[3.9] Some random troop tests

I ran a few tests on 3.9 yesterday to get a feel for some of the new and rebalanced units. In particular, I wanted to get an idea of the Marleons Heavy Cavalry's performance, as this is a unit I pretty much never use. Most of the tests came out how I expected, but there were a few surprises:


1) The new merc cavalry are great! I will definitely be using these guys if I come across them. In general, I love the balancing of the new mercenary units. Solid in the early game, when party size is at a premium, and a useful niche troop in the late game when you need to surge defenders into a castle for a siege defense.

2) Marleons heavies are basically on par with most faction knights and directional rogue knights. No real surprise there.

3) Horse archers do terribly in F1-F3, of course

4) Are D'Shar Noble Cavalry just terrible? I never use these guys, too hard to train... not sure what purpose, if any, they are supposed to serve.

5) Ghazi Reavers, on the other hand, are awesome, as we all knew already

6) I wouldn't put much stock in the relative performance of the different KOs. The tests are pretty random for one. Second, the test was against a D'Shar army of 175 men. D'Shar carry a lot of ranged weapons which can add to the randomness of the fight.

7) I wasn't surprised at the terrible performance of the Jatu Lancers in this kind of blob fight, but I was surprised at the poor performance of the JBR. I always assumed the Battle Riders were a KO-caliber sort of troop... I guess I was wrong?

The big takeaway for me was how great the new merc cavalry are. Great armor and weapons, plus a chance of throwing weapons leads to a great unit even though their horses are crap and their proficiencies are still a step below KO troops. Only question is, are they worth the wages??

Wow you also noted expected renown by the test that’s way useful.

Dshar noble cavalries s are Mtd-archer troops, if you order them charge, give the order “hold your fire” for compare, in fact in real battles you can benefit a lot from them as a Dshar lord

Just a two advices when doing such tests:
- watch for your FPS and if below 30 restart Warband - it impacts warriors performance,
- even in the same terrain, with the same opponents, you will get different results on different hardware configurations.

IconracI said:

Just a two advices when doing such tests:
- watch for your FPS and if below 30 restart Warband - it impacts warriors performance,
- even in the same terrain, with the same opponents, you will get different results on different hardware configurations.

Click to expand...

wtf really? What causes that to happen?

thermocline said:

IconracI said:

Just a two advices when doing such tests:
- watch for your FPS and if below 30 restart Warband - it impacts warriors performance,
- even in the same terrain, with the same opponents, you will get different results on different hardware configurations.

Click to expand...

wtf really? What causes that to happen?

Click to expand...

1. The low fps is due to too many reloads in a single game session which cripples down your performance (memory leak) by a lot even on high-end computers. The only solution is to restart the game. But you probably know about this already.

2. Different CPU-s go through the equipping routines differently, which may end up in troops having very different equipment ratios and melee weapon selection under the exact same circumstances. This impacts battles and test results by a lot.

The trends however are at least consistant on the same computer. So for example you won't get too different results on the same computer under the same circumstances and if you got more throwing weapons for the Empire Immortals (e.g. 25/50 on average), you'll also get more throwing weapons for Kraken Riddari on average compared to the guy who had lower amount of throwing weapons for these units.

Another simple example is that I had 34/50 Empire Pikemen with Empire Halberds on average (the remaining 16/50 had Ashwood Pikes), while Latis had the opposite ratio on average test results: 17/50 with Halberds and 33/50 with Ashwood Pikes.

But troops usually have a bigger weapon selection, which complicates things further. Konrad often had to average the test results of @Latis/Lag'ier Barlida, himself and me to get somewhat usable rankings for the different KO knights. Although I've mostly done equipment ratio and selection tests, while Konrad and Latis did the majority of troop tests.

Don't get me wrong, tests are always good to have, share and discuss about, but don't always tell the whole story.

