The "debate" about CMBB's Infantry Modeling

[Hat Trick]

From Battlefront:

This is the problem with computer code. It is inherently an asswipe about stuff like this Even using Fuzzy Logic we run into situations where it is very, VERY difficult to have the AI make a realistic approximation for what real people would do in a given situation. In one situation Sneaking even 5m would be the WRONG decision, while in another running 25m would be the correct one. Trying to establish rules for this, without unit memories, is very difficult

This seems a little defeatist to me (of course, I don't have to write the code, but still). The AI already must use some form of "fuzzy logic" to choose between (in my example) the nearby scattered trees and the farther away woods. A new option, stay put, need not have a fixed rule ("only stay put if cover is more than x meters away"); rather, the decision to stay or run to a particular piece of cover would use most of the same factors that the AI currently uses to choose between nearby scattered trees and farther away woods.

Would such a system be perfect? No. Would people still whine about how it is implemented? Yes, but hopefully less. Reasonable people can disagree about when units should (realistically) sneak for cover and when they should stay put, but I think that most people would agree that in some circumstances staying put is clearly the best option; an option that the AI does not consider now. Given how well everything else in the game is designed, I am sure that I (and many others) would be happy with whatever "weight" BFT put on "staying put" vs moving to cover, as selected by the AI.

Finally, perhaps I'm just dense, or I'm not communicating clearly, but I don't see how this solution is a memory issue. Units may stay put, then the next turn try to sneak to cover, then next turn stay put again, or whatever (not unlike real life, I imagine). This behavior would still be an improvement over the current model.

[Doodlebug]

Steve.

I only probably gained a different view as a direct result of you coming out with that well explained explanation of how some of the cogs fit together. Frankly a lot of the posts up to that point were never going to be seen as anything but inflamatory/defamatory by the FFG or BBG.

I appreciate you taking the trouble to start this post and then taking time out to comment on the various topics regularly.

I respect you for that.

Imagine, people, how different this discussion would be if CDV were involved.

( Sound of silence. Tumbleweed blows by)

[Fetchez la Vache]

Re: "fuzzy logic" and how to model infantry reactions under fire.

My suggestion is to allow the player to give a secondary 'evasive' move order to a unit which is only performed when they come under fire and the TacAI decides to take 'evasive manoeuvres'. The chance of the unit actually performing the evasive movement is dependent on current morale, fitness, HQ bonuses, etc. So that SMG squad moving across the open ground could be given a evasive movement order to head back to the clump of trees it came from. Or hit the dirt. Or assault the treeline up ahead.

This way you could micro-manage the actions of the units when they come under fire - to some extent. If they are tired and green then they might hit the dirt no matter what you tell them to do! (/me ignores the train of thought that says green troops might be more inclined to charge... )

Only a thought. Never hurts.

Btw, when I mentioned Sudden Strike before it was because my clan mate did used to play it. I wasn't trying to suggest that all SS players are dim . Not at all. It's one of the best RTS multiplayer games out there imo. But players first 'exposed' to SS are somewhat surprised to learn the Stürmtiger had a reload time of 5mins! Or was it 10? I forget. Whatever.

[Gary T]

Steve,

Okay, bit of an exaggeration on my part to say I won't play again...I'm sure I'll be tempted back at some point. Very shortly.

To say those are 'whining' about are finding the sneak command a problem due to their faulty tactics is simplistic IMHO. I'm not tactically inept...I know my tactics from experience in the real world. I do not go charging about across open ground a la CMBO. Maybe those who don't have a problem are simply not too concerned by the odd suicide sneak...maybe I've been concentrating on it too much myself. The fact remains however that there is a problem, small or large depending on who you ask.

As for staying put and not budging an inch...thats exactly what happens in real combat. Troops stay pinned down for hours on end. And I know, it wouldn't be fun for the player but I'd like that feature. The only way to extricate them would be by bringing large amounts of firepower on their assailants and/or outflanking them with another force or waiting for nightfall. Call me a sadist but as I said I think that would be realistic.

To take my previous example...if they had stayed put when coming under fire I could have given them a movement order backwards next phase - just a few yards. They may not have made it and got pinned down again. Fair enough.

If they sneak off I can cancel that and give them another order...what usually happens is they get fired on again and immediately revert to sneaking, usually in the opposite direction to previously. Therefore once your men start sneaking it is best to let then keep going otherwise they end up going round in circles.

