WEGO, RT and the blue bar that gets no lovin'...

[thewood]

I think the issue is that it adds time. In CM1, the typical scenario turn burned through the blue bar in 10-15 seconds. After that the replay started and if there was no activity, you just FF to the end of the replay immediately. So at the beginning of a scenario, you could get through a full turn in 20 sec.

Now you have to wait for the full 60 sec. RT segment to play out then maybe skip ahead. That means probably 65 sec. minimum to get through a turn. I would guess it takes me 2 to 3 times longer to get through a WEGO game compared to CM1.

[c3k]

Agreed.

Bigmac1281, the crux of the matter is that WEGO now works like RT. With mandatory pauses at 60 seconds. And a prohibition from intervening with commands unless at the mandatory pause.

The obvious question is, "How is what you just described different from CMx1 WEGO?" I'm glad you asked that...

In CMx1, after you plotted your commands, the execute button would freeze the screen while a blue bar, denoting computer progression, advanced. The computer would run all the calculations needed for the upcoming turn. It would do so using AS MANY COMPUTING CYCLES AS NEEDED. If it only took 5 seconds, great. If it took 3 minutes, great. However complex or simple the turn was just didn't matter.

Now, it seems that the computer MUST run the game in RT. Then you can rewind if you want.

The inherent assumption in this is that the computer MUST calculate all the turn's actions on the fly. This, despite all the computational advances, is inherently less capable than if the turn were given all the resources of the computer for as long as it needed. That begs the question, "What is being left out or short-cut?", and the corollary, "Does it matter that things may be getting left out or short-cutted?"

That is the crux. To me.

Thanks,

Ken

[dalem]

But I thought that whole "lack of blue bar = crappy grafx/LOS etc." was shot down by the BFC guys a couple of months ago quite convincingly.

-dale

[c3k]

dalem,

Hah! Don't listen to THEM. What do they know about this game?

Perception IS reality.

I have found that my enjoyment of CMSF, such as it is, has been increased by placing a blue strip of tape across the bottom of my monitor. Sweet joy!

Regards,

Ken

[Guest]

Heh. Blue tape. That's actually a good one. Maybe I'll sig it at some point in time.

Yeah, it was pretty convincing, dalem. I think we'd find people would be complaining about three minutes of blue bar for every one minute turn if BFC had actually gone with 1:1 everything.

[Guest Guest]

deleted per user request

[Guest]

WEGO is actually a fun paradigm to play around with. It certainly represents planning in a much less superficial way than IGOUGO (to my mind).

I wrote a simple company-level wargame a few months ago, no graphics unfortunately, that was primarily an experiment in WEGO. And it was fun, mostly *because* of WEGO.

I could throw some while loops in there to make the turns last ten minutes.

[SlapHappy]

Originally posted by Phillip Culliton:

Heh. Blue tape. That's actually a good one. Maybe I'll sig it at some point in time.

Yeah, it was pretty convincing, dalem. I think we'd find people would be complaining about three minutes of blue bar for every one minute turn if BFC had actually gone with 1:1 everything.

3 minutes? That's nothing. Oh wait, that's 90 minutes of overhead just for calculations for a 30 minute game.

Still, I would have thought it would be more like 10 minutes per turn the way Steve was describing it. I guess a lot of variables would contribute to the actual calc time.....

[SlapHappy]

I still think my Q6600 quad-core would make short work of the number-crunching if this game were 1:1 WEGO-only.

But making a game to run on the top 5 processors available on the market may not be the best marketing decision one could make.....

[Guest]

It might well be more like ten minutes per turn. I'd like to think that Steve and Charles looked at that problem pretty hard before going the way that they did, but only they know all of the variables for sure.

I can tell you that it would be a significant increase in computation time. Given the amount of AI/LOS/LOF CPU time that I currently think goes into a 1-minute turn, it could be quite a bit.

