Fixing Phalanx Behaviour

[Red Harvest]

Without the jumping horses I'm getting results such as, greek light lancers cannot beat Liberian spearmen, but Greek heavy cavalry can with two charges and Companion cavalry can with one charge. Poeni infantry can withstand charges from all of these three cav types, and even withstood 4 charges by Praetorian cav with about 1 to 1 kill ratio.

The 46% anti-cav bonus in RTW does seem rather weak. MTW spears had 200% and pikes 300% anti-cav bonus, and STW yari infantry had a 400% anti-cav bonus. In Samurai Wars, yari infantry have a 500% anti-cav bonus.

Puzz3D,

You are committing a tactical error by recharging cav vs. phalangites. The phalanx gets a charge bonus from their primary weapon when you charge at them. The Poeni are easily carved up by Praetorians with stock stats using a single charge followed by simple melee: I just rechecked by changing back the stats and it was 13 kills by the Punic phalanx vs. 78 by the Praetorians (all on large.) The penetration of the phalanx is the deciding factor. The phalanx is slowly driven back at last 30 meters. It doesn't behave as a phalanx after the first few seconds.

Even the Roman Cav are nearly even in their peformance vs. Poeni, winning one and losing one in my tests (and in the one they lost they killed as many as they lost.)

Macedonian cav won with 50 kills vs. 24 in my tests.

Lighter cav with low total defense do get carved up by the +9 primary and +7 secondary attacks of the Poeni. Considering Poeni are in between armoured hops and greek hops. They should be very effective vs. cav.

I have discovered something else interesting. Highly trained cav. have better combat behaviour than untrained. Now this is somewhat logical, but the text file implies that this should is instead a formation "tidyness" issue. I would expect morale to be more important to them than the training factor (since as far as I know there is no way to increase training level with experience or bldg. upgrades.) When I use Barbarian Noble Cav. they get mauled and rout with about 40% losses, despite their attack being considerably better than Roman Cav. (+3) and their morale also being (+2) better--defensive stats are identical.

Now regular spear units don't have the secondary weapon problem of the phalangites. So they are not handicapped in the same way as phalangites.

[Kraxis]

Red, what map are you using? It could be that the barb unit is getting a penalty and the others are getting a bonus perhaps. Remember those 4 figures that indicate terrain bonus? But of course a northern grassplain in summer should not pose any problems to any unit.

But if that is not the case, then it seems the trained and untrained stats are very important. Though I have not noticed anything like it, Triarii still don't seem all that powerful compared to weaker infantry, and they are highly trained.

[BeeSting]

I’m not a huge fan of phalangites. Many here know as well as I do that they were limited in their ground of operation. Anything such as a dead body on the ground could disturb the formation and open up a gap in the wall of spears for enemy swords men to hack through the ranks. As for the unit mass and being a push over, I have yet to observe a phalanx unit five ranks deep giving ground. What I’m wondering is if the number of rank factor was applied to a unit being able to better push their way through. More bodies behind you meant more weight for pushing. Or is this abstracted too in this game with the unit mass factor? If so, I would be really disappointed.

[BeeSting]

Okay, I'm disappointed. The depth of ranks has nothing to do with the stability of a unit in this game. How shallow! I expected more from CA than this.

[Red Harvest]

Red, what map are you using? It could be that the barb unit is getting a penalty and the others are getting a bonus perhaps. Remember those 4 figures that indicate terrain bonus? But of course a northern grassplain in summer should not pose any problems to any unit.

But if that is not the case, then it seems the trained and untrained stats are very important. Though I have not noticed anything like it, Triarii still don't seem all that powerful compared to weaker infantry, and they are highly trained.

Grassy flat land, summer, calm, midday, large settings, medium difficulty. I've got to figure out where to set all that to the default...I get tired of changing it in custom every time.

I just ran a set of tests of Barbarian Noble Cav (Dacian) vs. Barbarian Noble Cav (Gaul.) By inspection all their stats are identical, so I set the Dacians to "trained" and left the Gauls as untrained. I tried two different types of attack, and tried each from both sides. Results Dacians 4, Gauls 0. The key difference seems to be late in the combat where morale is influencing the actual combat results. Earlier in combat I'm not sure that it is doing much in this particular match up.