Anyhow, look forward to the patch.

Gary.

[James Crowley]

The infantry model is, IMO, light years ahead of that in CMBB. Not withstanding some idiosyncrasies (such as the dance of death and the exhausting sneak, which could well be moderated with the upcoming patch) the overall effect of one single change has done wonders to the realistic portrayal of infantry under fire. They hit the dirt!

However, while much is being made of the correct use of orders, tactics etc. one thing that puzzles me is that the time-frame of CMBB is the same as CMBO.

If we assume that the more realistic infantry model “slows” the pace (to allow more suppression, move to contact and so on) then why is the default time for QB’s still only 30 minutes and why do the designed scenarios still have average times similar to CMBO?

With an infantry attack, or worse still an assault, on a small map with difficult terrain 30 mins is hardly enough; a medium map starts to be a struggle and the large and huge maps close to impossible. (I know you can manually increase the time, just puzzled as to why the default hasn’t been altered)

Surely we should be looking at longer times to allow those historically correct tactics to be replicated?

On a completely different note, did any one else pick up on Steves’ little clue about muli-player in the engine re-write. If I read it correctly multiplayer will mean just that, not just two- player. So maybe I can command A company while some one else runs B and so on. If that is the case might it be possible to have the AI control some companies or platoons while I control the others. Now that would be something!

[rune]

Boo,

I am just helping out. Send it to Matt.

Rune

[Battlefront.com]

HatTrick,

This seems a little defeatist to me (of course, I don't have to write the code, but still).

And Hitler thought Manstein was a defeatest when he requested a pullback after Stalingrad Seriously, I am not being defeatist in the least. There is more that can be done with this stuff for sure. However, through experience with programming in general and CM's code specifically (indirectly, of course) I know how utterly painful it is to eek out even tiny behavioral improvements. What we are trying to make the computer do is mimic the behavior of masses of human beings, consistantly correct, in an infinite number of specific circumstances. We can't do it. Nobody can. Best we can do is sorta get it right and not go bankrupt in the process.

The AI already must use some form of "fuzzy logic" to choose between (in my example) the nearby scattered trees and the farther away woods.

Yes, the AI is for the most part Fuzzy. Most of the game mechanics, in fact, are Fuzzy. It is the only way to go for a system like this.

A new option, stay put, need not have a fixed rule ("only stay put if cover is more than x meters away"); rather, the decision to stay or run to a particular piece of cover would use most of the same factors that the AI currently uses to choose between nearby scattered trees and farther away woods.

True, but through direct experience (i.e. Charles making code changes available to us to play with) it simply didn't work. Units were too easily pinned. I don't know exactly why the game dynamics led to this end result, but it did. And it sucked very badly.

Reasonable people can disagree about when units should (realistically) sneak for cover and when they should stay put, but I think that most people would agree that in some circumstances staying put is clearly the best option; an option that the AI does not consider now.

The problem is that Charles was unable to come up with an adequate rule set to make this work well enough to be included. Again, I can only comment on this to a limited degree since it was a long time ago and I didn't code it. But the implementation now is far better than what it was like before when we tried to do just this.

Finally, perhaps I'm just dense, or I'm not communicating clearly, but I don't see how this solution is a memory issue. Units may stay put, then the next turn try to sneak to cover, then next turn stay put again, or whatever (not unlike real life, I imagine). This behavior would still be an improvement over the current model.

Memory allows the TacAI to know things like "I already tried to go that direction and it didn't work, so I won't go that way this time". That will stop the dance back and forth between two pieces of cover, for example. "I am taking fire, so I'm going to sit still and not make a break for those woods 25m away. (5 turns pass) Crap, I am still here, taken 3 casualties, and still getting shot at. Gee... maybe that 25m dash doesn't seem so bad afterall". Stuff like that.

Doodlebug,

I only probably gained a different view as a direct result of you coming out with that well explained explanation of how some of the cogs fit together. Frankly a lot of the posts up to that point were never going to be seen as anything but inflamatory/defamatory by the FFG or BBG.

Glad I could be of service

Fetchez la Vache,

My suggestion is to allow the player to give a secondary 'evasive' move order to a unit which is only performed when they come under fire and the TacAI decides to take 'evasive manoeuvres'.

