all the CM posts on this issue, and the overt lack of realism with the previous spotting, one would think there would be multiple topics/post from people wildly applauding the improvement.

So let me ask this question: Was the resolution of "Borg Spotting" actually such an immense improvement. Or, from a "fun", game design perspective, is its absence too much realism?

[Oh, and about whoever posted about Tobruck: it was a dice monster. And the problem was not helped by the rather bland desert maps. And a trivia question: What was the Avalon Hill WW1 naval game, at the ship level, that one played on the floor?--no map, cardboard tools to note ranges and fleet line turning.]

Edit: Right, JasonC, Jutland--I almost got up in the middle of the night to edit this post, when the name finally came to me...fortuntely, a burst of sanity allowed me to go back to sleep.

[ August 12, 2007, 04:58 AM: Message edited by: Rankorian ]

[JasonC]

Jutland.

[JasonC]

konstantine - one piece of a solution. If the tac AI is not going to succeed in finding cover effectively, then incoming fire impinging on the squad should be "probabilistically intercepted" by the cover in the squad's 8m tile, as though it were doing so rather well.

In other words and generalized, where the pure engineering approach breaks, substitute a "design for effect" approach that gives about the right outcome...

[JasonC]

Rankorian - borg was a problem mostly on the tac AI level. I do appreciate being able to tell what each of my units can see when I click on it, but since I can click around on the ones with the best view or de-select, of course I can still see the whole picture. This means the effects of being "un-borg" only appear through what the tac AI fails to do now, that it might have done if borg-ed. Since right now the AI is failing to do much of anything, this is not exactly prominent in the behavior seen...

[molotov_billy]

Originally posted by JasonC:

Blackhorse - personally I couldn't stand Airborne Assault, just as a game. One because I don't like real time, but also for the reason you mention. The major control variables on the outcome are simply taken out of the player's hands and put under the computational hood, so to speak. As a result, whenever I played it I felt like I was watching a (bad, flat) movie that someone else was directing, and that the game designer was sitting in my chair.

Though I wouldn't say that I can't stand it, especially Conquest of the Aegean, I do agree on the points in which it's lacking. It's a such a shame, too, because in terms of depth and realism it's absolutely lightyears ahead of anything else I've ever seen. It's the only game to come close to matching the experiences portrayed in operational level books that I've read. It exposes the same problem sets that real commanders faced on the ground.

Unfortunately, like you said, player interaction just isn't there. A single set of commands at the beginning of an operation can be carried out by your computer subordinates without so much as a further mouse click. Sure, it's interesting to see how things play out - but I'm not the one doing it. Further player interaction is actually punished by the realistic orders delay - there is no real way to adapt to the changing situation.

I'll buy every one of their games to support what they're doing, but I've been finding that I play them for no more than a few weeks before I move on.

[GreenAsJade]

I would have thought that one of the main issues to discuss with CMSF was the issue of how Battlefront tried to cure "Borg Spotting". Given all the CM posts on this issue, and the overt lack of realism with the previous spotting, one would think there would be multiple topics/post from people wildly applauding the improvement.

Hip Hooray. Honestly! It's great!

But... it's much more interesting to consider the underlying game design, and whether this is what's making CMSF less fun, than to sit around cheering.

If they could have added this borg cure to CMAK, and fixed up the dust and AI limitations, and added icons instead of bases ... we'd be all cheering wildly, some years ago, I reckon...

GaJ

[moneymaxx]

I think that 1:1 representation with an underlying 1:1 calculation, in theory, is superior to the highly abstracted CMx1 calculations.

To make my point clear I'd like to compare CMx1 with a "flawless" 1:1 system (and not CMSF at its present state).

In CMx1 you could kill an entire platoon in a heavy building (e.g. a church) by MG fire, because the probability to be hit is shared by all members of the squad. But that's not very realistic, most squad members should find perfect cover and while some unlucky guys standing at a window might be killed at the beginning, the rest should be 100% save for the rest of the engagement. CMx1 masks that shortcoming because it takes an insane amount of ammunition to do so and there is no graphical representation of that blunder (What does that squad do? Rotate squad members to an exposed position?).

