that Spartan guy is sure insane . his city keeper is top notch

I strongly support this recent movement on more rigorous testing of game mechanics. Well done!

You also need a test with none of the stats boosted as a control variable. To see what the general gain from each is

I’ve had this suspicion for a while and I kept telling my alliance but it’s nice to see some data behind it!

I say it again lockdown chisgule content is on fire. 10000000000000000000000/100

One factor that wasn't compared was how long it took to defeat the opposing army. Health took 2:57. Defense took 3:03. Attack took 3:02. The attack buff in AoO makes these numbers much closer than what they would be normally in KvK. These results mainly apply to a trade where you are just looking for value. Pushing a flag if you are on a time limit to defeat it before reinforcements arrive would still favor attack if you are able to keep the rally full.

I’m really digging these recent videos chisgule, your content is not only entertaining but super valuable for anyone trying to become better and more successful in ROK. You put a lot of work into it and it shows! This community is lucky to have you man, so thank you!!

Great video! You have to take into account, which special troop type you have and which type of troop you are using: If you for instance are using Rome the base defense on infantry is increased by 10 (on T4 troop). increasing defense here will therefore have a slightly bigger boost than the other stat-types (10% bonus defense would mean 1 extra defense compared to not having special-troop stat) This is something to think about through the different type of units used aswell, since the different troop-types have different basestats (look at your base attack on cavalry, its lower than defense and health) Btw. i dont think you mentioned this, but the type of fight is fairly important aswell. Fighting 1 vs many (being swarmed), defense will decrease damage you take from all the enemies. Fighting 1 vs 1 is on the other hand, benefitted by using high health.

What you mentioned in this video is something I've understood myself a while ago from self testing, I came up with very similar results. This is also why Constantine is such an absolute monster for giving 40% hp to Infantry. I'll delve a bit deeper on the math side of this below. Looking at the data that you've collected in AOO, it's very compelling evidence that HP is by far the best, the sev wound ratios are what is important in open field, and 1.60-1.52-1.35 says exactly that. Looking at the unskewed stats of defense and health, gaining 27% health over 27% defense had a (1.6-1.52)/1.52 = 5.3% impact on the battle, adding on the base stats, means in a situation where it was 183%hp 197%def vs 157%hp and 224%def, the former comes out 5.3% ahead. Using a bit more calculation, you find that 27% hp stat is 14.7%(27%/183%) of the entire 183% hp stat, but it positively impacts the battle by 5.3%. which also means that in this specific comparison, ratio-wise, having HP was 36%(5.3%/14.7%) more effective than having defense stat.

Your recent videos with insights in testing results has caught my attention and had me listen to you more. We do tests too so I understand how much time (and brain cells) these took, not to mention the time spent editing the video. Is it worth it? I would say YES! I hope more players appreciate these kind of intelligence. Great work Chisgule!

The lowest stat arguement might be right as for now. However the question to ask is more, where is the breakpoint for efficiency? Extreme example would be, is a unit with 100 HP and 500 defense better than a unit with 500 HP and 100 Defense provided they have the same attack value? If you figure out the damage calculations or have them by some other means, you should be able to find the cut off point where with X amount of Z stat you want more in stat Y and then when certain point is hit for that you want the last stat instead of previous two. Another game I play taught me that quite early, which is Black Desert Online. Damage reduction vs evasion vs straight up health. There will be breakpoints where overstacking is way worse in terms of value, resulting in diminishing returns. Finding it is gonna be tough, however whenever scenarios or buffs gives something like 200% damage, then yes your return value goes down for going attack compared to being able to take more hits with less loses from health and defense. Way to test this more would be 2 beginner accounts with 0 research and no alliance. Have T1 troops beat the shit out of one another and go get some runes with defense/damage/HP. Do note that you won't be able to use gear here, but it would only cost time and you could get more insight into the calculations by switching commanders. Granted new accounts would not have all commanders to test, but take Tomoe Gozen as an example. Her first skill gives 30% damage to all units for 3 seconds at max. You wouldn't even have to max it and just leave it at 10% (and she's fairly easy to get), with that you could figure out which unit type benefits the most from the attack stat at base value of stats. Could be interesting to see how different that is when you have everything maxed in both tech and alliance research. Of course no talents on her. For seeing how the health increases at least at higher end, you could get Ishida Mitsunari (universal health buff on 4th skill) as a secondary commander (HOWEVER DON'T EXPERTISE, adds randomness). However he would boost rage gain and both players would require same level on main skill and 4th skill to keep it on the same level to be able to replicate same/similar reports. If you really wanna get into small number changes and differences, looking at some of the talents in the talent tree you could add a few points of % HP/Defense/Attack and not get anything that procs a chance for something. However this requires gems to reset and respec to manouver around. Sorry for the long read, hope this helps in in some way :)