Keep in mind that Triarii are spear armed with the 0.73 combat modifier. That could have some impact on the outcome.

[Red Harvest]

Okay, I'm disappointed. The depth of ranks has nothing to do with the stability of a unit in this game. How shallow! I expected more from CA than this.

It is not all bad. Rank bonuses should apply for some things (particularly morale and defense) and they should definitely apply to the phalanx, but they should not be absent of mass or unit type. Sword "push" effect should be limited. Skirmisher "push" should be nil. Cavalry push with ranks doesn't make much sense (although individual mass push does.) The mass aspect has some good points, but without some specialized adjustments it causes some problems. Phalangites/hoplites were able to push legionaires back on their heels, even when they lost. Disordering of their ranks was the fatal flaw. Cynoscephale is an excellent illustration of both good and bad in the same battle. One of the wings of the phalanx was unable to properly deploy and form up in time on unsuitable ground, while the other advanced and pushed back the legions. This created a huge gap which the Roman multi-line deployment was easily able to exploit--throwing the reserve triarii (based on the numbers of men and maniples quoted) against the back of the advanced phalanx.

I've been trying to confirm whether or not ranks do anything directly to influence the push. I tried altering defense to very high values so that no killing occurred so I could see the push (def. skill tops out at 63 by the way...) I tried altering masses. I tried very thin formations down to 2 and 3, I tried very deep formations up to 16 ranks. I did about a dozen tests and my general impression is that mass determines the push, not ranks.

Now this isn't a super easy thing to see, because when you have narrow frontage against a wide formation, the flank individuals seem to crescent shape even in guard mode--they advance a bit while the rest of the formation is stopped. And deeper back in the ranks there is rank compression...then later they slowly uncoil backward if the front is stopped. Localized pushes can also occur when the phalangite individuals get into sword play, rather than spear/shield phalanx pushes. Plus deep columns tend to dig a corner in shortly after the fight begins, and then pivot to that side as their opponents shift while using swords.

But one thing I am confident in is that mass has a large impact on the push.

Another way CA could implement this is to modify the "mass" push effect for ranks to some extent. Say that a soldier has a mass of 1. Now for a spear unit the push of this soldier and the men behind him might be 1 * 1 + 1 * 0.75 + 1 * 0.5 + 1 * 0.25 + 1* 0.1 + .... So using perhaps 5 ranks would get all the benefits...or the formula might differ for long_pike (more terms for deeper rank effects.) A different formula could be used for sword (or non-phalanx spear) based pushing with a more rapid decline to zero: perhaps 1*1 + 1*0.5 + 1*0.25. And skirmishers might get simply the first term, and no others while cav might use two terms, elephants only 1, etc. As any unit takes casualties and its ranks decline it might be in danger of losing increments in each file's push.

[Kraxis]

Grassy flat land, summer, calm, midday, large settings, medium difficulty. I've got to figure out where to set all that to the default...I get tired of changing it in custom every time.

I just ran a set of tests of Barbarian Noble Cav (Dacian) vs. Barbarian Noble Cav (Gaul.) By inspection all their stats are identical, so I set the Dacians to "trained" and left the Gauls as untrained. I tried two different types of attack, and tried each from both sides. Results Dacians 4, Gauls 0. The key difference seems to be late in the combat where morale is influencing the actual combat results. Earlier in combat I'm not sure that it is doing much in this particular match up.

Keep in mind that Triarii are spear armed with the 0.73 combat modifier. That could have some impact on the outcome.

Interesting... Very much so. I think we need more testing of this to determine the strength of it and if it is only trained that is odd. Also I think large infantry units are better. Combat is slower there and the risk of flukes are smaller (if a single cavalryman is very lucky he might kill large percentages of the enemy, the same isn't true for an infantryman). Warbands seems to be good candidates (as long as neiter side uses warcry).

And best of all would be combat between non-general units.

