superior tactics video

[Belgolas]

While it seems like it was an easy win in low quality, he started with a full 20 unit stack:
519 Horse
480 Infantry
36 Gun crew
Total 1035

But finished (last frame where you can see unit numbers) with only 11 units in combat:
71 Horse
269 Infantry
12 Gun crew
Total 352

Lost his general & 2/3rds of his force.

Edit: beaten to the numbers...

Only if you don't count the men retreating.



One thing I don't like is at the very end it has a thing that says "plays best on Alienware". Come one they had to accept money from them . Some of the most expensive pieces of garbage ever. There so called water cooling sucks and are way over expensive. Just build your own and save a few thousand bucks. end rant.

Although I like that it says runs great on Core i7 So hopefully they will optimize it for 8 threads.

[Incongruous]

Um check again. The enemy AI had units in 3 ranks deep. Also they have a full battalion square and used it pretty well IMO. If this is what the AI will be like in the final version then it will definitely be the best TW AI I have ever seen. Watch the 720p version. You can actually see what is going on.

This is gonna be the best TW.

I saw a poorly executed battlefield manouvre, I was not impressed, if the AI finds it that hard to move around, it will prove no challenge. I mean I saw units of cavalry mulling around when they could have affected the outcome of the first phase of battle. This was highlighted even more by the insane movement speeds of all units, I knew that we would have to slow them for ourselves, but that was dire indeed.

The demonstrator has proved beyond a doubt that cavalry can dominate infantry, no matter the situation, it does not matter that they take heavy losses, they should not be able to do it full stop. Remeber RTW cav spamming? Players would take massive casualties but it did not matter because their all cav army could win, and that is what matters.

CA was reffering to the demonstrator's own formation, which was to deploy his infantry in very deep formations while on defence, which allowed him to beat back a counter attack

[lars573]

Anyone else notice the Green-coats about 3/4ths into the video?

Skirmishers. Look like 60th american rifles. Going by the forrest green jackets with red facings.

[CountArach]

Um check again. The enemy AI had units in 3 ranks deep. Also they have a full battalion square and used it pretty well IMO.

You misunderstand me completely. I don't have a problem with ranking 3 deep - that's fine. What I care about is there being 10 units stacked on top of each other on 1 hill that are in a bunched-up fight. That is completely ahistorical and not at all 18th Century.

If this is what the AI will be like in the final version then it will definitely be the best TW AI I have ever seen.
[...]
This is gonna be the best TW.

That's what depresses me.

CA was reffering to the demonstrator's own formation, which was to deploy his infantry in very deep formations while on defence, which allowed him to beat back a counter attack

No, actually, but I hated that as well.

The only reason that the Programmer took such high casualties was because he was using his army completely poorly. Charging cavalry headlong into teh enemy apparently does win battles - but you will pay in blood.

[Belgolas]

I saw a poorly executed battlefield manouvre, I was not impressed, if the AI finds it that hard to move around, it will prove no challenge. I mean I saw units of cavalry mulling around when they could have affected the outcome of the first phase of battle. This was highlighted even more by the insane movement speeds of all units, I knew that we would have to slow them for ourselves, but that was dire indeed.

The demonstrator has proved beyond a doubt that cavalry can dominate infantry, no matter the situation, it does not matter that they take heavy losses, they should not be able to do it full stop. Remeber RTW cav spamming? Players would take massive casualties but it did not matter because their all cav army could win, and that is what matters.

CA was reffering to the demonstrator's own formation, which was to deploy his infantry in very deep formations while on defence, which allowed him to beat back a counter attack

Well men can run pretty fast you know. I thought they were a little bit fast but not nearly as fast as RTW's infantry guys that could almost out run a horse. over half his army was cavalry and the majority of the enemy was infantry. The cavalry charged into men that were 3 ranks deep. Obviously they would win that.

Anyways in RTW a full army of praetorian cavalry would own a full army of whatever and maybe loose 100 men. This time he nearly lost all of his men. Although I do agree this guy didn't do too well but you have to remember that most people play at about his level. So the AI is still not as good as a human. Name one game where the AI is as good as a human without using cheats.