And if it broke a minute per turn, suddenly the current solution wouldn't look so bad, I think.

[acrashb]

Originally posted by Captain Adultery:

the blue bar would mean that you could ditch all the LOS/LOF problems because every calculation that needed to be done could be done???

and no horsepower trying to do graphics as well??

then the replay could use all the horsepower to render the turn without bogging the CPU with work??

In order: yes, yes, and yes.

[bigmac1281]

Originally posted by c3k:

Agreed.

Bigmac1281, the crux of the matter is that WEGO now works like RT. With mandatory pauses at 60 seconds. And a prohibition from intervening with commands unless at the mandatory pause.

The obvious question is, "How is what you just described different from CMx1 WEGO?" I'm glad you asked that...

In CMx1, after you plotted your commands, the execute button would freeze the screen while a blue bar, denoting computer progression, advanced. The computer would run all the calculations needed for the upcoming turn. It would do so using AS MANY COMPUTING CYCLES AS NEEDED. If it only took 5 seconds, great. If it took 3 minutes, great. However complex or simple the turn was just didn't matter.

Now, it seems that the computer MUST run the game in RT. Then you can rewind if you want.

The inherent assumption in this is that the computer MUST calculate all the turn's actions on the fly. This, despite all the computational advances, is inherently less capable than if the turn were given all the resources of the computer for as long as it needed. That begs the question, "What is being left out or short-cut?", and the corollary, "Does it matter that things may be getting left out or short-cutted?"

That is the crux. To me.

Thanks,

Ken

I didn't know any of that. Thanks for the explaination, had no idea what was going on under the hood.

[JonS]

Originally posted by c3k:

... WEGO (THE UNIQUE SIGNATURE OF THE CM SERIES WHICH WAS GROUNDBREAKING AND SET UP BF.C'S REPUTATION!!!!!!) ...

WEGO had been done in a computer game before CMBO came out. Years before CMBO came out. Like, 6 or more years. Did you know that, or are you just making stuff up?

FWIW, I only play CMSF via WEGO, either against the AI or via PBEM. I'm happy with it, and happy I don't have to watch the blue bar.

I get the "why do I have to wait 60 secs before I can hit rewind" thing ... well ... I sort of get it. But it seems very thin beer to be complaining about :confused:

[BFCElvis]

OK...This thread was just brought to my attention by a friend and I have 2 comments about it.

First, is Para theory that because the whole minute isn't procesed before you watch that the result is somehow watered down and not as complete. This fails to take into account that CMBO is now over 8 years old (from a serious testing point of view. I know the game was released 7 1/2 years ago). Is it possible that computing power has progressed enough in that time that not only can the same level computations be completed "on the fly" but actually many many more? I remeber bragging about my 500MB processor around that time. and 64MB of RAM or something. Think about it. Think about what can be done now that couldn't be touched in any previous version of CM in the processing of action both turn based and RT.

Second, and I think this is the harder one to really grasp, has anyone seen any war games lately that offer both turn based and real time? Hell, any programing company with half a bit of forethought would try to program allowing both turn based and RT. But I only know one company who did.

I don't post often out here but this one just has me shakin my head in disbelief. Like the game ...don't like the game...fine. It may not be for everyone....but this? Come on...There are better axes to grind. Like the mandatory turn numbering for PBEMs. Now that is something that offends me. I've got to page trhough 5 pages of turn files to get to the game file I want to load......Maybe I'll save that for a different thread.

[thewood]

The main gist of my thoughts on WEGO vs. RT is what you made a comment about. Can you do RT and WEGO as completely seperate game processes in the same game? You either are good at RT, good at WEGO, or mediocre at both. I think BFC had to base the game on one, and make the other one fit. That's how it feels when playing both.

[hoolaman]

Originally posted by Elvis:

OK...This thread was just brought to my attention by a friend and I have 2 comments about it.

First, is Para theory that because the whole minute isn't procesed before you watch that the result is somehow watered down and not as complete.

