Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[martinuzz]

Thanks for doing !!SCIENCE!! on this.

I think we might have forgotten one thing. Wrestling and shields. There's a little voice nagging in the back of my head saying 'wrestlers and shields are bad combo, less free grip to wrestle with'.
Remember reading that somewhere, but can't find it.
Wrestlers are probably more effective offensively without shields than with shields. More vulerable though, since they don't have a weapon to parry with, they'll rely solely on dodge (and armor).

[daggaz]

There is indeed a thread somewhere that talks about wrestling with or without a shield (might even be this thread lol), and how dwarves might take advantage or disadvantage of the tactics available in adventure mode, if they do at all.  Nothing was definitive.

I think Ulfarr, you also missing a basic control, which is to do dwarves with no shields, no weapons, and also no wrestling.   I noticed in my own trials that wrestling quickly became a dominate stat regardless in these scenarios, as the dwarves had little else to do, but my trial was more limited than your own.  In this instance, if wrestling is quickly auto-learned, it may be beneficial to seed embark dwarves with other military skills instead, like dodging, or armor.  Armor-user + wrestling is also a variation you are missing, I am also curious about limited armor, why not give them gladiator helms and nothing else and see how they do?

[Ulfarr]

Thanks for doing !!SCIENCE!! on this.

I think we might have forgotten one thing. Wrestling and shields. There's a little voice nagging in the back of my head saying 'wrestlers and shields are bad combo, less free grip to wrestle with'.

That would be the subject for my fort's youngest labrats citizens. When they come of age.

I think Ulfarr, you also missing a basic control, which is to do dwarves with no shields, no weapons, and also no wrestling.

As far as I know you can't forbid wrestling, and by simply assigning no weapon or shield you'll definitely get wrestlers. The only way I can think that MIGHT result in no wrestling is to create one dwarf squads so they don't spar, assuming they can still perform individual drills.

No no, fair enough, you answered my poorly-defined question suffieciently, thank you.  I asked because I want to train my smiths for stat-gain, and was curious how much in-game time it would take out of their smithing regimine.

If your goal is to minimize the time spent moving materials back and forth then I'd say you are better off with (quantum) stockpiles near your forges. Even the most optimistic projection of a 1.5-2 year training plan would still require your smiths to halt all production and focus only on training.

[daggaz]

The point of the "no-wrestling" control was to see if there was a significant advantage or not to starting with ranks in wrestling vs ranks in another, harder-to-train skill, like armor-user.  Since, as we know, they will definitely learn to use wrestling either way, and skills like armor-user can be effectively shut off by denying the related equipment. 

While high strength would minimize movement times, stockpile optimization is more effective.  I was thinking more along the lines of increasing all the relevant physical skills which affect actual production time, thus increasing the speed at which they reach legendary status.  If significant returns on attributes could be achieved in a short enough time-span, then an initial training regime would be indicated. 

[Ulfarr]

Embarking with some weapon/wrestling skills should result for your dwarves to start sparring sooner than if they didn't have any combat skills at all. Would that outweigh the time advantage of starting with, say, 5 armor user? I don't know.

About the production times, in the DFwiki attribute page, it is suggested that attributes do affect the completion time of some labors, but (unless I missed it) doesn't say anything about how exactly they relate to one another. It could be tested though.

[daggaz]

What I meant was, you obviously want to train wrestling, but it also will begin auto-training on its own quite quickly.  So you could give a dwarf armor-user (and no armor) or weapon-user (and no weapon) etc.. and let them train up wrestling naturally.  Now that will obviously take a bit longer than if you give them ranks in wrestling to start with, but how much time do you save on the next step, ie training armor which can, from my reading of these threads, be quite difficult to skill up in especially if the dwarves learn defensive skills first.  Shield skill improvement can be retarded by equipment denial, but dodging happens no matter what, so you have that issue. 

Might be best, if "natural" wrestling is nearly as good as "seeded" wrestling, to give them these other skill ranks and switch over once they are up to speed. 

Another factor that plays into this is the difference between sparring and lessons.  Dwarves will spar if they are close to the same level, and do lessons if the difference is large.  Lessons are more effective IF you have teacher/student ranks to multiply in.  Sparring on the other hand trains more skills simultaneously, and from what I understand, is what you need to do to get attribute gains. 

