Some thoughts concerning the present FPC engine

Matrix Games Forums

Forums Register Login Photo Gallery Member List Search Calendars FAQ

My Profile Inbox Address Book My Subscription My Forums Log Out

Some thoughts concerning the present FPC engine

View related threads: (in this forum | in all forums)

Logged in as: Guest

Users viewing this topic: none

Printable Version

All Forums >> [New Releases from Matrix Games] >> Flashpoint Campaigns Series >> Some thoughts concerning the present FPC engine

Page: [1]

Message

<< Older Topic Newer Topic >>

Some thoughts concerning the present FPC engine - 12/16/2020 10:08:04 PM

Panta_slith

Posts: 318
Joined: 4/4/2002
From: Montevideo, Uruguay
Status: offline

In the last weeks I have been playing again FPC, mostly with other players via pbem++ and I guess that some aspects of the game, flaws, actually, IMO, sholud be improved in terms to make the game still more appealing to players who, like myself, would choose a realistic approach over a simply funny one.

1-Some of the announced additions are mostly cosmetic, like the multiple layers and windows that can be displayed on the screen. Being a photographer and having worked in photo editing in an international wire services news agency for 17 years, I can say that I have some experience in screens stuffed with windows. Of course, that can be mostly solved using multiple monitors, though probably not everybody can indulge him/herself with one or more extra monitors. Anyway, those improvements will be undoubtely welcome.

2-The new AI may be something that will make the difference. With the present game engine, the lack of a suitable AI (can script codes be considered AI somehow?) the asymetrical turns hinder the player with longer times to wait between orders (WP), making their play unrealistic. After all, even with ther rigid chain of command WP units had, at least they had an SOP to attach too in case of doubt.

3-A particularly unrealistic aspect of the game is what I should call "recon by arty". The player with enough artillery (and ammo) just begin blasting the most probable locations of the unseen enemy and when an explosion indicates a hit, there must be an unseen foe, even if totally out of sight and far away. A common R/L tactic before an assault is to make a barrage (even a rolling barrage), but the attacker doesn't always get a feedback of the attack.

4-The engineers thing is another most wanted improvement, unnecessary to say.
So said, I must admit that I am eagerly awaiting the launching of the new engine. Meanwhile, let's play a little gamey...

< Message edited by Panta_slith -- 12/16/2020 11:00:42 PM >

Post #: 1

RE: Some thoughts concerning the present FPC engine - 12/17/2020 1:21:13 PM

pzgndr

Posts: 3170
Joined: 3/18/2004
From: Maryland
Status: offline

quote:

ORIGINAL: Panta_slith
2-The new AI may be something that will make the difference. With the present game engine, the lack of a suitable AI (can script codes be considered AI somehow?) the asymetrical turns hinder the player with longer times to wait between orders (WP), making their play unrealistic. After all, even with ther rigid chain of command WP units had, at least they had an SOP to attach too in case of doubt.

Whatever happens with C2, whether orders delay or ops points limits or whatever, needs to be applied to the AI too. I was stunned when we found out the AI isn't subject to delays.

Personally I don't care for the orders delay mechanic. I'd rather see set turn lengths and HQ command points like the Assault! series, where NATO had Bn-level HQs and WP had Regt-level HQs. That presented some challenging limits on WP and forced the player to use battle drill formations. More realistic IMHO. Ah well. I'm standing by to see what Southern Storm has to offer and how it plays out. If a revised orders delay works, then fine.

(in reply to Panta_slith)

Post #: 2

RE: Some thoughts concerning the present FPC engine - 12/17/2020 1:26:37 PM

CapnDarwin

Posts: 8467
Joined: 2/12/2005
From: Newark, OH
Status: offline

To clarify, the AI was always subject to command delays. That was the whole point of the asynchronous WEGO mechanics. What the AI could not do was deal with the limited orders option.

_____________________________

OTS is looking forward to Southern Storm getting released!

Cap'n Darwin aka Jim Snyder
On Target Simulations LLC

(in reply to pzgndr)

Post #: 3

RE: Some thoughts concerning the present FPC engine - 12/17/2020 3:51:01 PM

pzgndr

Posts: 3170
Joined: 3/18/2004
From: Maryland
Status: offline

quote:

ORIGINAL: CapnDarwin
What the AI could not do was deal with the limited orders option.

That's what I meant, sorry. So again, SS needs to apply limited orders to AI.