[CountArach]

So the AI is still not as good as a human. Name one game where the AI is as good as a human without using cheats.

You would be surprised what some games can achieve. For example Take Command: 2nd Manassas had a very tough AI that was capable of putting in place a good battle plan. You would be fighting one part of the enemy only to have the other wing loop around behind you and strike you in the rear. I had some nasty battles on that game engine.

[gollum]

This is the *improved tactics*?

The AI was rushed at the begining with the cavalry and responded by setting his infantry out and the cavalry in - that wasnt all that bad initself - altough it is debatable if he would be better off rearranging his musket infantry on a line formation (ready eventually to let deadly volleys out) while the Dutch cavalry bought them time instead of letting his musket infantry being charged by the British cavalry and then let his own cavalry in.

This is actually a typical mistake newbies do in mp - they let their gunners be *crunched* between the enemy line that charges them and their own melee line that goes to engage the enemy s. In this way they *lose* their gunners for nothing.

One of the signs to distinguish noobs and decent players in mp is the ability to readily protect your gun troops that become very important at the end phase of the battle because their powerful shots can rout depleted units. Protecting them can take place with fast cavalry (if the enemy attacks with fast cavalry) and with the gunners themselves firing in support OR if the enemy advances his melee line by advancing your own and engage in front of the gunners all the while having your gnners fire into the advancing enemy mass to tear apart their morale before engagement.

In the battle shown we had the first case - a massive cavalry raid - that succeded because the AI presented his msuket infantry instead of protecting it and having it fire while he engaged with his cavalry.

But the worst of it all comes after the *square* demonstration - the AI, like in previous games, ignores the fact that the enemy has the high ground and charges up hill let alone that he does that while being outnumbered. This is very familiar and there is nothing improved about it - its good old incompetent RTW/M2TW AI marching to his doom.

It seems CA programmers find the obvious enthrilling with the enthusiasm of a myops detective nowdays - so much for their *expertise*.

gah!

!it burnsus!

[lars573]

Your entire analysis is very flawed. One, in the 18th centry 90% of the time your shooting and melee infantry are the same unit. And it takes time to switch. Two, his army was heavy on the cavalry so he could take out the AI fast. He zipped in his light cavalry to pounce on the enemy infantry before they could fix bayonettes and form a square. Which is what they are for, and you'll note he took some heavy losses in the light horse units. Three, firing through your own lines with muskets is, simply put, not possible.

He saw that the AI was going to assume a defesive posture, as he had the high ground. He used his light cavalry to delay that until his infantry were in range. Then it was all over.

[Vuk]

Your entire analysis is very flawed. One, in the 18th centry 90% of the time your shooting and melee infantry are the same unit. And it takes time to switch. Two, his army was heavy on the cavalry so he could take out the AI fast. He zipped in his light cavalry to pounce on the enemy infantry before they could fix bayonettes and form a square. Which is what they are for, and you'll note he took some heavy losses in the light horse units. Three, firing through your own lines with muskets is, simply put, not possible.

He saw that the AI was going to assume a defesive posture, as he had the high ground. He used his light cavalry to delay that until his infantry were in range. Then it was all over.

Gollum did make one correct observation though, the AI's suicidal tendecies. During this time period most armies lost no more than 20% of their men in a losing engagement. Most battles were not decisive, and if faced with such a situtation, the enemy army would not have charged up hill at him, but pulled its army together and either waited for him to attack again, or make an ordered retreat. He fights and and runs away, lives to win the next battle he fights. That is what people did back then. They were not such fanatics.

[gollum]

Your entire analysis is very flawed. One, in the 18th centry 90% of the time your shooting and melee infantry are the same unit. And it takes time to switch. Two, his army was heavy on the cavalry so he could take out the AI fast. He zipped in his light cavalry to pounce on the enemy infantry before they could fix bayonettes and form a square. Which is what they are for, and you'll note he took some heavy losses in the light horse units. Three, firing through your own lines with muskets is, simply put, not possible.

