Matrix Games Forums

ORIGINAL: rhinobones

ORIGINAL: ralphtrick
By Sawtooth, I mean something like the following . . .

Well, that sure clears things up!! Can you draw a picture of 4-2-6 ?

Regards, RhinoBones

It's not just 4-2-6. Draw "dot-to-dot" from the 1 to the 2 to the 3 to the 4 to the 5 to the 6. Then, you'll see what Ralph is referring to as a "sawtooth".

For my attacking PO, I try to use smaller formations, with overlapping "Christmas Tree" objective tracks.

The primary axis of advance is the trunk, and the branches are trip-wire objectives in the rear, and flanks, that will pull back the formation if it is in danger of being cut off, or to assist nearby formations if their advance is being held up.

ORIGINAL: Szilard
Say I've got a long river line which I want a formation to defend along - what's the best way to set up objectives? Scatter them at equal intervals with a sawtooth back to a fall-back line? Do I need to have every hex in the line covered by an objective - I assume not, but is there any way of estimating the best density?

Check Jamiam's response and see if that makes any sense.

For several maneuvers, Elmer actually looks in a radius around the object, not just at the objective itself. For example, when figuring if he's flanked, and to modify his stance, he looks at the relative strengths in a circle of a radius of between 5 and 9 hexes around the center of your formation, depending on his intelligence. His old behavior was definitely triggered by objective hexes, his new behavior is much less programmed in general.

I'm still learning the code, so I haven't gotten all the subtleties down, so you'll have to experiment some. I'm hoping he's strong enough to be able to play against himself for rough play balancing.

There are two ways that you can try. I'm not sure which one will work better, or even if there is one 'best' way.

Method 1 would be the shotgun method proposed by Jamiam, small formations, all-inclusive objective tracks...
Method 2 would be the method that CommC switched to, which is larger formations.

Elmer's biggest problem right now is the fact that formations don't coordinate well yet. You can overcome that by either using larger formations, or more inclusive programming of the objective tracks.

I'm lazy[;)], I'd try the larger formations first, and if that fails, add in the extra work on the objective tracks.

Ralph

Well even this much has been super helpful. I just got one army to actually put up a nice semblance of a fall-back defense by manipulating the objectives. It is very tough to coordinate different formations on the defense though. Still, its performing far above where it was before you posted some of these hints.
 
Thanks again!

ORIGINAL: ralphtrick


For several maneuvers, Elmer actually looks in a radius around the object, not just at the objective itself. For example, when figuring if he's flanked, and to modify his stance, he looks at the relative strengths in a circle of a radius of between 5 and 9 hexes around the center of your formation, depending on his intelligence. His old behavior was definitely triggered by objective hexes, his new behavior is much less programmed in general.

I'm still learning the code, so I haven't gotten all the subtleties down, so you'll have to experiment some. I'm hoping he's strong enough to be able to play against himself for rough play balancing.

There are two ways that you can try. I'm not sure which one will work better, or even if there is one 'best' way.

Method 1 would be the shotgun method proposed by Jamiam, small formations, all-inclusive objective tracks...
Method 2 would be the method that CommC switched to, which is larger formations.

Elmer's biggest problem right now is the fact that formations don't coordinate well yet. You can overcome that by either using larger formations, or more inclusive programming of the objective tracks.

I'm lazy[;)], I'd try the larger formations first, and if that fails, add in the extra work on the objective tracks.

Ralph

Jamiam's message makes perfect sense for the attack, and is kind of what I do already; defense is a lot less clear.

I went back to the polog (not very enthusiastically; it's such a chore) & reconfirmed my earlier views on it - it can give you some hints, but only hints, and it there are big chunks of thinking which it doesn't reveal to you. Be great if you could look at beefing up the amount & quality of info you get from it at some point, and also adding in some kind of debugging monitor which would let you capture info from just specified formations, set watchpoints etc etc.

It did recall one of other thing I'd really like to see looked at sometime: the nexus between formation orders, loss tolerance and strategic bias. The two should be independent, IMO, but at the moment of course they're not.

I want to be able to set a force bias of "Very Cautious", meaning "care *a lot* about flanks; only attack at excellent odds; etc etc etc" while still being able to set a formation to defend at "Ignore Losses" - "If you do make a stand, defend very strongly". Hard to make the PO do that reliably, given that bias, orders and loss tolerance can all (?)modify each other for particular formations. The worst is when a formation's orders and loss tolerance and the strategic bias all come together to result in the AI telling a formation, "Do Nothing", which I think is rarely what is wanted.

ORIGINAL: Uncle_Joe

Well even this much has been super helpful. I just got one army to actually put up a nice semblance of a fall-back defense by manipulating the objectives. It is very tough to coordinate different formations on the defense though. Still, its performing far above where it was before you posted some of these hints.

Thanks again!

The next upgrade should stiffen the defense quite a bit. Right now the PO tends to charge out of the trenches a bit too much.

ORIGINAL: Szilard
I went back to the polog (not very enthusiastically; it's such a chore) & reconfirmed my earlier views on it - it can give you some hints, but only hints, and it there are big chunks of thinking which it doesn't reveal to you. Be great if you could look at beefing up the amount & quality of info you get from it at some point, and also adding in some kind of debugging monitor which would let you capture info from just specified formations, set watchpoints etc etc.

It did recall one of other thing I'd really like to see looked at sometime: the nexus between formation orders, loss tolerance and strategic bias. The two should be independent, IMO, but at the moment of course they're not.

I want to be able to set a force bias of "Very Cautious", meaning "care *a lot* about flanks; only attack at excellent odds; etc etc etc" while still being able to set a formation to defend at "Ignore Losses" - "If you do make a stand, defend very strongly". Hard to make the PO do that reliably, given that bias, orders and loss tolerance can all (?)modify each other for particular formations. The worst is when a formation's orders and loss tolerance and the strategic bias all come together to result in the AI telling a formation, "Do Nothing", which I think is rarely what is wanted.

Interesting idea. You're right in that those two concepts aren't the same thing. There probably show be separate settings for loss tolerance and for the force bias.

I want to add more dynamic debugging capabilities for debugging the PO later on. Something along the lines of being able to pause, and hover over a formation and it would display information about it's current state. In the meantime, if you have ideas on things that are missing from the PO log, let me know.

Ralph

Any ideas if/when the next upgrade will be released as a beta. It sounds like there are some fantastic and signficant changes coming with it. I know I've suspended working on my scenario until the AA bug is fixed so even an early beta with just that tweaked would be a great help.
 
Thanks again!

It will be released 'When it's done.' I can guarantee it will be a 'TOAD Tuesday!'[:'(]