About the Triarii, I had taken that into consideration, as they fought Libyan spears and experienced barb mercs. So maybe highly_trained is not stronger.

Oh and mass...
Good idea! Support it fully, if only I can add something.
This increases mass significantly for phalanxes... In fact it increases massively. This will cut back on the effectiveness of elephants, and the historical accounts of them is that they didn't seem to have much problem with plowing down phalanxes. Alexander had great trouble with them for instance.
So I propose that eles get a nullifying effect on the mass from ranks.

[Red Harvest]

Oh and mass...
Good idea! Support it fully, if only I can add something.
This increases mass significantly for phalanxes... In fact it increases massively. This will cut back on the effectiveness of elephants, and the historical accounts of them is that they didn't seem to have much problem with plowing down phalanxes. Alexander had great trouble with them for instance.
So I propose that eles get a nullifying effect on the mass from ranks.

I think you are referring to my mass effects post to Beesting? I don't really propose doing this with the mass values we input presently. I was thinking more of how the .exe should handle it. Don't take my numbers literally either, they are just meant as a "proof of concept" rather than some well reasoned and tested set of values. I propose that the .exe should generate the overall push back automatically. That way the unit will have a given mass, and deeper formations will benefit from it up to a point, but they won't benefit when in shallow formation--presently, they do get this undeserved benefit.

I agree about the elephants vs. phalanx. Everything I have read suggests Elephants worked very well against compact bodies of men, phalanx or legions. Rank effects should be non-existent vs. elephants. The thing that is hard to swallow at the moment is that both enemy horses and "friendly" horses are not nearly skittish enough around elephants. Heavy cav can halt an elephant charge in RTW. Elephants should cut right through them. This can be simulated with mass to some degree...but it really should be done internally, perhaps with a modifier "fudge factor" to apply to horse vs. elephant mass calcs. So while your elephant is 15 mass vs. infantry, it is 2*15 mass vs. horse units. If you look at elephant mass compared to horse mass, the elephant mass is much too low.

[Puzz3D]

Puzz3D,

You are committing a tactical error by recharging cav vs. phalangites.

I controlled the phalanx in my test not the cav. The AI pulls back to recharge because they are loosing the melee. When I controlled the cav I noticed right away that the AI doesn't use the phalanx properly, and the cav won easily. So, all these changes you are making are compensating for bad AI when it controls the phalanx.

[Red Harvest]

I controlled the phalanx in my test not the cav. The AI pulls back to recharge because they are loosing the melee. When I controlled the cav I noticed right away that the AI doesn't use the phalanx properly, and the cav won easily. So, all these changes you are making are compensating for bad AI when it controls the phalanx.

How is it not controlling it properly in this case? It stays in phalanx (except for in the case of very high melee stats for the phalanx), and that is what it should do. And in so doing it should easily beat the cavalry that I charge and then leave in melee. The shift to secondary weapons is happening man by man from what I can see, rather than an errant command. So I don't see anyway that I could be compensating for bad AI in this test.

The mistake is trying to withdraw the cav, yes the AI does this when it has the cav, but it does it at the point that it has a "win" on its hands vs. the phalanx in my control--this is a very bad AI judgement and proves immediately fatal. In some cases it will not try to recharge, when it does not it kicks my phalanx in the teeth. Perhaps I have this wrong, but the AI phalanx control seems proper in this test case, while its cav control is not. So my hypothesis is that your test method is failing to account for bad AI, while I am doing so in mine. If I use the method you are using to test the balance, I will instead end up compensating stats for inept AI. That might be pragmatic in some cases, until a faction leader or heir's oversized bodyguard cav unit slams into your pikes frontally and routs 2 or 3 consecutively. When doing multi unit tests, the AI does snowball with its cav.

The unfortunate part is that the AI control is so incompetent that it is difficult to even conduct logical controlled tests vs. the AI. Play testing and balancing are all compromised by the design of the AI.

[Simetrical]

"Push" should probably be calculated something like this:

1) Each unit would have a mass factor, proportional to its actual real-life mass.