A perfect 1:1 system however would simulate that most squad members find cover and further MG fire wouldn't do any more damage then blocking certain spots around windows. The squad wouldn't be entirely killed or routed like in CMx1.

For me comparing CMx1 and CMx2 is like comparing a matured technology with the possibility for only marginal improvements with a new technology that has much more potential but in its initial installment is inferior (like the first TFTs compared to CRT monitors, the first digital cameras, cars and carriages ....).

So for me the real question is if battlefront will be able to push the new engine beyond Cmx1, which would require many improvements that have already been mentioned in this thread. E.g.the TacAI must be up to the task to position the squad members according to threats and received information, otherwise all the theoretical advantages of an 1:1 system will be lost.

[GreenAsJade]

You don't have to have 1:1 representation to corretly model the response that you described of a squad in a building to MG fire on the building.

It would be interesting to find a concrete example where 1:1 representation gives you something you really can't have in abstract form, must have 1:1, and really do want to have in commanding a company of squads to do its job...

GaJ

[ August 11, 2007, 10:15 PM: Message edited by: GreenAsJade ]

[hoolaman]

Originally posted by Michael Dorosh:

</font><blockquote>quote:</font><hr />Originally posted by Hoolaman:

I disagree with MD's essay on the design of the game.

It's funny that you say you disagree with me, and then prove my point with everything else you say. </font>

[Battlefront.com]

Time to weigh in, I guess

First off, I do have to disagree with the notion that "abstraction has always worked, so why change it?" mentality. I've said it over and over again... CMBO broke more boardgame "abstraction" conventions than it retained. The ones it did retain were mostly out of pure necessity due to hardware constraints.

We didn't break abstraction conventions for the heck of it either, we did so to move computer wargaming beyond rules and design limitations that were inherently constrained by paper/cardboard and dice. I remember Steel Panthers guys rigidly defending the way that game handled ballistics. Tank has armor 5, you have gun with penetration 6, you get a good chance of killing. Why get more detailed than that with angles, strength of penetrators, velocity, etc? Keep it simple, keep it easy to understand and it is just as good. That was the logic we rejected.

We rejected it because an abstraction like ballistics had a lot of negatives that come with it. Mostly in terms of realism, but sometimes in terms of balance. Tanks in Steel Panthers, to me, never felt real. They felt like a cardboard counter with all the inherent limitations. Tanks in CMx1 and CMx2 don't feel that way because they aren't simulated so abstractly. And it pained a LOT of people to see this realism, especially the ones that were overly in love with German über tanks. Most of you here can attest to the reactions we got from people when they saw their Tigers killed by things like short barreled Shermans. "That never happened in Steel Panthers" was what we heard, amongst other things.

CMx1 was great because it was not arbitrarily bound to that which came before it. Hopefully nobody will disagree with that because I think a truer statement can not be made about the CMx1 series and its success. Going on the assumption that I've made that point clear, and its not in dispute, where does that bring us to with CMx2?

When we were making CMBB we ran into far more problems than you can realize with the game engine in its then current form. It was pretty much designed for Panthers and Tigers with infantry and ATGs in a West European setting. The engine was extremely brittle and sensitive to even small changes. Some of it were structural (i.e. the code wasn't easy to adjust), but some were conceptual. "Make Civil War, do Vietnam, how about Desert Storm, etc" was asked of us. Yet at the same time we were struggling just to shift fronts in the same war! In fact, we were so surprised by this that it took twice as long (2 years vs. 1) to make CMBB. Again, there were a variety of reasons, but one of them was battling the abstractions.

It was during this time of frustration and annoyance that we started understanding the design flaws under the hood. Abstractions were one of the prime limiting factors. It was right there and then that we decided we needed to ditch the CMx1 engine and completely start over. We even told you guys this at the time, stating that CMAK would be the last of the engine.