P.S.: those are some interesting test results, thank you for sharing them!
- Knights of the Phoenix did very well for you, I guess it's due to D'Shar units having only cutting melee weapons, which performs worse against heavily armored troops
- Radiant Cross and Dragon Knights also did a surprisingly good job on your PC (especially the latter)
- Lady Valkyries, Jatu Battle Riders and Marleons Heavy Cavalry did worse than expected indeed (Valkyries were ranked #3 in mounted KO knight tests overall. Including sieges too though...)
- as for horse achers, Latis is right, compare them to melee cavalry only when you order them to 'hold fire'. Otherwise they will suck due to Warband's default horse archer AI. But turning on the advanced horse archer AI in mod options will make them god-tier, which will pull them to the other side of the ranking, which I don't suggest to do for tests either

I did know about the save/load memory leak. I didn't know about that CPU randomness though. That's super interesting. As you guys know I've been posting my own troop tests since POP 3.6--interestingly, all of my tests until this one were on a different computer! My old tests all showed Dragons as terrible troops. I guess the computer I'm using now gives Dragons better equipment

Unless Dragons got buffed since 3.6, that is!

thermocline said:

My old tests all showed Dragons as terrible troops. I guess the computer I'm using now gives Dragons better equipment

Unless Dragons got buffed since 3.6, that is!

Click to expand...

Dragons were buffed a little in v3.8 and then considerably in v3.9. Almost all other KO knights were buffed as well beside receiving a visual overhaul. In order to tighten the gaps between them, i.e. make those who sucked before not suck anymore:

PoP Wikia said:

- All Gold & Iron Gauntlets changed to Steel Gauntlets
- Cross Hilt Noble Sword changed to Ravenstern Longsword, Knight's War Axe and Great Sword (now each will have one of the 3 weapons rather than all have the Cross Hilt Noble Sword)
- Charger changed to Grey Draped Warhorse (now has this one and Ravenstern Draped Warhorse)

Click to expand...

I think this test are missing the point. I look on many horse archers troops not as a cavalry, but a dual purpose unit. A unit that could do well in fields battle and at the same time been a reasonable archer unit in the siege and then when run out of arrows been a good melee unit. Adventures are probably the best example of this.
That how I use them. I add them into my army when I expect to take castles, in addition to melee unis and pure archers. They give you the flexibility to perform well in both types of battle.

They tend to be extremely expensive, so if I expect only to attack lords on the field, I use pure nonmissile cavalry in addition to a mixture of shield melee with polearm melee and some archers.

So, test you did does not give enough information.

Polearm melee does not perform well in sieges, so If I only expect to siege I use pure shield melee.

Of course horse archers are more versatile, but the point was to get actual numbers on brute force.

I was most surprised by the fact the even mtd-rangers did better than ghazi stalkers. I suppose the singular charge+blob grind also greatly punishes cavalry with only spears and no swinging weapons. Such units might do better to charge-retreat-charge. This includes d'shar noble cav, and rangers of clarion call as well, I think.

Battle of the level 30 trash...


I wanted to see how the various level 30 trash units that upgrade to something good (adventurers, heretic worshipers, sarleon squires, etc) matched up. No huge surprises, but I did notice how well the Cobra Warriors did at least in an environment where they had a numbers advantage over the enemy. Adventurers were also less useless than I remember them being.

One important thing to note is that infantry overperformed due to the AI. When I F1-F3ed cavalry, the AI kept all its forces in a blob and my cav were fighting the whole 88 unit army at once. With infantry, the AI charged their cavalry first, which were eaten by my infantry, then their infantry, and finally it was just their archer line against my blob.