He saw that the AI was going to assume a defesive posture, as he had the high ground. He used his light cavalry to delay that until his infantry were in range. Then it was all over.

Thank you for your comment Mr lars - out of curiosity how many mp battles have you fought in TW games and in which ones?

!it burnsus!

[Monk]

Gollum did make one correct observation though, the AI's suicidal tendecies. During this time period most armies lost no more than 20% of their men in a losing engagement. Most battles were not decisive, and if faced with such a situtation, the enemy army would not have charged up hill at him, but pulled its army together and either waited for him to attack again, or make an ordered retreat. He fights and and runs away, lives to win the next battle he fights. That is what people did back then. They were not such fanatics.

That doesn't sound like a very fun game if you ask me.

[gollum]

That doesn't sound like a very fun game if you ask me.

I guess you got what you wanted in ETW then Mr Monk - have *fun* beating outnumbered counterattacks uphill.

!it burnsus!

[Vuk]

That doesn't sound like a very fun game if you ask me.

lol, I disagree. What would often happen is after the first few stages of the battle it would quickly become evident that if an army sat there and fired it would be wiped out, so they would create an ordered retreat. If the enemy pursued them they would lose the advantage of their superior position, so it was hard to have a decisive victory. A good general COULD get an enemy in a position where he could have a decisive victory though, but it was not common. Also, enemy armies would often surrender rather than being wiped out.
Not only would it be more realistic, but it would put more of an emphasis on precombat maneuvers and less on standing and watching.

[Monk]

I guess you got what you wanted in ETW then Mr Monk - have *fun* beating outnumbered counterattacks uphill.

!it burnsus!

Actually what I want for ETW would be an AI that rivals a human, but I realize that desire probably isn't going to be realized.

An AI that doesn't fight and instead retreats off the map at the slightest bit of casualties, however, is not my idea of fun. Total War games have always gone overboard in casualty ratios due to the relatively small unit scale that we have in regards to what was actually used. Assuming that your average army will have around 1300-1500 men, then 20% means 280 men. When 280 men die, the AI retreats. I don't find that idea fun. Would you?

[Vuk]

I guess you got what you wanted in ETW then Mr Monk - have *fun* beating outnumbered counterattacks uphill.

!it burnsus!

Gollum, your manner of address to both Lars and Monk is very sarcastic and inappropriate. You have a habit on all the threads I have seen you post to get very excited about this game. Maybe you should just relax. I would prefer it totally realistic as well, but it IS a game, and that IS how most people want it, and CA is making the game to be fun for MOST people. We can wait for mods when we want realism. Also, many of your suggestions for realism are flawed anyway, so I doubt that the game would be much more realistic if you were at the helm. If you want a game in your image, wait till it comes out and make your own mod. In the meanwhile, treat other members with respect.

[Vuk]

Actually what I want for ETW would be an AI that rivals a human, but I realize that desire probably isn't going to be realized.

An AI that doesn't fight and instead retreats off the map at the slightest bit of casualties, however, is not my idea of fun. Total War games have always gone overboard in casualty ratios due to the relatively small unit scale that we have in regards to what was actually used. Assuming that your average army will have around 1300-1500 men, then 20% means 280 men. When 280 men die, the AI retreats. I don't find that idea fun. Would you?

Armies did not retreat because they took those casualties, and they should not. They retreated when they knew that they could not win, or that if they did it would be at too great a cost. I am not sure if they could get the AI smart enough to be able to judge this accurately, but it would be nice. Again though, I do not think that they should retreat because of taking casualties, but instead because either they think that they cannot win the battle, or they think it will be at too great a cost.

[gollum]

An AI that doesn't fight and instead retreats off the map at the slightest bit of casualties, however, is not my idea of fun. Total War games have always gone overboard in casualty ratios due to the relatively small unit scale that we have in regards to what was actually used. Assuming that your average army will have around 1300-1500 men, then 20% means 280 men. When 280 men die, the AI retreats. I don't find that idea fun. Would you?