This what we call SOPs, and they will be in the engine rewrite. But no way no how can we get them into the current code. Far too low level a change to make at this point. We even rulled it out 2 years ago as too difficult. Fresh code will make this a piece of cake to add, so it will go in no question about it.

Gary,

To say those are 'whining' about are finding the sneak command a problem due to their faulty tactics is simplistic IMHO.

Oh, I agree. However, you are forgetting classic comments that are being made like these from Kraut on Page 4:

"I just got done playing a solid tcp/ip game with CMBO and it was great! Infantry actually returned fire instead of running away! You could count on your infantry to perform. Infantry were actually able to move up to a treeline and create a base of fire to let another platoon advance. You think you can do this in CMBB? Hah! A few bursts from an enemy MG will send a whole damn platoon running for home.

Noone can tell me that this infantry is worth a damn. No matter how fundanmentally functions flabberheiner they are."

This is the extreme position I discussed on Page 1. It is a position I have seen over and over again. The Sneak thing is only a tiny outcome of such players' problems, not the cause of them. To them they do want to streak over open terrain and overrun MG positions. They don't want to have their troops flinch any more than they would in CMBO. This is quite different than people just taking issue with the sneak thingy.

Maybe those who don't have a problem are simply not too concerned by the odd suicide sneak...maybe I've been concentrating on it too much myself. The fact remains however that there is a problem, small or large depending on who you ask.

True, but is it SO bad that the game is unplayable as some have charged (including you, initially)? This is my point. We are all playing the same game, so how can it be so bad and at the same time not much of an issue when the latter group claims they are winning battles just fine?

As for staying put and not budging an inch...thats exactly what happens in real combat. Troops stay pinned down for hours on end.

This is true, but not necessarily staying pinned in one physical location. Yes, it CAN possibly mean that, but in a firefight situation it is rare. Staying put generally equals eventual death. This is not the same in other non-CM type combat situations (waiting out a bombardment for example).

To take my previous example...if they had stayed put when coming under fire I could have given them a movement order backwards next phase - just a few yards. They may not have made it and got pinned down again. Fair enough.

Units do have the ability to stop dead in their tracks and not move. If they are already in good terrain, they will do that. Open terrain is a different matter because GENERALLY staying put is not something a unit would do.

Steve

[Battlefront.com]

Oh!! BTW, something I totally forgot about until now. I had Charles tweak some AI weights a bit so that both the StratAI and TacAI will not be as snobbish about open terrain as it was before. This might help out a bit regarding troops favoring sneaking to distanat cover rather than staying put. But we probably won't know that until the larger audience (i.e. you guys) gets to see it first hand.

Steve

[Grisha]

Originally posted by dalem:

Approximately 200-230m. (EDITED to add And EFOW.) My MGs are in cover (scattered trees, not under fire, in command, and not hiding. My advancing platoons were rapidly dead, of course, but I am not discussing that. I am focusing on the behavior of my "support" platoons.

Now, there is light snow, BUT I can spot and target no problem, therefore I expect the AI to. This is what I mean when I talk about feedback in the game. Behavior X yields Outcome Y? I have NO problem with that. I am not asking for unworried strolls through hails of bullets. I am not asking for laser accuracy and instant action. I am just looking for consistency.

I love Area Fire and have used it since my first CM:BO QB. But if I have 4 MGs and I decide I want to hose a key building with two of them and keep the other two free for targets of opportunity, the two that are not firing should fire when ToOs appear. That's their purpose.

Unless I have a critical misunderstanding going on which is clouding my view. That is certainly possible.

-dale

Dale,

Part of the answer may lie in the programming, since we've been told it's fuzzy logic. I just did a short test. It's late summer and raining, steppes, defending Germans in houses, Soviets between 200-300m out in the open. On the first turn, the Germans saw all but one of the squads ftom the beginning, but no one fired. Two Soviet squads advanced mostly in a flanking manner, so they didn't close with the Germans much, if at all. Not one German fired on turn one. However, the very next turn they did open up on the Soviets.

Are you familiar with fuzzy logic? It's a method of creating states traditionally based on distinct conditions and muddying it up to where these states are based on percentages of a number of possible conditions. This is much more realistic, because it's how we as humans generally do things, balancing a lot of variables in our heads to come to some decision at that given moment - a decision that might be different the next time around under similar conditions.