Quality content and information on this one, Chisgule! The effort shows through. I look forward to this level of content in the future from sponsored content providers ;)

I've always just gone in Expedition and compared the first white damage number to compare health vs defence stats. Never seen any randomness on turn 1 damage. This method produces the same conclusion that health is superior point for point.

Chisgule you are revealing Lilith secrets won't they be mad🤣🤣🤣🤣

Daya to support my wondering. I am a 30 mil power city. I was focused on def as a primary stat. It seemed I could never get enough. So I started focusing more on health...suddenly my troops are beasts! In osiris I noticed a big change in my points I was getting. I was pretty convinced health is super super important. Now I'm sure of it. This video is awesome.

When I tested this I did it in expedition. To remove randomness I tested using a few commanders who had no randomness in their skills, (charles for non-dps, sun tsu for dps, and Richard for healer,) just to see if different stats were preferable for different commander types. (had heard healers would prefer health.) Also, I tested against a tank and a dps by testing on number 25 against charles, and on number 50 against ysg/sun (there are no archer troops in 50 so ysg/sun has minimal variability in this instance.) Also, I brought a reduced number of troops to stretch out the fights, and tracked results by screen recording and comparing the remaining troops as well as the length of engagement. Probably also not a perfect method, but it was before AoO practice matches were available.

I'd like to see the test again if you divided up the health and defense and compare that with all health all defense and all attack. Thanks for all you do.

I'm a university professor, and I really appreciate both the approach and the presentation of your methods and findings.

TLDR: 1. Boosting your lowest stat is what you'll benefit from the most. 2. In his trials, HP came out on top

Hey Chis, thinking about a more well-rounded approach. We've ruled out variability of tiered units, and repeated test results. You could run one of each of these tests with 50k t4 units and get a reliable comparison

Chisgule you have upped your content creator game big time recently. I think this analysis qualifies as actual science. Great job!

Yes please please keep making these kind of in depth videos, love it 🤟🤟🤟 id really like to see u test this in regular game play. For 1v1s/ open feild war.

Everyone knows Load is the best stat in the game smh...

Please test Health and Defense stats when swarmed. Because when youre in a war you wont expect to be one on one with another player. Anyways, this video tells us Health is the best when 1vs1.

These are by far my favorite types of videos that you make! One thing that I was wondering is if these tests remain consistent for different units, since they have different flat values. I noticed that only infantry was used in these tests.

nice video , simple, because the formular use mulitple each other ,the lowest statistics bring most effectiveness .

Hello, I've enjoyed this kind of video, where testing is highlighted. And I do agree with Chingsule that health it is indeed a great stat to boost and I'll explain how I think and did some testing as well. However, there are some considerations regarding troops stats that I'd like to make as well as some suggestions for future tests. I'd say it all depends on the user focus. To summarize this, it is notable that attack does increase some skill (and overall) damage, defense reduces damage taken and health grants more survavibility. So, if you give a defensive commander a boost to attack, it won't be much effective as other stats would be. If you give defense for damage dealer commander might even help him mitigating the damage taken. However, the health it works fine for every kind of troops. Defensive commanders will get more survavibility and damage dealers will do more damage since they will have more troops over time dealing damage. That being said, the question everybody might be thinking: "Is overbuffing one troop stat (such as attack) bad?". And honestly, I don't think that's a bad idea. If you want to demolish a target as quickly possible and you got the alliance backup and time is relevant, attack might be the thing. And yes, you'll have more loss, but maybe you've accomplished your goal faster. My suggestion for future tests are: Test the different commanders roles with the attack, defense and health. And do include in statistics "time", which might be relevant in different situations. Thanks for everyone that has read my comment, Gustavo Bessa.