It depends - is he outclassed in quality? What is the battle for in strategic terms?

And in addition - even if the AI has to stay and fight - then he is better off using tactics that can win him victory; i dont know what it says to you, but to me being outnumbered and downhill it says retreat somewhere that i have flat ground or even perhaps a slight advantage in height and wait. In no circumstances does it say charge.

Even if the AI was even in numbers it is well known that high ground rushing can succeed when this happens with all the melee line advancing at once - that is maximise frontage and time impact to the enemy.

Well known to all except CA that is i guess.

!it burnsus!

[Mailman653]

In some instances the battle seems to go into a mini cutscene to highlight an action like a charge, and then go back to the game. I wonder if thats a feature of the game or just some nifty editing.

[lars573]

Gollum did make one correct observation though, the AI's suicidal tendecies. During this time period most armies lost no more than 20% of their men in a losing engagement. Most battles were not decisive, and if faced with such a situtation, the enemy army would not have charged up hill at him, but pulled its army together and either waited for him to attack again, or make an ordered retreat. He fights and and runs away, lives to win the next battle he fights. That is what people did back then. They were not such fanatics.

True enough. However, this is Total War! Casualties, along with army sizes, have always been wonky and out of sync with how it was. Truthfully the AI made a bad plan B (consisting of kill'em all! ). Honestly one 6 minute demo is far enough to make any kind of real judgement.

I am however hopeful from what I did see.
-The AI tried to from a defensive position when it was faced with mostly cavalry who had positon on them
-That said attempt could be foiled with a good use of light horse on unprepared infantry

Thank you for your comment Mr lars - out of curiosity how many mp battles have you fought in TW games and in which ones?

Never once have I played mp battles. And there is a good reason why....

[Monk]

It depends - is he outclassed in quality? What is the battle for in strategic terms?

And in addition - even if the AI has to stay and fight - then he is better off using tactics that can win him victory; i dont know what it says to you, but to me being outnumbered and downhill it says retreat somewhere that i have flat ground or even perhaps a slight advantage in height and wait. In no circumstances does it say charge.

Even if the AI was even in numbers it is well known that high ground rushing can succeed when this happens with all the melee line advancing at once - that is maximise frontage and time impact to the enemy.

Well known to all except CA that is i guess.

!it burnsus!

Strategic repositioning is one thing. I was referring to the idea that an AI army should give up if it takes a certain % of casualties. Or at least such a low one in regards to the scale. An AI army retreating to save face if it suffers 60-70% is much more feasible, given the scale of the TW games, imo.

Armies did not retreat because they took those casualties, and they should not. They retreated when they knew that they could not win, or that if they did it would be at too great a cost. I am not sure if they could get the AI smart enough to be able to judge this accurately, but it would be nice. Again though, I do not think that they should retreat because of taking casualties, but instead because either they think that they cannot win the battle, or they think it will be at too great a cost.

Perhaps it was the way you worded it then, I apologize for taking liberties with your argument if this was your original position.

I'm not sure that you can tell the AI to run if the battle looks like it will be too costly to win. I'm sure casualty ratio triggers could easily be built in. (if you suffer 90% casualties, retreat no matter what. ect) but I'm not sure if the battle AI could speak to the strategic AI in the sense to know what is at stake for a specific battle, on any given battle map. With luck one of the CA people will peak in and comment but it sounds really complex, maybe though that's just because i'm ignorant, I dunno. If not, you might get situations where an English army retreats when they are fighting for their capital, ect.

For what it's worth. The AI i saw in action, imo, did look improved than M2's and RTW's mess. It just seems that I had lower expectations than some of you.

[gollum]

Then we approached this from different angles mr Monk - if you mean by strategic repositioning the ability of the AI to make strategic decisions that in turn trigger tactical routines - then this is precisely what i am talking about.

This is the *brain* of the AI on the battlefield - his judgement if you will. What i am saying is that if this judgement is unable to choose from a situation of <outnumbered, in low ground> the obvious that is take a defensive posture - then there is little scope for the developer in making *improved tatcics* videos.