Although the thread has subsequently been hijacked in that direction, Para never actually said this.

All he said (and I agree) is that not being able to fast-forward the video on the first playthrough is annoying.

Trivial maybe, but worth a rant if you ask me. I have ranted about it months back too.

I agree that the whole "blue-bar makes for a better simulation" argument is pretty dubious without knowing how the game is programmed and how much processor power goes to graphics.

[Battlefront.com]

Philip is correct... we've had this discussion before. Er, more than once too

The "Blue Bar Brigade", as someone jokingly labeled the pro-blue bar folks, do have a point. I do not disagree that it would be a good thing to have the ability to not watch a turn if the player doesn't wish to. No problem there at all, even though I for one don't understand the passion behind the request when taken in the context of the game itself.

The problem is that, as Thewood points out, we had to choose to make the engine inherently WeGo without ANY chance of RealTime, or inherently RealTime with a perfectly viable WeGo. Since one choice was vastly more inclusive than the other, and far more marketable to boot, the choice was stupidly easy to make There was no hesitation at the time and there are no regrets about it now. It was the ONLY viable decision to make and so that's the one we went with.

As is very typical, when a player sees a feature missing he only sees the glass as half full. The things gained from whatever led to the feature being left out are completely ignored at best, viciously attacked at worse. Putting aside the obvious "I hate RealTime, therefore everybody should hate RealTime, therefore the game should be WeGo only" line of thinking (because it is arrogant and rather insulting at best), there are other advantages to the RT game engine.

Many of the oddities and limitations inherent in CMx1 were the result of taking WeGo as far as it could go without it being RealTime. Things like not being able to track the actual flightpath of a shot, which resulted in things like tanks being shot up after moving behind cover. Artillery shells falling to hit the ground AFTER the 60th second passed. Not being able to have viable TacAI memory from one turn to the next. These things, and probably hundreds others that Charles could come up with, collectively presented us with a choice of either arresting CM's potential or allowing it to go into new territory. Not just now with the first release, but in subsequent releases.

So yeah, there are many things that were gained by going with a RealTime engine even if the game is played in WeGo mode. The problem is that they don't have blue bars so it is a little harder to notice them

I would also remind people to stop comparing apples to oranges. Turn crunching routinely took more than 60 seconds to perform on the hardware the software was originally played on some 6-7 years ago. I remember Rune making a scenario (Battle of the Bulge) for CMBO that took nearly 20 minutes to crunch a single turn. And I had a better computer than most people who bought CMBO! So I have no choice but to dismiss the notion that CMx2 on today's hardware would play as fast as CMx1. It's simply a nonsense argument.

Could CMx2 have a lot fewer abstractions and more time for AI and still crunch turns in less than 60 seconds? Over all, not much. For sure not the sort of fantasy concepts I've seen tossed around about how far we can go with 1:1 with present hardware. There is nothing wrong with wishful thinking unless it is mistaken for informed expert opinion. I've been over that ground enough times I don't wish to go over it again. Simply put, we know what we're talking about, you don't... so our opinion is of more relevance than yours if you think CM can do a fantastic number of extra things at no additional cost (I'll not even bring in programming time that would be needed and that we don't have, because that's killing a dead argument with a nuke ).

So that's really that. There will never be another Blue Bar in CM ever. It's therefore useless to lobby for one. So don't.

One option that Charles and I have discussed doing is compressed time. In theory it might work, but we don't think it will. Two problems:

1. Most games that use time compression are fairly straight forward sims, usually vehicle sims. Often they have known "worst case" scenarios that can be used to benchmark tradeoffs between visual performance and number crunching accuracy. Combat Mission is too open ended for that, therefore we don't think we can gauge this in a way that would work well.