(in reply to CapnDarwin)

Post #: 4

RE: Some thoughts concerning the present FPC engine - 12/17/2020 4:55:07 PM

fluidwill matrix

Posts: 140
Joined: 12/7/2006
Status: offline

Allow me a short rant - this is not in any way personal - just my opinion for all its meagre worth and not limited to Flashpoint.
What is it with these purista's and the idea that an AI should obey same rules as a player?
The job of a decent AI, and God knows its hard enough to find one, is to provide the human with a challenging fun game opponent - until the recent advent of Alpha Zero given the level playing field, programmed AI opponents were, are and continue to be absolutely substandard compared with a human opponent who has any understanding of the mechanics.
It makes me laugh, an AI will not spend the whole game sneaking a recon unit round the edge of the map to snatch 3 VP hexes on the last turn. A human will do this. Repeatedly. And boast about it afterwards.
If you want to win all the time and (again my opinion only) the AI is beating you, just put it on easy. The AI cannot cheat, it doesn't care if you win or not. It only exists to provide you with fun opposition, if your only enjoyment comes from winning just lower the difficulty.
Rant over - fix the PBEM++ end game thing, thanks x

(in reply to pzgndr)

Post #: 5

RE: Some thoughts concerning the present FPC engine - 12/17/2020 6:43:14 PM

pzgndr

Posts: 3170
Joined: 3/18/2004
From: Maryland
Status: offline

Allow me a short counter-rant.

This is not about beating the AI for the sake of beating the AI. I expect a computer opponent to play by the same game rules and be competent at most of them. Not brilliant, of course, but at least competent. And then difficulty settings can provide known bonuses or handicaps as a player sees fit.

The issue with the limited orders option is to try to get inside the enemy's decision making cycle to gain an advantage. But that could not be done in this game with this AI. Big surprise. And a big disappointment for this former mech infantry company commander in Germany 1987-90 where we trained to use our C2 advantages against our WP opponents.

As for providing a challenging fun game opponent, I have a different perspective. One example is my Advanced Third Reich mod for Strategic Command which has challenging AI scripted for both Axis and Allies for the entire ETO conflict. Another example is my code development work for Empires in Arms which has the AI much more challenging now than it was. It's not about winning, unless it is hard fought against a competent and challenging computer opponent. And what makes you laugh about a human opponent playing "gotcha" with a gamey tactic that no real-world opponent would be doing is the stuff that makes me cringe and avoid wasting my time with some PBEM players. You can have that.

(in reply to fluidwill matrix)

Post #: 6

RE: Some thoughts concerning the present FPC engine - 12/17/2020 7:51:18 PM

Panta_slith

Posts: 318
Joined: 4/4/2002
From: Montevideo, Uruguay
Status: offline

I concur with pzgndr. Realism in wargames can vary a lot, from beer and pretzel games to have a little fun, to complicated and hard to play military wargames trying to comply both with reality and military doctrine. My wife plays majong on her laptop when she is tired of working, that relaxes her. On the other hand, yo can always play chess, there is no cheating or gamey playing in chess, it is a closed game.
If you play wargames, tactical or operational wargames in particular, you will probably want to replicate the r/l challenges and problems as much as possible. Otherwise, there are plenty multi-player tank, naval or air combat games on line for free, though they are mostly of the arcade type with better graphics and interfaces.
I have also noticed that most PC WG players play against the AI. It is far more rewarding, IMO, to play against another fellow human, but then , as pzgnr accurately pointed, a lot of glory-seeking guys use whatever gamey approaches they found to win.

(in reply to pzgndr)

Post #: 7

RE: Some thoughts concerning the present FPC engine - 12/18/2020 10:59:39 AM

jmlima

Posts: 782
Joined: 3/1/2007
Status: offline

quote:

ORIGINAL: fluidwill matrix
..
What is it with these purista's and the idea that an AI should obey same rules as a player?
The job of a decent AI, and God knows its hard enough to find one, is to provide the human with a challenging fun game opponent - until the recent advent of Alpha Zero given the level playing field, programmed AI opponents were, are and continue to be absolutely substandard compared with a human opponent who has any understanding of the mechanics....

I think it starts with a common confusion of naming a computer opponent as AI in the first place. Most (could be argued all) games have scripted opponents, not AIs, ie, the computer will not learn from mistakes and develop his own solutions to the challenges posed by an human player, but will remain operating within the constraints of the scripting provided. You can easily see this in some games via the 'gigantic flanking move' effect. Some games are quite good at providing an electronic opponent, but rarely will this opponent provide a total surprise to the human player via (for example) a totally unexpected move that takes advantage of that temporary weakening that the human player created on his forces. Going back to my initial point, (IMHO) calling a computer opponent 'AI' often creates unrealistic expectations on players about the challenge this computer opponent will provide but, most of all, provides very unrealistic expectations about the degree to which this opponent will adapt and react to players actions and even to game conditions.