They are better off in letting the bomb explode at the last minute - when the hullaballoo of enthusiam will drown everything into oblivion - until the major bugs find their way to the surface of course.

!it burnsus!

[gollum]

Never once have I played mp battles. And there is a good reason why....

Thank you for clarifying mr Lars - reading your analysis, i advise you to stay on the SP side of the game with the improved AI and for a good reason why too.

regards

!it burnsus!

[Vuk]

Then we approached this from different angles mr Monk - if you mean by strategic repositioning the ability of the AI to make strategic decisions that in turn trigger tactical routines - then this is precisely what i am talking about.

This is the *brain* of the AI on the battlefield - his judgement if you will. What i am saying is that if this judgement is unable to choose from a situation of <outnumbered, in low ground> the obvious that is take a defensive posture - then there is little scope for the developer in making *improved tatcics* videos.

They are better off in letting the bomb explode at the last minute - when the hullaballoo of enthusiam will drown everything into oblivion - until the major bugs find their way to the surface of course.

!it burnsus!

Again Gollum, while I agree with you that the AI needs to think better about how and where to fight, I think that you are being disrespectful and completely unreasonable. If you are not, let me tell you, that is the way you are coming off. I think that you need to calm down.

Strategic repositioning is one thing. I was referring to the idea that an AI army should give up if it takes a certain % of casualties. Or at least such a low one in regards to the scale. An AI army retreating to save face if it suffers 60-70% is much more feasible, given the scale of the TW games, imo.



Perhaps it was the way you worded it then, I apologize for taking liberties with your argument if this was your original position.

I'm not sure that you can tell the AI to run if the battle looks like it will be too costly to win. I'm sure casualty ratio triggers could easily be built in. (if you suffer 90% casualties, retreat no matter what. ect) but I'm not sure if the battle AI could speak to the strategic AI in the sense to know what is at stake for a specific battle, on any given battle map. With luck one of the CA people will peak in and comment but it sounds really complex, maybe though that's just because i'm ignorant, I dunno. If not, you might get situations where an English army retreats when they are fighting for their capital, ect.

For what it's worth. The AI i saw in action, imo, did look improved than M2's and RTW's mess. It just seems that I had lower expectations than some of you.

There is no need to apologize Monk, it was a simple misunderstanding. The reason that I mentioned the statistics is to demonstrate how indecisive many battles were. Of course they would have to take overall goals into account, so retreating from Capitol defenses would not happen much. (though they have happened before, even with the US, though, granted, there was no real defensive army)
While I see improvements also, I still think it is deeply flawed and has most of the same problems. I do not think it is game breaking, but certainly realism breaking.

[Monk]

Honestly it sounds like we are on the same side, just taking the argument to different degrees.

Then we approached this from different angles mr Monk - if you mean by strategic repositioning the ability of the AI to make strategic decisions that in turn trigger tactical routines - then this is precisely what i am talking about.

Yes that is exactly what I was referring to. The ability to identify what he's fighting for and fight according to that idea, I just don't know if it's possible or is attempted in ETW. I don't know if strategic and tactical AIs are linked (I know Military and diplomatic ones are on the campaign map, but dunno about the former) nor do I even know if it's possible. AFAIK we've never had this. There was always campaign AI and battle AI from my understanding.

It would certainly be nice though.

This is the *brain* of the AI on the battlefield - his judgement if you will. What i am saying is that if this judgement is unable to choose from a situation of <outnumbered, in low ground> the obvious that is take a defensive posture - then there is little scope for the developer in making *improved tatcics* videos.

They are better off in letting the bomb explode at the last minute - when the hullaballoo of enthusiam will drown everything into oblivion - until the major bugs find their way to the surface of course.

!it burnsus!

I was impressed with the AI up until it charged the hilltop position. The reason I was impressed is because it demonstrated to ability to change its tactics and adapt to a changing battlefield. I think it was disheartening that it chose to put itself in such a disadvantageous situation, however, and that is where my enthusiasm was cut short.