2a) Any two friendly soldiers who are sufficiently close would be considered to be supporting each other, and supporting soldiers would have part of their mass added to the supported soldiers for pushing purposes. However, the angle between the supporting soldiers and the direction of push would cause serious penalties—a soldier at a 90° angle to the push wouldn't help at all.

2b) Alternatively, if this would be too hard on the processor, you could just have a mass multiplier proportional to the number of ranks and the closeness of ranks. Hoplites and phalangites, who stand shoulder-to-shoulder, would have a much better chance of resisting pushing than skirmishers, who would probably get almost no rank multiplier at all due to their wide formation spacing.

3) When a unit charges, its mass factor and rank factor would be multiplied by its speed. If a unit isn't charging, the factors will be multiplied by a different, lower factor (because then its pushing ability comes from actual shoving, not velocity).

4) Horses might gain a bonus to push resistance due to their extra legs, and would get a general bonus to pushing because of their superhuman strength. Elephants would gain a huge bonus to pushing ability, on the basis that they're vastly stronger than anything else on the field, and they'd definitely gain a bonus to push resistance from their extra legs.

-Simetrical

[Puzz3D]

How is it not controlling it properly in this case? It stays in phalanx (except for in the case of very high melee stats for the phalanx), and that is what it should do. And in so doing it should easily beat the cavalry that I charge and then leave in melee. The shift to secondary weapons is happening man by man from what I can see, rather than an errant command. So I don't see anyway that I could be compensating for bad AI in this test.

Staying in phalanx formation isn't the only issue that determines how well the phalanx does against cav. There is a huge difference in perfornace vs cav if the phalanx is not stationary, facing the cav and in guard mode. When the AI controls the phalanx, it will often be moving it when the cav charge hits, and that is very bad for the phalanx.

The mistake is trying to withdraw the cav, yes the AI does this when it has the cav, but it does it at the point that it has a "win" on its hands vs. the phalanx in my control--this is a very bad AI judgement and proves immediately fatal.

I don't see this in my tests. The AI doesn't withdraw the cav when the win is certain. It presses on with the melee and wins. It withdraws for another charge when the win is uncertain which seems correct to me since the cav then tries for a flank charge, although, it's not a full flanking maneuver.

In some cases it will not try to recharge, when it does not it kicks my phalanx in the teeth. Perhaps I have this wrong, but the AI phalanx control seems proper in this test case, while its cav control is not. So my hypothesis is that your test method is failing to account for bad AI, while I am doing so in mine. If I use the method you are using to test the balance, I will instead end up compensating stats for inept AI.

I can beat 54 man Praetorian cav (840 denari, 12/22 att/def) with 81 man Poeni infantry (540 denari, 9/18 att/def) and the only change I made was to remove the horse jumping. So, you are going further with the changes so that when the AI has the phalanx it can beat the cav, but then when the player has the phalanx isn't it going to be too effective against AI cav?

That might be pragmatic in some cases, until a faction leader or heir's oversized bodyguard cav unit slams into your pikes frontally and routs 2 or 3 consecutively. When doing multi unit tests, the AI does snowball with its cav.

The on the spot battlefield upgrades makes it impossible to balance small units. The purpose of those upgrades is to affect a unit's perfomance in the next battle. This was confirmed by the programmer who designed this when he removed battlefield upgrades from MTW/VI v2.01 multiplayer.

The unfortunate part is that the AI control is so incompetent that it is difficult to even conduct logical controlled tests vs. the AI. Play testing and balancing are all compromised by the design of the AI.

I wonder how Creative Assembly did the balancing. I can appreciate that you are trying to get the SP game to play better and the mass effect results are interesting, but I got a huge improvement in the SP campaign by always using auto-resolve. That's the way I play it now, and it gives me a workaround on the siege/savegame bug because I can play enough turns in one session to get to a point where I can save with no sieges on the map.

[GTS]

can you please tell me how to change the colors of some units (roman´s archer auxillia in carthaginian colors). thanks for help

by gts

[tai4ji2x]

I wonder how Creative Assembly did the balancing. I can appreciate that you are trying to get the SP game to play better and the mass effect results are interesting, but I got a huge improvement in the SP campaign by always using auto-resolve. That's the way I play it now, and it gives me a workaround on the siege/savegame bug because I can play enough turns in one session to get to a point where I can save with no sieges on the map.

that's fun enough for you?