(side note (also IMHO) not realistic to compare wargames to chess in terms of what a programmed opponent can achieve, better to look at something like go)

(in reply to fluidwill matrix)

Post #: 8

RE: Some thoughts concerning the present FPC engine - 12/18/2020 11:56:51 AM

pzgndr

Posts: 3170
Joined: 3/18/2004
From: Maryland
Status: offline

quote:

ORIGINAL: jmlima
rarely will this opponent provide a total surprise to the human player via (for example) a totally unexpected move that takes advantage of that temporary weakening that the human player created on his forces

I invite you and others to try my Advanced Third Reich mod for Strategic Command WWII War in Europe. Besides the generic 'fuzzy logic' for most routine game functions, the variability in the events and AI scripting makes the computer opponent not totally unpredictable but close enough to seem like a human opponent. I scripted things to mimic how I play, with multiple and diverse strategies, so that yeah even I am occasionally surprised by an unexpected situation as a result of my own scripting.

However, to acknowledge your point, it is 'rare' to find a game where the developer or modder has invested sufficient time and effort into scripting a nefarious computer opponent. My A3R mod evolved over about the last 15 years through various iterations of Strategic Command with more and more scripting capability along the way. So, it 'can' be done, but sadly we wargamers often don't see such AI improvements and enhancements over time. Developers do what they can and move on to their next game; they have to, you don't make money off years of AI improvements. But hey, if CapnDarwin and company can keep at it and help make this AI as nefarious as possible, then that would be great!

(in reply to jmlima)

Post #: 9

RE: Some thoughts concerning the present FPC engine - 12/18/2020 1:41:00 PM

Panta_slith

Posts: 318
Joined: 4/4/2002
From: Montevideo, Uruguay
Status: offline

A few years ago I remember having the same exchange with James Sterrett, who I know from the times of the TacOps list (presently he is Chief of the Simulation Education Division in the Directorate of Simulation Education of U.S. Army University/Command & General Staff College) and who has always been for me a reference in wargaming. When I ventured the same criticism about the wargames AI, he just told me to check the behaviour of a then recent wargame they were using at work, Panther Games' Command Ops. In his opinion, CO's "AI" was something different and far more efficient. I don't know if CO employs scripting or uses a more evolved AI, I should ask him about now, but it seems that even then, there were different types of AIs.
In a recent Matrix teaser Rob Crandall refers to their upcoming version of FPC and its totally renewed AI, and I hope that they make a good job, otherwise a complex game like that could be a failure in some aspects.
BTW, another much needed improvement will be the possibility to assign SOEs to the units, something that an old game as TacOps had since its release back in 1994, and an idea that for some reason nobody else seemed interested in imitate.

(in reply to pzgndr)

Post #: 10

RE: Some thoughts concerning the present FPC engine - 12/18/2020 4:29:09 PM

pzgndr

Posts: 3170
Joined: 3/18/2004
From: Maryland
Status: offline

quote:

ORIGINAL: Panta_slith
another much needed improvement will be the possibility to assign SOEs to the units, something that an old game as TacOps had since its release back in 1994, and an idea that for some reason nobody else seemed interested in imitate.

Exactly! TacOps was great in that the OPFOR would enter along either the north, center, or south avenues and you, the player, did not know which. So you needed to play the recon/counter-recon battle and shift your forces accordingly in preparation for the main battle. Having variable entry would be a very nice addition to this game. Should be doable.

Offhand, the whole semantics gripe about AI vs CO is largely irrelevant. Certainly there is a difference between a learning AI and a scripted AI, and we're not likely to see a true learning AI in any niche wargame. At best, a game could maybe save counts of certain activities and use some weighted averages to fine-tune the scripted AI to adapt it to a player's style, but I'm not holding my breath. Call it whatever you want, but at the end of the day it's a computer opponent for a player to play a computer wargame at his convenience and with whatever difficulty settings he chooses to play the kind of game he wants to play. For fun. Some folks forget that.