With this in mind, it might be that when targetting an actual unit in CMBB you are ordering your team/squad/vehicle to get a good shot at the target itself. Hence, if the target is defending in a building, then the times where you have the target in solid sight are fragmented and brief. Given the sporadic nature of the sightings, your unit may not 'feel' it has a good enough shot. I use the word 'feel,' because this is how one could interpret fuzzy logic within the game. Now, tell your men to just open up on that leftmost window on the second floor (area fire) and they'll probably just plaster it without too much consideration.

[ November 21, 2002, 12:39 AM: Message edited by: Grisha ]

[rune]

Dalem,

Looking at the scenario, if you did not move the maxims, then the buildings are on the cusp of visibility. Light snow and fog, about 225 meters, the absolute closest building is 204 meters. If you were targeting the second row of houses, out of visibility range completely.

If the AI MGs fired at the second row of houses, then the AI is simulating firing at the gun flashes [Yes, this IS coded in] Also, a lot of the houses have a partial viewpoint, where terrain blocks part of the house, where part of the house can be seen.

You can look at this by looking at the scenario in the editor. However, if you have another case, let me know...as I am still looking into this.

Rune

[Battlefront.com]

Hi James,

If we assume that the more realistic infantry model “slows” the pace (to allow more suppression, move to contact and so on) then why is the default time for QB’s still only 30 minutes and why do the designed scenarios still have average times similar to CMBO?

With an infantry attack, or worse still an assault, on a small map with difficult terrain 30 mins is hardly enough; a medium map starts to be a struggle and the large and huge maps close to impossible. (I know you can manually increase the time, just puzzled as to why the default hasn’t been altered)

Good question. I don't have a good answer except that testers didn't bring this up as an issue and, as of yet, I havent seen it as a big one since the game was released.

I think that part of the reason is that in CMBO you generally spent the last 10 turns or so hunting down the other guy's stragglers. At least that is what I did This in effect means that the battle really only needed 20 turns.

Another possible reason is that certain changes have made games more difficult to play out to the "last man" stage. Well, at least more so than with CMBO.

Surely we should be looking at longer times to allow those historically correct tactics to be replicated?

It is a tough call to make. Depends on the nature of the scenario. For example, look at the turn counts for the two Demo battles. The one with less forces is a lot longer than the one with more. The reason is the depth of the map and the nature of the terrain.

We expect that the new engine will allow us to have QB turn counts be "intelligently" assigned based on the battle's parameters. In CMBO/CMBB it is pretty much just a default variable.

On a completely different note, did any one else pick up on Steves’ little clue about muli-player in the engine re-write.

Nice of you to notice This is actually not new news, but it is something I'm sure many people do not know. The next CM will support multi-multi player of some number (not determined yet). Cooperative play is something we think can bring CM into an entirely new world of possibilities.

Steve

[Grisha]

Dale,

It's also entirely possible that BFC have coded a fuzzy logic algorithm that determines better abilities to sight a target given the time a target spent in your LOS. For example, in my short test above, it's possible that the Germans hadn't spent enough time spotting the Soviet units to get a good bead on them, but after watching their movements for a time, getting a sense for their silhouette against the rainfall out in the steppes they were able to get a good bead. Those Germans you had trouble with in those buildings would've been difficult to shoot at, because they would've been popping in and out as they ducked under a window or other aperture. Thus, it might've been better to just saturate their immediate area with lead or shrapnel via area fire.

[James Crowley]

Steve

Thanks for the almost instantaneous reply!

Good to hear that the re-write will tune the number of turns to the parameters.

And the multiplayer thing was news to me... sounds good. Mind you I can see the "battles" behind the battles..." C Co. commander, you blithering idiot, where was your supporting fire?" and so, and so.

And thanks to yourself and the team for a marvellous and unparalled gaming experience (warts and all).

[dalem]

Originally posted by rune:

Dalem,

Looking at the scenario, if you did not move the maxims, then the buildings are on the cusp of visibility. Light snow and fog, about 225 meters, the absolute closest building is 204 meters. If you were targeting the second row of houses, out of visibility range completely.

If the AI MGs fired at the second row of houses, then the AI is simulating firing at the gun flashes [Yes, this IS coded in] Also, a lot of the houses have a partial viewpoint, where terrain blocks part of the house, where part of the house can be seen.