[Red Harvest]

Puzz3D,

I've rerun the test through a full 3rd series with vanilla stats. I think I understand the problem, there is something you might be missing about the way the AI uses the cav. It is actually enhancing the AI problems. The Praetorian cav cav are failing to charge in AI control until they are nearly at the end of the pikes (or in some cases in contact with them.) They are getting almost no benefit from that kind of charge. It doesn't matter whether I am walking with the phalanx, or stationary, I seem to get the phalanx charge bonus for the contact. Despite this blunder by the AI, they eventually get past the pikes, and start inflicting casualties at an ever increasing pace. When they get down to about 40 of their own, they inexplicably withdraw, a fatal blunder. At that point they will lose. Note: I have the horse jumping fix, there are no jumps happening.

When I control the cav I smash the Poeni phalanx on contact at full charge (again, no jumping horses.) I am doing nothing fancy. I issue a single attack command at the start of battle, and the unit marches up then automatically charges at the right moment--no other input by me. The Poeni are in phalanx, and my stat tests show that they get a charge bonus vs. the cav, but the cav charge and mass is so high that it breaks through the spearwall on contact. The AI loses about 15-20 Poeni before the cav charge bonus stops.

Yes, we are both facing the same problem and we have a difference of interpretation, but my testing while both using and facing the phalanx suggests that the closest I am getting to fair and square match up is when I use the cav and not when I have the phalanc. And it seems to match how they behave in campaign mode.

P.S. I am going to try resetting the static charge distance to see if that fixes the cav AI charge behaviour. It could be that the AI is issuing a march order, then an attack, then maybe a charge order separately. The command delay would completely bugger their charge if this were true. By comparison I issue a single command and everything else is automatic.

EDIT: Tried increasing the charge distance out to 60 (vs. 40), the cav still marches up onto the pikes. No change in AI behaviour, so the AI is doing this intentionally. The battlefield AI really sucks--it doen't even fit THIS game's combat mechanics. The AI cav control is noticeably worse than the AI phalanx control in 1vs1. In multi unit mode the lack of cohesive phalanx use tips the scales heavily the other way, favoring cav.

[Red Harvest]

that's fun enough for you?

Puzz3D's interest is/was MP primarily (forgive me if that is wrong, Puzz). Since he can't get even mildly competent play from the AI in battle mode in SP, I can understand the move to autocalc to try to play the strategic game.

[BeeSting]

It is not all bad. Rank bonuses should apply for some things (particularly morale and defense) and they should definitely apply to the phalanx, but they should not be absent of mass or unit type. Sword "push" effect should be limited. Skirmisher "push" should be nil. Cavalry push with ranks doesn't make much sense (although individual mass push does.) The mass aspect has some good points, but without some specialized adjustments it causes some problems. Phalangites/hoplites were able to push legionaires back on their heels, even when they lost. Disordering of their ranks was the fatal flaw. Cynoscephale is an excellent illustration of both good and bad in the same battle. One of the wings of the phalanx was unable to properly deploy and form up in time on unsuitable ground, while the other advanced and pushed back the legions. This created a huge gap which the Roman multi-line deployment was easily able to exploit--throwing the reserve triarii (based on the numbers of men and maniples quoted) against the back of the advanced phalanx.

I've been trying to confirm whether or not ranks do anything directly to influence the push. I tried altering defense to very high values so that no killing occurred so I could see the push (def. skill tops out at 63 by the way...) I tried altering masses. I tried very thin formations down to 2 and 3, I tried very deep formations up to 16 ranks. I did about a dozen tests and my general impression is that mass determines the push, not ranks.