(in reply to Panta_slith)

Post #: 11

RE: Some thoughts concerning the present FPC engine - 12/18/2020 5:01:55 PM

Panta_slith

Posts: 318
Joined: 4/4/2002
From: Montevideo, Uruguay
Status: offline

Actually I am not worried about the AI handling the opponent, but about the way the players' own units react in different situations during the period he cannot intervene, in particular in the case of the player with the longest turns (WP).
Doctrine (for everybody) dictates, for instance, that a dug in vehicle, after firing or being fired at, can expect after a lapse that will change according to several variables (type of foe, enemy arty threat, etc.) to receive a rain of arty shells. Whoever played Steel Beasts Pro, even if he or she is a civilian, has learnt that the hard way. Well, in FPC units remain in place awaiting for long minutes for the unavoidable fire and steel rain. No tanker in his wits would do that in R/L. That is an unrealistic aspect of FPC that a well designed AI could help to prevent, for instance.

(in reply to pzgndr)

Post #: 12

RE: Some thoughts concerning the present FPC engine - 12/18/2020 6:05:25 PM

CapnDarwin

Posts: 8467
Joined: 2/12/2005
From: Newark, OH
Status: offline

Guys, I agree a lot with this AI discussion. Game "AIs" are nothing more than complex if-then engines. You can add in weighted values and some fuzzy RNG to make them less chess like and predictable. Red Storm AI does a lot under the hood to replicate the OODA loop process for the units and formations on the battlefield. It has a mission, it reacts to the seen enemy (no all seeing AI) and it generates a number of Courses of Action (COAs), and decides on what to do. We have been working hard to improve on this model, add more control to the various units (SOPs), and teach the AI how to use these new tools to better fight and maneuver. As noted above with the unknown nature of an enemy attack (or defense), the new game engine will have scenarios with even more replayability than Red Storm based on a number of new features we have added. We have also done a lot of work to fix and improve the C2 aspects of the game as well. So recon units will back off when the enemy gets too close and units will run minefields instead of sitting still, and many other improvements to the fidelity of the system. We may not hit everything that everyone wanted, but it is a much better, more refined and realistic game than Red Storm was. In my humble opinion.

_____________________________

OTS is looking forward to Southern Storm getting released!

Cap'n Darwin aka Jim Snyder
On Target Simulations LLC

(in reply to Panta_slith)

Post #: 13

RE: Some thoughts concerning the present FPC engine - 12/18/2020 6:46:12 PM

Panta_slith

Posts: 318
Joined: 4/4/2002
From: Montevideo, Uruguay
Status: offline

Good news, Cpt! To illustrate the SOP point I was referring to, I had an image of old TacOps SOP that was good enough then, I presume that today it can be done far more comprehensive, but I did't find the way to post it here. I suppose that it isn't possible.

(in reply to CapnDarwin)

Post #: 14

RE: Some thoughts concerning the present FPC engine - 12/18/2020 9:13:30 PM

pzgndr

Posts: 3170
Joined: 3/18/2004
From: Maryland
Status: offline

quote:

ORIGINAL: Panta_slith
Good news, Cpt! To illustrate the SOP point I was referring to, I had an image of old TacOps SOP that was good enough then, I presume that today it can be done far more comprehensive, but I did't find the way to post it here. I suppose that it isn't possible.

Here it is. Yes, good enough then and sorely missed now. I am optimistic that Jim has heard our pleas and will show us an improved SOP menu in SS.

Speaking of AI and TacOps, there were many games I got surprised by that reinforcing tank bn that kinda showed up right where I wasn't prepared...

Attachment (1)

(in reply to Panta_slith)

Post #: 15

RE: Some thoughts concerning the present FPC engine - 12/18/2020 9:51:31 PM

Panta_slith

Posts: 318
Joined: 4/4/2002
From: Montevideo, Uruguay
Status: offline

quote:

ORIGINAL: pzgndr

Here it is. Yes, good enough then and sorely missed now. I am optimistic that Jim has heard our pleas and will show us an improved SOP menu in SS.

Speaking of AI and TacOps, there were many games I got surprised by that reinforcing tank bn that kinda showed up right where I wasn't prepared...

Oh yes, it was exactly that one! And btw I realize that you are another ancien combattant from the good old times, probably also from the Major's private list, where simple amateurs like myself were kindly allowed to exchange thoughts with seasoned servicemen...
I also hope that the new engine will fulfill our mos cherished secret wargaming desires!
And you are probably mentioning the unspeakable Phantom Battalion that was always there to remind you about the need to guard your flanks at all times.

< Message edited by Panta_slith -- 12/19/2020 11:31:56 PM >

(in reply to pzgndr)

Post #: 16