You can look at this by looking at the scenario in the editor. However, if you have another case, let me know...as I am still looking into this.

Rune

Rune, you rock! I was only concerned with the very front houses (the one with the cemetary in front and one other that looked juicy), and I would say that none of the targets I am whining about were beyond the 220m mark - I had moved the MGs up the edges of their woods, etc.

I had Germans pop up at the forward edges, top floor, of those two buildings, and the AI would not target them on its own, and generally dropped fire after a few bursts, even if the target aspect did not appear to change to me (i.e go prone, turn into a "cross", etc.

I have also experienced this at closer ranges and clear conditions, but that was a QB, etc.

Grisha-

No, I am not knowledgeable about Fuzzy logic per se, but I can follow your example well enough.

I can see your point (and Rune's). Maybe the targets are too far away. Maybe the exposure changed by a point or two during the turn. I don't know. I've seen it happen in a few varied situations where that wouldn't be applicable as far as I was able to discern. Maybe the lesson I am supposed to have learned is that I should have had ALL my MGs (in that situation) Area Firing ALL the time. That's certainly what I resorted to. In that case color me fooled by a fat red target line that looked good to me.

But I really think it's something else, and if it's not, now I can't trust the AI's targetting feedback. Which is the conclusion I came to. Add that to the fact that I already didn't trust the AI's ability to not insanely drive my tanks out of cover into LOS when I never gave them a Movement order of any kind, and that's why I pretty much stopped playing.

Awww look. Now you've gone and made me all depressed again.

-dale

[dalem]

Originally posted by James Crowley:

And the multiplayer thing was news to me... sounds good. Mind you I can see the "battles" behind the battles..." C Co. commander, you blithering idiot, where was your supporting fire?" and so, and so.

This is one of my favorite things to encounter when playing a miniatures game in the Horse and Musket period. Me: "Hey, my flank is caving in, get your cav over here now." Him: "Screw you - I want to countercharge his cav over here." Me: "But that's not part of the main battle!" Him: "It is for me." Ahhhhh, personalities.

-dale

[Battlefront.com]

Dale,

It is true that the TacAI is a bit more cautious with expending ammo than in CMBO. If the unit thinks it has a good shot, it takes it. Otherwise it holds back. Different conditions will dictate what is or is not a good shot. This is very similar behavior to being low on ammo, but for slightly different reason.

As far as trusting the TacAI, you shouldn't have shaken faith. The TacAI is doing its job, and that is managing ammo consumption inline with potential payback. If my units spot something and my guys don't open up, instead of assuming the TacAI is being a moron I assume it knows better than I do and is witholding fire for a reason.

On long distance shots I rarely override the TacAI. When fighting gets in close at critical ranges, I take more control because coordination of fire is far more important and the TacAI is not supposed to be able to do this as well as I can.

There is feedback about the quality of the shot, but nothing overt like text displayed saying "Very Crappy Shot". The reason is that on the very milisecond the turn ended, perhaps it is a decent shot. But one milisecond into the next turn the unit drops and then becomes a totally crappy target, so fire is either withheld or one volly is sent over.

Personally, I think this is big improvement over CMBO's modeling. It was too easy to blow through ammo on fruitless actions IMHO. CMBB is more stingy and I think that is good because when it really counts, I have ammo

Steve

[Boo Radley]

Originally posted by Battlefront.com:

Dale,

It is true that the TacAI is a bit more cautious with expending ammo than in CMBO. If the unit thinks it has a good shot, it takes it. Otherwise it holds back. Different conditions will dictate what is or is not a good shot. This is very similar behavior to being low on ammo, but for slightly different reason.

As far as trusting the TacAI, you shouldn't have shaken faith. The TacAI is doing its job, and that is managing ammo consumption inline with potential payback. If my units spot something and my guys don't open up, instead of assuming the TacAI is being a moron I assume it knows better than I do and is witholding fire for a reason.

On long distance shots I rarely override the TacAI. When fighting gets in close at critical ranges, I take more control because coordination of fire is far more important and the TacAI is not supposed to be able to do this as well as I can.

There is feedback about the quality of the shot, but nothing overt like text displayed saying "Very Crappy Shot". The reason is that on the very milisecond the turn ended, perhaps it is a decent shot. But one milisecond into the next turn the unit drops and then becomes a totally crappy target, so fire is either withheld or one volly is sent over.

