What's the point in elite units?

[The Stranger]

I do disagree, because that isn't my experience. Elites are rarely much better at doing the same job, and they certainly aren't better than two regular units at doing it. Two units of Hoplitai will hold a line much more effectively than one of Hypaspistai, not least because one can pin and the other flank. Especially when you have the force multiplier of your own cavalry, doing the unit-breaking from the rear added into the mix.

The better morale is often irrelevant compared to the impact the general is having, and the difference between 11/12 (the usual morale of regulars) and 15 (usual morale of elites) not anywhere near as important as between 8-10 (usual levy morale) and 12.

thats not a fair comparison, you should compare 2 elites to 4 regulars, and not 1 elite to 2 regulars. manouvers are another extra thing to consider, and its not fair to take that out of the eqation for 1 unit but noth for the other.

No, clearly you don't. Levies are an irrelevance since they are not capable of doing the same job. Whereas elites are only at best 25% better than regulars stats-wise, for double the price.

The only economy elites are providing is in taking up less slots in a 20-unit stack than the equivalent cost in regulars.

im not comparing elites to levies, im comparing how good elites do vs something and how good regulars do vs that same thing. its the obvious choice for that something to be levy units, because i can't balance the ratio of regulars vs another regular or an elite vs another elite. so i let 1 regular fight 1 levy and see what the casualties are and what the cost is for that unit, and i do the same for the elite, and then i compare the two. then i look at what point the levies will overwhelm and start to defeat the regular/elite and again i compare the results.

the conclusion was that when for example you can put 1 regular vs 1 levy, then cost-wise the regulars will suffer less casualties than the elite. but when you are in a chokepoint where you can't put two regulars but you can use 1 elite, then the elite (some of them atleast) will do better than even 2 regulars (put behind each other) because the elite will break at a much much later point.

its not that hard.

[vartan]

Some four years ago, we had similar questions to that posed in this thread. Tests proved to be much more efficient than theoretical discussion. Hypothesis. Method. Data collection. Analysis. Conclusion. Good times.

[The Stranger]

i did the tests way back. cant do any again even if i wanted to, disc is broken.

[Ludens]

More to the point, two units of regulars will do a better job of it than one elite. I'd rather have two units of Hoplitai than one of Hypaspistai (or for that matter a unit of decent cavalry - more cavalry is better for flexibility). That's more tactical flexibility. Furthermore, they're recruitable all over the place, and replacable with an identical unit of mercenaries, who are again available all over.

And that's the third reason why elites aren't as useful in EB as they were in reality. In the R:TW engine, experience has a major effect. The stat difference between regulars and elites isn't that big in EB, so regulars with a few bars of experience are as good as elites. And because losses in a regular unit are easier to replace, a veteran regular is actually more useful than an elite unit. IIRC regulars also gain experience faster than elites.

(Personally, I think the R:TW XP mechanism is unbalanced. We should give all units two or three bars of experience straight out of training, and slash base stats accordingly. This will mean that units will gain experience more slowly, and the max. effect of experience is lower. IIRC this idea was considered for EB1 at some point, but the team decided not to go with it.)

sigh, you dont get the concept of cost-effectiveness.

I am pretty sure that elites haven't been balanced on cost-effectiveness. I believe it's the reverse: you pay a premium for the extra ability - so it's only worthwhile to use them in critical positions or when you want maximum combat power in a 20-unit stack. In other words: they're only worth the premium in a one-on-one fight. If you're not fielding large (> 12 unit) armies, regulars are a better buy, since you you get twice as many men for only a ~20% reduction in survivability power. That's how it should be. Where it goes wrong is that, for a variety of reasons (poor A.I. tactics, the effect of XP), regulars will perform nearly as well as elites in critical positions, and are far easier and cheaper to replace.

[QuintusSertorius]

And that's the third reason why elites aren't as useful in EB as they were in reality. In the R:TW engine, experience has a major effect. The stat difference between regulars and elites isn't that big in EB, so regulars with a few bars of experience are as good as elites. And because losses in a regular unit are easier to replace, a veteran regular is actually more useful than an elite unit. IIRC regulars also gain experience faster than elites.

(Personally, I think the R:TW XP mechanism is unbalanced. We should give all units two or three bars of experience straight out of training, and slash base stats accordingly. This will mean that units will gain experience more slowly, and the max. effect of experience is lower. IIRC this idea was considered for EB1 at some point, but the team decided not to go with it.)