Now this isn't a super easy thing to see, because when you have narrow frontage against a wide formation, the flank individuals seem to crescent shape even in guard mode--they advance a bit while the rest of the formation is stopped. And deeper back in the ranks there is rank compression...then later they slowly uncoil backward if the front is stopped. Localized pushes can also occur when the phalangite individuals get into sword play, rather than spear/shield phalanx pushes. Plus deep columns tend to dig a corner in shortly after the fight begins, and then pivot to that side as their opponents shift while using swords.

But one thing I am confident in is that mass has a large impact on the push.

Another way CA could implement this is to modify the "mass" push effect for ranks to some extent. Say that a soldier has a mass of 1. Now for a spear unit the push of this soldier and the men behind him might be 1 * 1 + 1 * 0.75 + 1 * 0.5 + 1 * 0.25 + 1* 0.1 + .... So using perhaps 5 ranks would get all the benefits...or the formula might differ for long_pike (more terms for deeper rank effects.) A different formula could be used for sword (or non-phalanx spear) based pushing with a more rapid decline to zero: perhaps 1*1 + 1*0.5 + 1*0.25. And skirmishers might get simply the first term, and no others while cav might use two terms, elephants only 1, etc. As any unit takes casualties and its ranks decline it might be in danger of losing increments in each file's push.

Great idea!

Also this may fix the wedge formation to really penetrate. But the formula seems a little.... albeit I think I know where you are getting at. I would propose however to use the affect of mass = number of men per given area. Hence it would reflect density as it should, and whatever point that single mass is headed should reflect the weight per area exposed on impact.... times the speed at which it was traveling at the point of impact. I'm sure there's more to it than this, but the point I'm making is to implement reflect real physics and not just abstract it out of the air. So this would largely fix the problem in that you don’t have to individually click the targeted unit to get the abstract affect of a charge bonus in the current game, hence fixing the AI problem of scattering its formation all over the map, leaving gaps for exploitation. Instead, you could make the whole line of units just run and upon impact you get the benefit of modeled real life physics of a charge. The same mass density could be applied to the pushing. A heavy infantry in close formation should have greater mass than a light infantry. So too should a closely formed cavalry unit over an infantry.

[Puzz3D]

The kinetic energy of an object is directly proportional to the square of its speed. To implement realistic physics, objects have to move at realistic speeds and have realistic mass, and right now in RTW they don't. The way I see it the faster a man charges into a phalanx of pikes or the more a man get pushed into them from behind the more impaled he should get. Nothing gets impaled in RTW. If the pike doen't kill him in the game, he stops or slides between the pikes. You can see the three level pikes are close together with a space between the groups of three which is what is allowing enemy men to get into the pike formation.

[BeeSting]

The kinetic energy of an object is directly proportional to the square of its speed. To implement realistic physics, objects have to move at realistic speeds and have realistic mass, and right now in RTW they don't. The way I see it the faster a man charges into a phalanx of pikes or the more a man get pushed into them from behind the more impaled he should get. Nothing gets impaled in RTW. If the pike doen't kill him in the game, he stops or slides between the pikes. You can see the three level pikes are close together with a space between the groups of three which is what is allowing enemy men to get into the pike formation.

I wouldn’t want to sort out that detail without pay. But wouldn't you love to play a battle that implements the mechanics of physics, instead of adjusting some odd numbers in the stats to accurately reflect it?

[Red Harvest]

Update:

I have been conducting more tests and now I'm forced to retract my conclusion on the phalanx charge bonus applying *against* charging enemy units. I had pretty solid looking results in a couple of tests, but the retests have forced me to completely abandon this. Jerome says the stat is doing nothing, and as best I can tell, he is correct.

I loathe posting incorrect test results, so I wanted to get this info out straight away.

[econ21]

Red Harvest, I don't know if you replied to this point in which case I missed it, but I would like to second the request made earlier in this thread that you make public your tweaked unit stats file(s). I really like the way you think - both your observations on the game as shipped and your research-based approach to trying to improve it.

Personally, I am not a great fan of mods that change everything, creating whole new units etc. to make a new game. But I really appreciate ones that take the vanilla game and just make it play better.

I guess it is a work in progress but I suspect you've already made a number of significant tweaks (horse archers, now phalanx) that many people including myself might like to use but are reluctant to thrash around under the car bonnet trying to do ourselves.