2. In some situations it is impossible to free up CPU cycles because what is going on within the game is so intensive that all resources are already maxed out. Such situations are not only inconsistent from one block of 60 seconds to the next, but also within those 60 seconds. Forcing the game to "hurry up and move on" means shortcutting on the calcs which almost always leads to problems. Shots not going where they should, etc.

We'll keep the possibility of time compression open, but if we don't feel we can make it work it won't get added. Plus, you need to keep in mind that CMx2 will continue to use up computing resources as they become available for new features, therefore "crunch time" won't go from several minutes down to several seconds like CMx1 did. That CMAK engine was basically what we released 3 years earlier, so yeah... it was a lot faster because it wasn't really doing anything new. Uhm, which is why we chucked the code and moved to CMx2

Steve

[Mud]

On a related... tangent... of idle thought,

...would it be possible to go the other direction, and voluntarily run through action (in RT) at slower-than-realtime? This might be useful in particularly complicated situations, so the user isn't quite as hampered by reaction time, without the stutter of rapid pause/scroll/unpause cycles.

Might also be more pleasant than full pausing in MP, as long as it's done in intervals of some length, perhaps limiting the total amount of slow-time per game. I'd prefer to let the MPers judge whether or not it'd be useful for MP, 'tho.

[Battlefront.com]

Yes, I think it would be possible to slow down the action. It's not something I think has been mentioned before, so I don't have a good sense if it would be all that useful to people, but I think technically it can be done.

Steve

[Barleyman]

Since we've got Comms back online, there are some excellent suggestions made vis-a-vis UI here.

I mean, just in case Steve doesn't wade through to page 5 of every bitchin' thread in the forum..

Basically, if you want to go toward RT(s), you should ste.. *ahem* study some of the UI innovations made in the genre during last decade or so.

One additional feature of the blue bar era that would be nice is the ability to replay a bit, especially if you get pause on events -function (such as abrams getting smoked) to see wtf happened.

[c3k]

JonS,

This is a forum. On the internet. About a game.

I certainly do not let niggling things such as "facts" get in the way of my posts. If there is something I posted that you think may be erroneous, please give me the Wikipedia entry in question and I'll be sure to bring it into line.

Regards,

Ken

[DaveDash]

I have to agree with Steve actually about time compression. I remember playing Falcon 3.0 and if you'd speed up time all sorts of weird AI behaviour would occur.

Don't really want that in CM type games at all, to be honest.

Interesting to hear about the limitations of precalculating. It seems what has been done now is in fact more optimal than the past.

Cheers.

[Other Means]

How about TC until contact?

[DaveDash]

Hmm have you seen vehicles try and navigate when in close proximity in normal time?

Image the inaccuracies at speed. They'd end up like the old C&C harvesters who went out of their way to go through the enemies base before returning to yours.

[Steiner14]

Good, that these aspects were raised once again:

Many of the oddities and limitations inherent in CMx1 were the result of taking WeGo as far as it could go without it being RealTime. Things like not being able to track the actual flightpath of a shot, which resulted in things like tanks being shot up after moving behind cover.

IMO that's not a valid argument, because that case was extremely rare and when it happened, i just thought: the tank was one or two seconds to slow. So that caused, at least for me, not the slightest problem.

Now we have bullets regularly flying through terrain.

So where is the advantage?

Artillery shells falling to hit the ground AFTER the 60th second passed.

Not perfect, sure, but no deal breaker, too.

You are exaggerating not so perfect things in CMx1, to make CMx2 look better, imo.

Not being able to have viable TacAI memory from one turn to the next.

From a programmer's point of view, i do not understand that. Unit memory is an aspect to give units a memory, where they can store the kind of threat, it's position, direction, time, severity,...

Why should WEGO forbid to give units a memory? It's only additional storage-space. It was not possible to expand CMx1 in that way, due to code restrictions. But now argueing that RT is the prerequisite for unit-memory, does not seem logical to me.

[ December 04, 2007, 05:25 AM: Message edited by: Steiner14 ]