Personally, I think this is big improvement over CMBO's modeling. It was too easy to blow through ammo on fruitless actions IMHO. CMBB is more stingy and I think that is good because when it really counts, I have ammo

Steve

OK, now I'm understanding a bit more. In the "Pavlov" scenario where my troops aren't firing even though the Germans are within their cover arcs, it may be that at 230m, the AI figures they'd be wasting ammo and so therefore, holds off on firing. Correct?

[JonS]

Originally posted by Battlefront.com:

</font><blockquote>quote:</font><hr />A new option, stay put, need not have a fixed rule ("only stay put if cover is more than x meters away"); rather, the decision to stay or run to a particular piece of cover would use most of the same factors that the AI currently uses to choose between nearby scattered trees and farther away woods.

True, but through direct experience (i.e. Charles making code changes available to us to play with) it simply didn't work. Units were too easily pinned. I don't know exactly why the game dynamics led to this end result, but it did. And it sucked very badly.

Reasonable people can disagree about when units should (realistically) sneak for cover and when they should stay put, but I think that most people would agree that in some circumstances staying put is clearly the best option; an option that the AI does not consider now.

The problem is that Charles was unable to come up with an adequate rule set to make this work well enough to be included. Again, I can only comment on this to a limited degree since it was a long time ago and I didn't code it. But the implementation now is far better than what it was like before when we tried to do just this.</font>

[akdavis]

bah...misunderstanding

[ November 20, 2002, 08:21 PM: Message edited by: akdavis ]

[dalem]

Originally posted by Battlefront.com:

Dale,

It is true that the TacAI is a bit more cautious with expending ammo than in CMBO. If the unit thinks it has a good shot, it takes it. Otherwise it holds back. Different conditions will dictate what is or is not a good shot. This is very similar behavior to being low on ammo, but for slightly different reason.

Okay. So I'm not crazy. I am a little surprised that this thread is the first I've read of something I would consider fairly major.

As far as trusting the TacAI, you shouldn't have shaken faith. The TacAI is doing its job, and that is managing ammo consumption inline with potential payback. If my units spot something and my guys don't open up, instead of assuming the TacAI is being a moron I assume it knows better than I do and is witholding fire for a reason.

I will carve that on the headstones of the soldiers that expected some suppression fire, I assure you.

-dale

[Massattack]

James Crowley, good point about the need for slightly more game turns per scenario to accomodate the more realistic infantry model. I have found it a little difficult in some of the scenarios to go in for the finish after carefully creating an attack, though some if this may be down to me being TOO cautious!

Anyway good to hear this feature will be improved. BTS, great thread, learned quite a bit from it.

Regards

Massattack

[Grisha]

Originally posted by Battlefront.com:

Dale,

It is true that the TacAI is a bit more cautious with expending ammo than in CMBO. If the unit thinks it has a good shot, it takes it. Otherwise it holds back. Different conditions will dictate what is or is not a good shot. This is very similar behavior to being low on ammo, but for slightly different reason. . . .

Steve,

Glad to see I wasn't too far off the mark. Fuzzy logic is a good thing

[Warren Peace]

Steve,

Since turn limits were brought up, I was wondering something. Does the strategic AI take into account how much time is left when performing actions. For example will it be more cautious in games with large maps and long time limits relative to games with large maps and short time limits? My guess is no, but I am not certain.

Warren

[Michael Emrys]

Originally posted by dalem:

As I advanced my rifle platoons forward (admittedly clumsily-enough) and they came under fire from those buildings, I noticed turn after turn where the Maxims (in command, not Hidden, and not under any fire) displayed the following behaviors:

1) refused to acquire targets by themselves. I could target successfully some Germans but the AI rarely did. And I mean "rarely" - I was paying attention this time.

2) when I would specifically target a unit, and that unit's target became unspotted, that MG would not acquire another target that turn.

Dale, you've just jogged my memory. I used to see this literally all the time in CMBO. It never mattered all that much there, so I never thought all that much about it. I may have even felt it was realistic in some way. Obviously it is an issue of much greater import in CMBB and that's why we are suddenly hearing about it a lot. But I claim it is an artifact of the coding of the first game.

Which isn't too likely to help get a cure for it, but I thought I would mention it just for the record.

Michael

[Bruno Weiss]

Personally, I'm a counter-whiner.