Can we have an AAR from CMC ?

[Urban Shocker]

...just a wafer-thin pamphlet would be satisfactory.

[Bamse]

*bump*

[jim crowley]

Originally posted by Moon:

Yes, good idea and definitely possible, but it will take a little bit to write up something...

Martin

I'm glad it will only take a "little bit". I dread to think how long it would take for a "big bit"

[Slappy]

Of course, it will be ready when it's ready.

Seriously, I'd love to have an AAR, even a turn or two in the middle of an opp/campaign to get a feel for how things flow.

In the interest of full disclosure though, I'll buy the damn thing either way; just the way I bought all the others sight unseen.

Throw me a bone or somefink.

[Steiner14]

Announcing it and then silence for weeks is not really nice.

[Salkin]

Battlefront isn't here to be "nice". They are here to bring us kick-ass games !

//Salkin

[ December 17, 2005, 06:22 AM: Message edited by: Salkin ]

[Bamse]

It would be nice too have some kind of response....

[KNac]

Originally posted by Steiner14:

Announcing it and then silence for weeks is not really nice.

Who announced it?!

The only thing I can read is "it would take a little bit to write something like that", but that's not an annoucement, more the contrary hehe

[Tactical Wargamer]

Originally posted by Bamse:

[qb] Remember the CMBO Alpha-Beta AAR with Moon and Fion?

Now that was a classic!

[JasonC]

What this is, is a sign that their campaigns take too long to be realistically playable, as presently configured. I can sympathize. I've run several CM campaigns, and the two longest ran several months. One finished 2 days of in game action, the other finished 1 day - and was supposed to continue (and may yet). They generated 25-50 tactical fights apiece. I can give you something of an AAR of day one of the Kursk campaign, and I have the scenarios. Maybe some would be interested in those.

[Sergei]

I would find it more likely that the beta test simply hasn't started yet, so there's no one to write an AAR. Then again, what would I know.

Your campaign AAR would be a good read, though, so please do go ahead!

...hmm, re-reading the FAQ, it seems to have been updated:

Q: Do you need beta testers or people willing to create new campaigns for the final release?

A: Not at this time. While we did seek for candidates earlier, we now have enough interested people offering to help. We appreciate the enthusiasm and excitement that this title has generated and will be contacting those people with more details shortly.

So let's hope to hear more soon.

[juan_gigante]

I second the desire to read a bit of an AAR from JasonC.

[Michael Dorosh]

Originally posted by JasonC:

What this is, is a sign that their campaigns take too long to be realistically playable, as presently configured. I can sympathize. I've run several CM campaigns, and the two longest ran several months. One finished 2 days of in game action, the other finished 1 day - and was supposed to continue (and may yet). They generated 25-50 tactical fights apiece. I can give you something of an AAR of day one of the Kursk campaign, and I have the scenarios. Maybe some would be interested in those.

Errr.....the difference being that everything you did by hand will be done by the computer....

[Peterk]

I don't see the campaigns taking that long to playtest because of the auto-resolve feature. Testers will NOT have to go into CMBB and play every battle.

They just have to make sure that, in general, the CMBB battle set-up is good and that the data transfer at the end is good.

CMC campaigns might even end up being easier to test out and balance than an operation in CM is.

[securityguard]

Originally posted by Peterk:

They just have to make sure that, in general, the CMBB battle set-up is good and that the data transfer at the end is good.

And that would require playing a battle with hundreds of different units...?

I don't think beta testing is as easy as you think it is, sorry. I would know from experiences that have left me a shallow husk of a game editor

[Peterk]

That part of the game probably does not have to be tested for each individual campaign. You can test it standalone. And once it's working properly you can start to trust it.

[JasonC]

Peterk - um, if you don't playtest it with CM resolution of the fights, you will get a distinctly unrealistic sense of whether the system is working or broken. Because tape and bailing wire in the auto-resolve routines can easily mask broken subsystems, by not stressing the flow of the campaign the way actual player resolution of the fights, will.

Variance in the outcomes, contested locations, weird retreats, full annihilations, exitings through enemies left intact etc will all come up in player resolved cases and can all be wished away with a magic wand in auto-resolved ones. Also you need to test the playability of the formation and maps sizes, the ragged leftover bits or how they are consolidated, etc.

Until you playtest a campaign with actual player resolution, you've got an outline of a system but not an actual game.

[Sergei]

While auto-resolving battles will yield some information, it doesn't mean that the properties of the campaign map tile correctly portrays the underlying CMBB map and its tactical possibilities.

It IS a good way for preliminary testing, however. Especially since you can pit AI against itself - if it can play the campaign through couple of times with reasonable results, it might be ready for testing by players.

[Peterk]

Yes, but CMC was in all likelihood a playable standalone article (ie. an actual game) before the links to CMBB were added.

Beta-testers will most definitely have to run lots and lots of CMBB battles in the early stages to make sure the fusion is working, but will the end user/designer? After it's released and determined to be working well? Probably not, if he doesn't want to.