When we started designing the CMx2 engine we had to think ahead a lot so we didn't get ourselves stuck in the same boat we found ourselves in with CMBB (i.e. locked into an inflexible design). We don't get paid enough to sit around for 3 years to develop new engines and then get 3 games of declining sales numbers off of them, so we had to do the next engine to last a lot longer and be much easier to work with. As I've said many times over, CMBO sold the most, CMBB sold significantly less, CMAK sold even less than CMBB. You guys can have rose colored glasses about how great our past games are, but we can not. A business based on putting in inordinate amounts of time and expenses to get less for each hour worked is going to fail. We don't like failure

OK, so there we are... 2003/2004 and realizing that the old engine is dead and we need something that will take us to about 2010 without major reconstruction, 2015 with portions replaced as needed. Remaking CMx1 was never considered, but we also recognized we didn't want to abandon it either. So we made the overall design goal to reduce the abstractions and that meant going 1:1. This served two purposes:

1. Eliminated many of the issues that prevented us from being able to cover different subject matter efficiently

2. Gave us an opportunity to start out ahead of the market curve instead of behind it as CMBO was (the novelty was not repeatable)

If you look at that that year 2000 thread Dale linked to you'll see that I posted four objections to going with 1:1. So what changed? Here are my four points from then and my answers as they would have been from 3-4 years later:

1. Today's hardware, even the fastest systems, would have a hard time with full 1:1 representation. So much so that people would turn it off because the framerate would be so poor (at least for medium to large infantry battles).

Very true for 2000 when the average person was having trouble getting unmodded CMBO to work at acceptable framerates. Hard to imagine, eh? But by 2003 we saw that there was a long awaited boom in processor speeds and a rapid decline in their price, accompanied by better video cards, RAM, and other features that were also becoming quite inexpensive. Therefore, what was true in 2000 was already no longer true in 2003/2004. Projecting forward a few years was easy to see 1:1 working.

2. The people with the best hardware ALWAYS make up a small percentage of the customer base. In our case, an even smaller percentage since wargamers are generally slow to upgrade.

Not as true any more due to faster hardware becoming relatively inexpensive. My 2007 near top of the line $2000 Mac Laptop would have cost me $4000 in unadjusted 2000 Dollars. Adjusted it would probably be around $4500 (rough guess).

3. Without putting in all sorts of very complicated and time consuming game features, 1:1 figures will actually detract from the gaming experience. This will make the people that requested it turn it off and demand that we then add in the features they previously said they don't care about.

Still true to some extent, but people's expectations changed over the years. At first people were just floored by CMBO and were willing to overlook, to some extent, the abstracted depiction of soldiers and terrain. There were hardly any other large scale outdoor 3D games at the time, for example. As the years went on we saw that the number wanting 1:1 was increasing as hardware increased the possibilities of doing it and other games actually did do it. By 2003 we were convinced that we had to go 1:1, and therefore had to make the sacrifices necessary to simulate it. In our case it probably came at the expense of adding one year to the development process.

4. Putting in any major feature like this takes time. Time that is not spent doing other things. Things that have far more direct impact on the gaming experience AND (most importantly) are practical, positive additions to the game for EVERY customer.

Partially covered in my previous comment, but it goes further than that. After I made the 2000 statement we started working with the engine to make CMBB. We thought it would be easy, but we found it to be very difficult. As I stated further up in this post, a lot of the reasons had to do with the very nature of the abstractions. Fewer abstractions would have made the move to another theater, not to mention another epoch, much easier. So AFTER I said that it wasn't worth doing we found out that wasn't really the case. In fact, we found the opposite to be true.

To sum this all up:

Based on what we saw at the time we started designing CMx2, and what we predicted the future would hold for us, we viewed avoiding 1:1 as being a suicidal strategy. How do we feel about it now? Our convictions are even stronger that 1:1 was the right way to go, even with the teething problems. We can get past teething problems, we can not get past a fundamental error in judgement. Ironically, some of you see things in the exact opposite way. Since you are not standing in our boots that's a bit too easy for you to say since you're livelihood doesn't depend on it.

We came to a junction in the road and saw two possible turns. The road we took wasn't the easiest one to take, but the other one deadended into a brick wall. Perhaps because you're in the passenger seat you couldn't see that wall, but we in the driver's seat did. We took the right turn, of that we have no doubts.

Steve