Overall, all of these filler units seem pretty well balanced, which is really kudos to the devs, especially for a tier of units that isn't paid much attention by players or analyzed closely. There are trade-offs everywhere, which leads to interesting gameplay and choices for the player. Some will produce prisoners for you, but are either worse in absolute performance (Sarleon MAA) or less cost-effective (Mercenary sergeants). Some upgrade to something nice (D'Shar reavers), but perform worse than a unit that doesn't upgrade (Cobra Warriors). Some upgrade into something nice and are decently powerful, but are super expensive (Adventurers). Some perform poorly on performance and cost (the second-tier mercs) but are the only ones that can be mass recruited instantly. Some are pretty bad, but can basically be recruited for free (Maiden cavs). And so on.

You should add the level they train from.

So, out of my head after a pause of 1 month or more ... looking at those tables i would say that :

Merc Cav. are ok, no suprises there but if you want potential you go for Adventurer, even if some might die before HA.

I think Cobra perform so well because of solid gear and some shock throwing from horseback before they actually engage in melee.
I don't remember if they level up from armsman snake but if they do i still prefer the snake xbows to cobras at least in some situations.

Heretic worshipers are ok cavalry, on par with most other.

Maiden cav. aren't worth keeping if you think Adventurers->HA cause they level harder than those and perform worse.

Jatus are better at lancer level.

Huscarls are much harder to train than kierguard and from fierdsvain i would make spearmaiden. Kierguards are ok. Sarleon man at arms
only good in siege defence for choke points they otherwise don't have the armor.

For some empire/vanskery/d'shar infantry you gotta use their gear by telling them to stop a bit before the enemy, let them throw what
they have. Or keep them in squares. I think that if you tell them to charge or if they engage too soon, they don't get to throw and some
skill/gear is wasted.

For the elite table:

No way phoenix are better than HA. You're prolly seing that because both are well armored (with phoenix maybe being better)
but you're proly not using the HA from up the hill before engaging to let them use the arrows they have. HA has much more offensive pottential
and if you're using them properly 30-70% enemies never get to hit them. Phoenix are forced to do one thing and they performed it ok (the F1-F3).
As you said, just doing F1-F3 is not a fair thing. Because Phoenix are made for F1-F3, sort of juggernauts just to charge.

A couple of the others are also made for F1-F3 like lions/sarleon/dragons(but worse gear)/SLC.

Clarion call and even empire knights  ... not F1-F3. Empire knights look sturdy and they are but still they come from lvl 15, very fast to lvl up
and i think very good for how fast you get them.  And think about going CKO later, very fast conversion cause they train fast noble->knights->CKO.
Also consider their horses, some are solid horses other are fast horses. Those with fast hoses are for flanking rather than for frontal assault.
I let one troup front charge while I move fast jatus or d'shar on flank. When the 1st troup has engaged the enemy I let the other one attack as well.
Maybe also try to enable horse archer AI, see if the results will be different.

Some units are better used for flanking or from up the hill or from holding position even.

Yep, trade-off exists in PoP’s balance everywhere

Nice!

Would be particularly interesting to see the results from prisoner-takers...

Some of those are prisoner-takers. The right two columns are number of prisoners taken (total), and % of units out of 88 that were taken as prisoners. I tested Mystmountain Shamans exactly for that reason--they are not level 30, and they don't upgrade to or from anything, but I wanted to know how many prisoners these other units will capture in comparison to the Shamans.

Does anyone have a troop testing script that unlocks Knights of the Unicorn?

thermocline said:

Does anyone have a troop testing script that unlocks Knights of the Unicorn?

Click to expand...

Cheat menu- testers cheat- get 100 troops, you can find them there.

Hmm, I didn't see them but I'll look again. How do they perform, roughly?

thermocline said:

Hmm, I didn't see them but I'll look again. How do they perform, roughly?

Click to expand...

Weaker than any KO knights for sure.

Why is that? At first glance their stats look decent--PS7, WPF 350. Trash weapons I guess?

thermocline said:

Why is that? At first glance their stats look decent--PS7, WPF 350. Trash weapons I guess?

Click to expand...

Yes, trash weapons, besides there is no KO knights have worse stats than that, the only one is KoRC, like at their weapons and armor