So.... you can give dwarves different ranks in harder-to-train skills and if those skills are equipment related, you can turn them off completely.  In the mean-time, dwarves will then train spontaneous military skills, including wrestling, striker, fighter, dodger and biter, and as they are all starting at zero (assuming no ranks given here), they will tend to spar and level it all up while gaining attributes.   Once they are well rounded in hand-to-hand combat, you then turn on the other skills by allowing progressively more equipment.  At this point, if you gave different skills to different dwarves, they will focus on lessons and more rapidly bring the other dwarves up to speed on their individual skill-focus.  Doubly so if you spent a few ranks on teacher/student and even organizer.  Once they are all starting to level out again with eachother, they begin to spar again and now you have some seriously well-rounded military dwarves.

If you are careful, you could even figure out a way to tier the system so that you start with boot camp spar-training, have some master-level skill teachers that focus a military class on one specific skill through advanced lessons, and the classes would graduate upwards until they come together into a large sparring group, at which point they would be seperated again into weapon-type for final training. 

That's what I have been working on at least, looking at Teacher 4 Student 1 Organizer 1 Skill_X 4 as an embark seed for my master teachers.

[Staalo]

Convincing science in this thread. Time to rethink militia training somewhat, although it might not provide appreciable benefit. In my forts it seems that when I give dwarves some weapons and armor and get them sparring in the beginning, they'll level in just few years up to Legendary in all combat skills, including Biting. How they do that, I don't know; they don't even use bite that often while sparring.

[Ulfarr]

ok, now I get what you are saying daggaz

I would definitely give them hard-to-train skills instead of weapons/wrestling. In terms of time saving, the pre embark bonus, is very big for skills like armor user/dodger, and HUGE for other (arguably less useful) skills like bitter/kicker. Following is the time (minimum-maximum, in months of constant training) needed for a squad to reach an average level of  4 and 5 from having no skill at all.

Skill       lvl 4     lvl 5
Armor     17-23    20-28
Dodge     16-19    18-24
Bitter      35-46    42-50
Kicker     27-36    35-42
Weapon   8-12     12-14
Wrestler  14-19    15-23

The idea of using veterans, to teach others, is very interesting. In all my fortresses so far, I was always reluctant to use them that way, mostly because I tend to rely on them for fortress defence (damn you weapon-breaking-in-trap bug ) and I wanted to keep them ready for action at any time.  How effective are demonstration as a form of training? Will the students' level reach that of their teacher through demonstrations or will they always need to spar after a point? How much do teacher/student skills affect the results of demonstrations?

[daggaz]

OMG I fucking love you for posting that data man.

Search the forums, there are quite a few interesting threads on the topic.  Ive basically paraphrased everything I have read in the post above, however, but I havent been searching the topic for more than a week (I do read fast).

Toadyone is on the record saying that teaching and student directly modify the efficiency (xp gained) of lessons.  I posted some of my preliminary results on the matter in my thread on this forum section about optimized embarks, check the last few pages..  I didnt run it very far but one thing I noticed was that student and organizer was learned slowly from zero, (with the exception that my primary teacher, the only one with a teachable embark skill in that setup, got a couple organizer ranks).  But if I gave them each a rank each of student and organizer, then both skills progressed in short time to competent which was significantly more than the earlier trial... i did not check for behavioral problems but I tend to pick my military dwarves to avoid those anyhow so I doubt that is a major issue. 

In this case, I wanted the teachers to teach each other all their skills, but the more I think about the setup the more I wonder if that is necessarily a good thing.  I am pretty sure that teaching is dominated by having a large skill gap and by the highest skill available, so better to have a strong focus perhaps than round these guys out.  They are never meant to see battle after all (which allows you to recruit female dwarves who lust for families, as well as not worry about traits like slow healing).   So you could forgo the student rank, perhaps (they will learn it anyhow eventually) and you could also not have them train with each other period, but wait for new students to arrive and slowly build them up that way, keeping their abilities seperate.   Organizer you would want to keep, it cuts down the time wasted setting up the teaching lessons and I imagine the gains become significant with time due to compounding problems that can be avoided by small reductions in time taken to start the lesson.

Ideally I would want one teacher who was extremely good at armor (first class), one teacher who was extremely good at shields (second class), and one teacher who was extremely good at dodging (third class).   Doging is a funny one because all students (and teachers) will learn it spontaneously, but as long as they suck at it compared to the last teacher, then it should be ok.  Its important because as we all know, it is a major source of defense. 

For second generation teachers, I would pick the most advanced teacher/organizer student in a given squad and have them teach future squads in their given weapon of choice. 