The way I see it CMC is 3 big black boxes...

1) The CMC engine in which the campaigns get run/loaded.

2) The combat engine that simulates CMBB results

3) The transfer mechanism between CMC and CMBB

If I'm the coder, I'm going to test each of those things damn well individually and write dedicated programs to automate a lot of it for me if possible.

I'm pretty sure it's possible to write a routine for #2 that simulates convincingly most of the end results of a CMBB game. Squads will have lost men and some may have broken; some AFV and equipment may have been lost and some flags may have changed hands. Yes, there's lots of combinations, but it's not outlandishly difficult. Will simulated results ring 100% true and look "human". Maybe not, but they don't have to...they just have to be close.

These are games which are intended to be played by multiple players over the long term. There is no reaonable way for a single campaign designer to extensively check all (or even most) possibilities with CMBB and still release his creation for us to enjoy during the lifespan of CMBB. I think he will have to trust that black box #2 will do a decent job of simulating the results that can occur in CMBB + black box #3...otherwise we're fried.

[JasonC]

Of course the designer can't check all of it alone, that is why you have playtests. In case everybody forgot, a playtest is what this thread was advertized as being about.

[General Bolt]

Jason, Good point on the need to test the remote CMBB battles in order to fully test the CMC/CMBB interface. One solution would be to have the CMC beta-testers ask us common folk to play a battle or two of CMBB for them. Parallel fighting of the generated battles could speed testing up immenesly.

[JasonC]

We always used both parallel resolution and TCP/IP resolution. I kept the fights in the 1500 point range or less to make that easier to handle, and the battle lengths to 20-30 turns. We resolved one operational move per week that way when all went well - with, naturally, occasional slippages when they didn't.

[juan_gigante]

I don't think that we can say whether or not auto-resolve will be viable for play-testing until we've seen and used it ourselves. Maybe it will do a shoddy job of replicating human play, and be invalid for playtesting, or maybe it will be super-awesome, and be able to play-test great.

Originally posted by JasonC:

Variance in the outcomes, contested locations, weird retreats, full annihilations, exitings through enemies left intact etc will all come up in player resolved cases and can all be wished away with a magic wand in auto-resolved ones.

See, until we see the final version of CMC before us, we won't know how much of this "wishing away" the auto-resolve will do. Maybe there will be a lot. Or maybe there will be a little. Until we know, it is difficult to judge the quality of the CMC auto-resolve as a campaign play-tester. That said, I think that the best way to do things would be to auto-resolve while in the process of creating the campaign, and then once it has reached a final draft stage, human play-testing it once or twice to work out any kinks the auto-resolve might have missed.

[JasonC]

j-g - yes, I do. I am telling you, it is not possible to program an auto-resolve that will do things as wacky as players sometimes will. You wouldn't even want it to. It is mostly going to give reasonable, expected type performances. Players are not. Things that it would make no sense to program in as possible occurances for an auto-resolver, will happen when there are two humans behind the wheel. You can bank on it. I don't mean every time - of course not. But they will do wacky things. The state space they are exploring is vastly larger than the auto-resolver can possibly be. 15 sigma events from the standpoint of the auto-resolver, will happen in the first human campaign.

[juan_gigante]

But I am hypothesizing that no play-testing, human or otherwise, can account for the limitless things that might happen, a problem multiplied many times over in a campaign setting, when no individual battle will be setup the same way twice. Are you saying that human play-testing could find "things as wacky as players sometimes" do? If human playtesting as opposed to auto-resolving could identify 5% more of the wack things humans do, is it really worth the many extra hours of playing? Call me undedicated, but I'm not going to put in an extra 15 hours on playtesting to find another 5% of wacky possibilities.

Besides, even with auto-resolve, a human is in charge of the greater strategic or operational picture. Even using auto-resolve for battles, it is still a human making decisions like to group all his tanks and to send them into the thick forest. That, I think, is an example of a non-standard command decision. I think that for CMC, the focus of playtesting would be on making sure that the greater operational stage functions - after all, if you had 10 years, I doubt you could playtest every single possible battle that could occur in your campaign. And auto-resolve can check things like overall balance and seeing that there aren't any clearly broken areas or forces. In fact, the time saved using auto-resolve means that one could playtest even more at the operational level - trying out more of those wacky human decisions we want to investigate.

In essence, I'm saying that the onus will be on operations, with the individual CM battles in the background. Hence, one should put the majority of playtesting effort into the operational stage. Auto-resolve makes that process faster and easier. I just don't think that for an operational level game, the rare vagaries of human on human play for individual battles will have too much of a difference. And besides, for crucial battles, you can choose to specifically human on human those. Here's my point - when playing through a campaign, people would probably auto-resolve a battle that is platoon scale and play anything bigger. What I'm saying is that for play-testing, you auto-resolve reinforced company scale or less. I really don't think that so much of a difference would be made.