[Vuk]

Honestly it sounds like we are on the same side, just taking the argument to different degrees.



Yes that is exactly what I was referring to. The ability to identify what he's fighting for and fight according to that idea, I just don't know if it's possible or is attempted in ETW. I don't know if strategic and tactical AIs are linked (I know Military and diplomatic ones are on the campaign map, but dunno about the former) nor do I even know if it's possible. AKAIK we've ever had this. There was always campaign AI and battle AI from my understanding.

It would certainly be nice though.




I was impressed with the AI up until it charged the hilltop position. The reason I was impressed is because it demonstrated to ability to change its tactics and adapt to a changing battlefield. I think it was disheartening that it chose to put itself in such a disadvantageous situation, however, and that is where my enthusiasm was cut short.

There I agree with you. What is at stake should play a huge role in how they fight. Also, the main problem I had with the AI was the charging up hill thing.

[gollum]

Honestly it sounds like we are on the same side, just taking the argument to different degrees.

Indeed.

Yes that is exactly what I was referring to. The ability to identify what he's fighting for and fight according to that idea, I just don't know if it's possible or is attempted in ETW. I don't know if strategic and tactical AIs are linked (I know Military and diplomatic ones are on the campaign map, but dunno about the former) nor do I even know if it's possible. AFAIK we've never had this. There was always campaign AI and battle AI from my understanding.

It would certainly be nice though.

Well Mr Monk, i certainly would be very dissapointed if that is not in - because it was advertised
from ages ago and repeatedly.

I was impressed with the AI up until it charged the hilltop position. The reason I was impressed is because it demonstrated to ability to change its tactics and adapt to a changing battlefield. I think it was disheartening that it chose to put itself in such a disadvantageous situation, however, and that is where my enthusiasm was cut short.

I wasnt (really impressed) - historical formations existed in TW since the MTW era - the routines were scripted in txt files in the main game folder. In RTW for example the senate was fighting with the Camellian battle order (hastati-principes-triarii). This is nothing new - and the formation on the video is in all probability something on this vein.

The true test for the AI came after the charge was beaten off - the programmer says *now they have to come up with something else* and the *something else* is a piece meal charge uphill.

!it burnsus!

[Vuk]

Indeed.



Well Mr Monk, i certainly would be very dissapointed if that is not in - because it was advertised
from ages ago and repeatedly.




I wasnt (really impressed) - historical formations existed in TW since the MTW era - the routines were scripted in txt files in the main game folder. In RTW for example the senate was fighting with the Camellian battle order (hastati-principes-triarii). This is nothing new - and the formation on the video is in all probability something on this vein.

The true test for the AI came after the charge was beaten off - the programmer says *now they have to come up with something else* and the *something else* is a piece meal charge uphill.

!it burnsus!

It is impressive I believe because they are decided HOW they are going to use the formation instead of doing everything the same all the time. I will agree with you though, the AI had an epic failure after that. They say that they have given the modders more freedom this time around though, so hopefully there will be something that they can do.

[gollum]

It is impressive I believe because they are decided HOW they are going to use the formation instead of doing everything the same all the time.

Yes Mr Vuk - however remember that this was a public demonstration of the AI using the square formation. If i was doing such a thing, i would script the AI to use the formation i demonstrate by default. In other words i doubt that the AI had a say in what formation to use at the start of the battle.

!it burnsus!

[Monk]

Yes Mr Vuk - however remember that this was a public demonstration of the AI using the square formation. If i was doing such a thing, i would script the AI to use the formation i demonstrate by default. In other words i doubt that the AI had a say in what formation to use at the start of the battle.

!it burnsus!

The other alternative is that the AI automatically adopts that formation if it sees the enemy has a massive detachment of heavy cavalry, as was the case here. Though I won't get my hopes up.

Like i said before, I seriously hope CA wouldn't tease the community by playing with scripted events after so much hype.

[Vuk]

Like I always say, hope for the best, expect the worst. ~;P That is what I am doing with ETW. Like I say though, modders will probably be able to do a lot for it.