[RJV]

Red Harvest, I don't know if you replied to this point in which case I missed it, but I would like to second the request made earlier in this thread that you make public your tweaked unit stats file(s). I really like the way you think - both your observations on the game as shipped and your research-based approach to trying to improve it.

Personally, I am not a great fan of mods that change everything, creating whole new units etc. to make a new game. But I really appreciate ones that take the vanilla game and just make it play better.

I guess it is a work in progress but I suspect you've already made a number of significant tweaks (horse archers, now phalanx) that many people including myself might like to use but are reluctant to thrash around under the car bonnet trying to do ourselves.

Seconded (or is that thirded). If you have anything that you would feel comfortable with sharing that would be great. To be honest, I would have a lot more faith in the battles I'm playing if I were using your unit stats. Your ideas and tests suggest you're desperate to get the vanilla game as good as it can be, and you seem to have the time and enthusiasm to try and get it right. At least then I would feel confident that they have been tested, tested, tested.

Cheers,

Rob.

[Kraxis]

I have been mirroring Red Harvest and his developments, ven done a little change myself. But overall I think you will find that our changes are similar.

I have done more or less what he has, but I have also:
Changed a number of names (Naked Fanatics to Gaesatae, Eastern Infantry to Sparabara...)
Given the horde formation to all low tech barb units (warbands and swordsmen, not spear warband) and low tech horse archers (actually they seem much easier to use now).
Changed the stats of certain units (Iberian Infantry buffed, Hastati nerfed and various barb units tweaked slightly to make more diversity).
Nerfed archery to a base of 5 for non-elites and nerfed range for elites to 150, all except Cretans who are also the strongest archers now (with lower chance of apearing).
Various changes to mercs and their placement.
Upped upkeep of cavalry (tried to let the Scythians off easier) and archers.
Buffed javelininfantry with range and power and certain of them with ammo (you will love Illyrians and Mercs).
Buffed javcav with ammo (to 8).
Added officers here and there, changed a few but in general have not done it too much.
Changed descriptions so that they now fit the names and in some cases so that they fit the tweaks of the barb units.

I'm finding it much more to my liking, but my tweaks to the game is far from done as I have yet to finish a campaign with it and I'm not much in the mood with the siegebug around.

[Red Harvest]

Simon and RJV,

I haven't forgotten the request. I don't mind posting files once I get to a state that seems satisfactory. Like Kraxis says, our tweaks are very similar. His comments and suggestions, along with those of many others contribute to my balancing efforts. My main concern is that some folks might not like the depth to which I've changed things...I've done away with head hurlers, wardogs, screeching women, flaming pigs, and flaming arrows for example.

I'm still testing and trying to play through campaigns on some of this. I have *heavily* modified the descr_strat so that I can face better armies, larger towns, better starting barracks in many--I'm running this as a "provincial campaign." The one negative this has is that I can also build better armies sooner...and unlike the AI, I know how to use them. So in some ways it is different, rather than being that much better.

There are some base level units that I would like to replace/move in the queue (town watch and town militia), but there are some limitations that have prevented me from doing so. As a result I've boosted the town militia (Spain and Carthage) to something short of base warband quality, but at the smaller unit size. I've also increased the cost of both to discourage their use by the AI.

Unfortunately, there is nothing yet that I can do about the mismatch of command stars, or the siege lifting bug, the battlefield AI, or the lack of coordinated aggression by the AI as you eat away at its territory.

[Kraxis]

I think the best you huys can do is to mod the game yourself. That way it will fit you and not me or Red.
I don't like to mod units out, Red does that with a glee. So eventhough we have much the same game there are vast differences in gameplay, and I'm sure we would rather like our own job than the other's (not a bad word about your job Red).

Since I have made several small mistakes alongthe way I can tell you that all my changes (and there are many more than those I have posted) takes about 4 hours to do, as long as you don't make the mistakes (so keep checking the game out).