I hadn't even considered experience, that's a good very point. It's trivially easy to get units you can recruit at one or even two chevrons at the beginning (using a Type II or Type IV government, combined with a Gymnasium/it's equivalent).

In all those migrated faction games I've played, I almost never had the option of building a Type I (Homeland) government, so as a side-effect I was always getting experienced troops (usually with a Blacksmith in the recruiting centre, too).

I am pretty sure that elites haven't been balanced on cost-effectiveness. I believe it's the reverse: you pay a premium for the extra ability - so it's only worthwhile to use them in critical positions or when you want maximum combat power in a 20-unit stack. In other words: they're only worth the premium in a one-on-one fight. If you're not fielding large (> 12 unit) armies, regulars are a better buy, since you you get twice as many men for only a ~20% reduction in survivability power. That's how it should be. Where it goes wrong is that, for a variety of reasons (poor A.I. tactics, the effect of XP), regulars will perform nearly as well as elites in critical positions, and are far easier and cheaper to replace.

Indeed, the most illustrative unit to demonstrate that they're not costed on the basis of effectiveness is the Hoplitai Haploi. A levy unit which is a lot better than many other levy spearman units (such as Pantodapoi, Pontic/Slavic Light Spearmen and others), yet is also cheaper. I try to avoid using them for anything but garrisons, because it feels like an exploit, but for some factions they're the most cost-effective line troops.

[vartan]

(Personally, I think the R:TW XP mechanism is unbalanced. We should give all units two or three bars of experience straight out of training, and slash base stats accordingly. This will mean that units will gain experience more slowly, and the max. effect of experience is lower. IIRC this idea was considered for EB1 at some point, but the team decided not to go with it.)

Where were you when Creative Assembly was developing Rome: Total War? Where were you, Ludens, where?!! Is this really something that could have been modded into EB1 on the Rome engine? I didn't know that. The entire experience mechanism could have been designed better...what is it in RTW, linear? With a small slope at that.

[The Stranger]

And that's the third reason why elites aren't as useful in EB as they were in reality. In the R:TW engine, experience has a major effect. The stat difference between regulars and elites isn't that big in EB, so regulars with a few bars of experience are as good as elites. And because losses in a regular unit are easier to replace, a veteran regular is actually more useful than an elite unit. IIRC regulars also gain experience faster than elites.

(Personally, I think the R:TW XP mechanism is unbalanced. We should give all units two or three bars of experience straight out of training, and slash base stats accordingly. This will mean that units will gain experience more slowly, and the max. effect of experience is lower. IIRC this idea was considered for EB1 at some point, but the team decided not to go with it.)



I am pretty sure that elites haven't been balanced on cost-effectiveness. I believe it's the reverse: you pay a premium for the extra ability - so it's only worthwhile to use them in critical positions or when you want maximum combat power in a 20-unit stack. In other words: they're only worth the premium in a one-on-one fight. If you're not fielding large (> 12 unit) armies, regulars are a better buy, since you you get twice as many men for only a ~20% reduction in survivability power. That's how it should be. Where it goes wrong is that, for a variety of reasons (poor A.I. tactics, the effect of XP), regulars will perform nearly as well as elites in critical positions, and are far easier and cheaper to replace.

i know they havent, thats why i have been saying some elites. their cost has nothing to do really with how good they are, as roman elites (though they dont really have any real elites) are considerably cheaper than most other factions but alot better for their cost. the dacian komatai are also very expensive and pretty bad (or dont work well anyway). there are a few more examples too.

i didnt say that elites are cost effective, ive always said that regulars are more cost effective, ive pretty much said what you have said, that elites are worth their cost in special cases. and in such cases, theyre worth their weight in gold. the reason why i said he doesnt get cost-effectiveness, is because hes insisting that i test how well elites do vs regulars, while that doesnt necessarily say anything about their cost-effectiveness. its way easier to test them both vs 1 levy and compare how effective they slaughtered the levy for their cost. this guy tested it in regular vs elite, https://docs.google.com/document/edi...axITfr1Q&hl=en

anyway, its also subjective, and since he also plays on medium battle difficulty, i would probably argue you wouldnt ever need anything else than levy units, since depending on the civ, theyre often as cost effective + all the easy access bonusses. the ai is so bad you could conquer half of the world with 5 low tier horse archer, 1 general and 1 infantry merc you recruit on the spot to conquer cities.


also a reason that elites arent much better is because how EB balanced units, a spear wielded by an elite doesnt have much higher attack or charge than one wielded by a regular. nor do they often have alot more defensive skills or a better shield value. they really differ much when it comes to morale and armour value. i think atleast in defensive value there is some room for better balance (some elites have a lower value there than even some levies)

as for gaining experience, ye the system is flawed, it should be a procentual bonus on the base skill of the unit, so when a elite ranks up 1 experience point, his boost is bigger when compared to the boost a levy gets from his experience. or something in that general direction

[The Stranger]

did some more testing, not real tests but just 1 time battles.