In the end, it might not be possible or even optimal to try to seperate their skill sets in this manner, but here you could at least control lessons thru equipment restriction.  The problem being dodge, of course (which will make shield blocking and armour user lessons less efficient). 

Speaking of armor-user, one poster here uses a cool pillar scheme for his barracks, so dwarves (MUST use full armor!!!) fall down 2 z levels (fighting level - free space- landing level) and learn armor the hard way.  Results in miner bumps and bruises.  He also put water down in the path back to the barracks to force swimming lessons.   All in all it gave a slight slow down on sparring efficiency but boosted armor-user skill which was the big problem in the first place. 

[Timmi]

I just find out in this new save that starting training with wrestling and armor using is not very faster than  those ordinary training troops (with weapon and shield at the first place). When I get a lot of level7-9 wrestlers, i also got those ordinary soldiers with wrestle skill level 5-7. Samt thing to the armor using.

[Ulfarr]

Give them a couple more years to train 

On the other hand, if you just want to get some good-enough soldiers in a somewhat short time, then weapon+shield is perfectly fine.

[Colonel Sanders Lite]

Disclaimer:

This experiment does NOT test the combat effectiveness of each build, only their training capabilities. For example, just because using a shield stuns the progress of the Armor User and Dodger skills,it doesn't mean that you shouldn't give shield to your dwarves when they do have to fight. In fact, i would argue that the very fact that it does that, is a testament to the effectiveness of shield user as a defensive skill.

FWIW, I suspect that this is a very overlooked statement in this thread.  It may be a personal philosophy thing, but my goal is *always* to have soldiers capable of being combat effective with their actual combat gear as quickly as possible.  Doing the weapon only method and then training shield later may result in a higher overall level of training over 4 a period of years, but that doesn't much matter if your guys can't survive fighting off an attack in the middle of the second year.

That being said, you're making me very tempted to try out training starting with weapon + shield, and then switching to weapon only once I feel their shield use is reasonably competent.  Their shields can be stored in the barracks which will let them grab their shields very quickly in the event that there's a military emergency.

The idea of using veterans, to teach others, is very interesting.

I haven't measured the results in any kind of science, but I can say that I *always* do this these days and, subjectively, it seems to help a lot.  I pick the best teacher for each weapon and give them the special job title "Drill Instructor" and assign them to lead training squads when I'm training a batch of fresh recruits.  Once I feel that the recruits have reached a certain level of combat readiness, I mix them into the squads with the veterans so that they will continue to learn from them as their military careers progress.

I'll tell you what though.  I have some raw recruits I need to train up soon.  I can post some comparative numbers.

[Splint]

I would just like to say (I haven't read the whole thread in detail, but I did skim it) that when units were made able to skid across the ground and take damage, accidents with wrestling specifically could occur at least in 34.11 if one unit threw the other and a body part bounced off the floor without at least some kind of leather armor - I know this was a fact because I had it occur on multiple occasions to a sergeant.

He got thrown on three separate occasions while sparring and on each occasion something would skid, resulting once in a broken wrist, once in broken ribs, and once in a broken ankle. After this last one, he did the same tot he offending recruit who kept using wrestling and his unhelmeted head skidded on the floor and his brain was damaged, killing him. I stopped having these incidents occur once I added some leather armor to the squad's uniform, with the armor deflecting the impact.

No incidents happened to the squad with their weapons, only when doing wrestling throws and they lacked any sort of protective gear.

Now, that being said, I feel like I need to start a fort and just train up a bunch of unarmored wrestlers to see what happens and if anyone ends up seriously injured or killed durring sparring sessions

[daggaz]

I'm on 44.05 and havent seen a single injury yet in any of my trials, all of which can involve wrestling and none of which have armor equipped. 

[Ulfarr]

It appears I've made a mistake with my interpretation of Dwarf Therapist's attribute scale. As feelotraveller explained to me, it ranks attributes in relation with the rest of your population. That means that a dwarf with a stable score throughout the years hasn't necessarily reached his maximum in that attribute (as I thought) but that he remains in the same rank compared to his fellows. Due to the way that I've collected my data (taking screenshots), I sadly don't have access to the ingame attribute values so my observations on attribute groth is inconclusive at best and plainly wrong at worst.

With that in mind, I will set up a new experiment in order to study attribute growth. However I have a tight schedule these days so it might take a week or two before I can give you any results.