I have for instance had great trouble with the building file. Apparently the building bonusses needs to be placed correctly with a vengeance. I had a hard time making it possible for the Greeks and Romans to build all of their culture's temples (after level 1 so they can only start their own). I had hoped to make it possible for the eastern faction to do the same but apparently that conflicted with the Parthians so I had to abandon it.

Anyway, I have nerfed the bonus that phalanxes get vs mounted units. I noticed that I was very reluctant to engage enemy phalanxes with my chariots, even from the flanks and rear. That was not the intention. I'm thinking of removing the bonus vs chariots and elephants. Or else it seems the phalanx might become too much of a wall.

[econ21]

Thanks, Red Harvest! I'll keep reading this forum to follow your progress.

I was going to include Kraxis in my begging post, as you are both doing very similar work. I would like to make the changes myself, Kraxis, but right now I'll be lucky to find the time to play a campaign let alone mod or playtest changes.

My main concern is that some folks might not like the depth to which I've changed things...I've done away with head hurlers, wardogs, screeching women, flaming pigs, and flaming arrows for example.

All sounds good to me! It sounds that what you are doing is getting rid of the most ahistorical aspects of the game, but keeping it recognisable.

I kind of lost interest in the otherwise wonderful MedMod when WesW started changing the stats of everything, then introducing whole new units and fundamentally rejigging building tech trees etc. I am sure he made a great game, but I just wanted vanilla MTW with a more competitive AI. With RTW, I am rather resigned to the AI but I like the sound of your tweaks to make it less blatantly ahistorical. [EB may be one root and branch mod I can accept, as I really appreciate the teams focus on historical accuracy.]

[Red Harvest]

Kraxis,

Good point about the phalanx bonuses and one I keep in mind. What I think all of us are after is getting really substantial phalanx bonuses vs horses & camels frontally, but nothing when struck from the rear or flank. There is no way that I can think of to do this. Reading Jerome's comments about defensive skill it might be possible to use defensive skill to make them tough from the front but very vulnerable from behind. Unfortunately, it is not selective in applying to cav. However, right now you might find in vanilla form, stat-for-stat that infantry will beat a phalanx frontally. (And I would really like to find out if def skill and armour are stacked on the shielded side...or if only one applies to each side. I suppose they stack in front.)

About removing units: I'm only adding comments in front of each line to remove them from the build list, so that they should be easy to put back in. (One reason is that I'm considering replacing some of the units so that their place in the queue is still used, but for a more rational unit.)

[Kraxis]

Yeah I was rather surprised about the defensive skill bonus. That means that certain quite good units are absolutely crap from the left (if tehy have no shield). I have a hard time no believing him, but my gutfeeling is that he is not right. I haven't noticed the difference between left and right yet.

Btw, I see a lot more infantry in my game, especially the Egyptians bring lots of infantry. Sadly they are mostly Nubians, who are after all only Militia Hoplites with another skin. Also there are fewer horse archers of the Parthians and lots more Eastern Infantry. But at least they have been very aggresive (taking Dumatha for instance).
A great result has oddly enough been a limit on the usage of Peasants, and I have even lowered their upkeep to 50. But then again I have lowered their mass to 0.6.

I actually think the AI takes more than just cost into account. And that might explain why I only saw my first Pontic Phalanc Pikemen in this game as they were cosidered too weak by the AI due to their small size, but now with the benefits the phalanx units have gotten it might alter it a bit.
Also, the Egyptians have let back a bit on the chariot archers and brought plenty more normal chariots. It is highly interesting to face a massive army of phalanxes with a sprinkle of cavalry and chariots. They even act as hammer and anvil at times, just too bad that the anvil is too weak.

[hrvojej]

Has anybody tried my earlier suggestion about changing the attack frequency/lethality for pikes only, while leaving the secondary attacks intact? IMO it works great, since phalanx is able to waste anything frontally. Most of the things are not able to get past the pike points that way.

[Kraxis]

Has anybody tried my earlier suggestion about changing the attack frequency/lethality for pikes only, while leaving the secondary attacks intact? IMO it works great, since phalanx is able to waste anything frontally. Most of the things are not able to get past the pike points that way.

You mean to something like 1.5 or what?