Test 1 with me commanding the dosidataskeli
2 dosidataskeli vs 4 thorakitai (upkeep: 3796 cost: 949 vs upkeep: 1910 cost: 478 for a single unit)
243 vs 649: 566 kills and 62 losses

The battle was pretty straight forward, 3 units attacked my dosidataskeli head on, 1 unit went for a rear attack. All I did was keep my units in a line and turn around at the last moment to catch the flanking enemy. This test is a bit skewed because the AI is uncapable of using troops that throw missiles before charge. But I'm pretty sure that the Dosidataskeli would still win this fight even if javelins were properly used.

Test 2 with my in charge of the thorakitai:
2 dosidataskeli vs 4 thorakitai (upkeep: 3796 cost: 949 vs upkeep: 1910 cost: 478 for a single unit)
649 vs 243: 135 kills and 507 losses

At the end my general got killed and the units broke and fled. The battle was already lost however, with 1 dosidataskeli down to 40, and another still left at 80, there was pretty much no way to win it anymore. Around 60 my thorakitai started routing one by one, but they all rallied again to rejoin the fight. I didnt try to go for a full surround and I used my javelins sparsely to simulate the AI. I must add that the dosidataskeli didnt go for the most direct route to attack me and as a result started the fight when already "very tired". This undoubtedly affected the fight in my favor and is probably why the losses are alot higher.

Tried again with 1 dosidataskeli vs 2 thorakitai and the reverse of that. Results were the same, the dosidataskeli win, even when surrounded. They will have around 50 to 90 men left, and the thorakitai will start routing around 40-60 men left in a unit.

Need to test more and actually vs a player, but at face value it seems that some elites, atleast at huge settings, can be more cost-effective than a regular, even when facing regulars, as long as a full surround can be avoided. So elites are worth it when theyre flanks are being protected (by elites or other units) or when fighting in chokepoints or when fighting small battles where manouvering isnt as important.


Test 3 with my in charge of the epilektoi:
3 epilektoi vs 7 hoplitai (cost: 3123 upkeep: 781 vs cost: 1367 upkeep: 342 for a single unit)
367 vs 1135: 981 kills and 226 losses

It was a bit hard to balance these, as in upkeep the ratio was 1:2, but in recruitment cost it was more 1:2,25. I decided to do 3 vs 7, which may have influenced the battle a little bit. However the AI didnt really use their general much, they tried to keep it out of combat in most of the battle, so I think it was ok. They did go for a full surround, which means that numbers offer more advantage. The hoplites started breaking later than the thorakitai despite having 1 morale lower, perhaps because of their disciplined trait. When most of the hoplitai had broken the general started to join the fight, but it was already a bit too late. The elites lost 62% of their men, the regulars 82%. This was pretty close, if all units wouldve been used by the AI and with a full surround, the hoplites may just win the fight, of no surround would be achieved, I'm confident the epilektoi would clinch the victory.

Test 4 with my in charge of the hoplitai:
3 epilektoi vs 7 hoplitai (cost: 3123 upkeep: 781 vs cost: 1367 upkeep: 342 for a single unit)
1135 vs 367: 315 kills and 985 losses

The hoplitai got defeated, but it was very close. I went for the full surround, and unlike the AI i used all my units to do so. At first it seemed like I was going to win, but the more men died, the less likely this became as there werent enough soldiers left to keep the full surround intact. When at about 80% losses for both sides I lost my general, 2 hoplite units had already routed at that point (both only started routing at around 15-25 units left). It didnt cause a chainroute though, so I dont think it affected the battle much. In the end all my units routed (all around 15-25 except for the last one) and the battle was lost.

It seems epilektoi are also worth it, even when surrounded, but theyre alot less cost-effective than the dosidataskeli (which, granted, have insane stats and are easily the best infantry unit in the game)

[athanaric]

the dacian komatai are also very expensive and pretty bad (or dont work well anyway). there are a few more examples too.

I take it you mean Komatai Epilektoi, seeing as the regular Komatai are very cost-effective.

[The Stranger]

I take it you mean Komatai Epilektoi, seeing as the regular Komatai are very cost-effective.

ye meant those


Test 5 with me commanding the Rhomphaiaphoroi
2 Rhomphaiaphoroi vs 4 hoplitai (cost: 2899 upkeep: 725 vs cost: 1367 upkeep: 342 for a single unit)
241 vs 649: 535 kills and 87 losses

2 hoplites cost slightly less than 1 Rhomphaiaphoroi to recruit and also slightly less upkeep. To balance this out i gave all 4 hoplites +1 attack upgrade, this basically made the recruitment cost completely equal in multiplayer. The fight was pretty straight forward, i had both Rhomphaiaphoroi charge 2 hoplites, which got pretty easily massacred and the fight was over pretty quick, the general didnt die but routed as one of the first. I also tried with 1 Rhomphaiaphoroi vs only 2 hoplites, this was alot closer, but still the Rhomphaiaphoroi won with about 40 units left.

Test 6 with me commanding the hoplitai
2 Rhomphaiaphoroi vs 4 hoplitai (cost: 2899 upkeep: 725 vs cost: 1367 upkeep: 342 for a single unit)
649 vs 241 : 153 kills and 492 losses

The ai did a significantly worse jobs in using these units somehow. But they still managed to win. I engaged with 2 hoplites head on and with 2 coming from both flanks. The ones that charged head on got massacred while the flankers stayed relatively intact. At around 20 units left my general unit lost his life and broke. The other center hoplite unit also broke at around 15. The flankers broke around 60 or so, but there were still about 90 Rhomphaiaphoroi left and only about 120 hoplites. so the Rhomphaiaphoroi wouldve won anyway.


just discovered a gem of an elite unit

Test 7 with my in charge of the Cordinau Orca:
3 Cordinau Orca vs 7 hoplitai (cost: 3151 upkeep: 788 vs cost: 1367 upkeep: 342 for a single unit)
364 vs 1135: 914 kills and 124 losses

Balanced them similarly to the epilektoi with at 1/2,5. Which is slightly in favor of the hoplitai. The cordinau orca did a much better job than the epilektoi though. Their stats are simply amazing at 15 attack 30 defense and they were at 30% losses while the hoplitai had suffered 70% losses already. At this point the general died and it caused a rout which later rallied. I dont think it mattered too much, the cordinau had already won. I think the ai did a poor job at engaging though and if they werent so clumsy the cordinau wouldve suffered more losses for sure. But not over 200 i think, perhaps not even over 150. Amazing unit.

another gem

Test 8 with my in charge of the kluddargos:
2 Cordinau Orca vs 6 hoplitai (cost: 3109 upkeep: 777 vs cost: 1367 upkeep: 342 for a single unit)
243 vs 973: 813 kills and 100 losses

1 kluddargos tied down 3 hoplites each and chewed through them. They were at 40% losses when the hoplitai were at 80% and the enemy general died. All were routing not much later. They took on 6 hoplites with ease and may have been able to even take down 7. Which i think says enough about their strength and cost-effectiveness. Ofcourse the hoplites are heavily armoured and the kluddargos have an armour piercing bonus, probably haploi wouldve done better relatively.


well maybe its my game, and there is something wrong with it (would be strange cuz clean install)

but for me alot of elites are beating levys and regulars both actually winning the battles and being more cost-effecienct. elites are winning even when its a single elite vs 2 regulars or so, but do significantly better when its 2 elites vs 4 regulars or so. also they get increasingly better when the unitsizes dwindle, because of diminishing returns.

some elites are not much more cost effective than regulars, but some are alot more, even in the open field. Notable names are: Gaesatae/tindonatae, the Dosidataskeli/TAB/Dubosaverlacica, the cordinau orca (solduros/rycalawre/hypaspistai have similar stats but have a secondary spear, which screws them up), kluddargos, all roman elites (not so much because they have amazing stats, but because they are amazingly cheap for their stats and unitsize), any horse archer elite. i havent really tested cavalry, but horse archers pwn all, and armored horse archers pwn unarmored horse archers.