Great video! Like i wrote in the equipment video, I always go for health equipment!... It would be really cool to see the difference between commander with vs without equipment.

Thank you for making this video, much appreciated, I've started focusing on pushing health for commanders that melt easy, like khan.. results have been much better and i can use them on field again, it may not be the best result, but at least i have more commander options now.

You have tested with infantry only it might change if you used other types of troops, maybe the best for archer will be attack and for cavs defense

Great vid chis. We also tested on garrison against atila and having More health worked better by far. If you want to change equipment and talent tree on ark you need to force quit the game and then change it

This is some nice information. This is going to help me a lot. Thanks Chisgule

Another quality effort. Love to see the data behind the battles. Thanks for the effort and time you are putting into these

thanks chisgule....appreciate you taking the time and effort to test this...enjoying this technical videos alot..keep it up👍

Hey chisgule, let me say good job there. And i hope what i am gonna say will be helpful. First, if you have played many other games like i did. It is a baseline thing in literally every single game that boosing your lowest stats is the best to do. However, in certain situations and certain jobs it could be ghe wrong thing to say. And please hold this thought. if you are askibg what other test can we run then the answe is as follows, you need to run a test where you attack one march with 2 marchs, where you will control the stats of the march that is not tanking and more tests where you also play with the stats of the one solo march. From your video and from what i the simulations i ran in my brain, this is what i conclude. First. Attacs stats value will be greatly enhanced by the second march that ia not tanking and lets call it march X . If The enemy of X which is focused on reflecting damage rather tuan pure survival. then the attack stat for the damage march that is not tanking will be greatly reduced, although j would still assume it is going to stay the best. This concludea that im actual battles, the fastest march speed should be combined with the highest attack. Because you will be able to run when you are focused and attack when you are not. Second In conclusion, the higher the enemy attack is the better your defence becomes, and the higher your defence is the better health will become. But considering how increasing defense had more effect on the number of severes for both marchs. This means defence stats might be considering superior to the tanks in the battle field, especially healers like richard. Third, taking in consideration what was previously mentioned. This also concludes that the best stat for support players should be health. 4th, the value of attack stat should increase1 vs 1 scenario for the march that generates more rage. This does not conclude that attack will better than def or health. I am only saying the march that will hit a skill first will benifit from attack more. Lastly and a little unrelated. I don't know how healing works in this game, lets say you have a temporary buff of 30% health. Does richard heal heal less troops because of the increased health? Or does the nukber of troops remain the same ? If by any chance the increase health is nad for richards healing it will put defence for healers into a whole new different level.

great video chisgule your videos has improve my gaming experience alot and i am very greatful for your efforts.

I am beginning to wonder if this is all a game of diminishing returns. Health is by far the lowest of stats by default & considering each stat loses its return value at a similar rate, HP should easily be the best. Thanks for these epic tests Chis! My God this changes everything....I never even thought about diminishing returns in RoK despite being very familiar with the concept!!

7:30 in the morning and doing math so this will go great! The formula isn't known but from what we know it's the ratio between attack and defense and health is raw subtraction so... if you have 100% attack and 50% defense you have a ratio of 2. If however you gain 10% more attack (110% attack and 50% defense) your ratio becomes 2.2. If you have 10% more defense (100% attack and 60% defense) your ratio becomes 1.67 (rounded). So in this case defense would give you a better ratio, .33 change over a .2 change. Health though is just raw, so if you increase your health by 10% then you will have 10% more health, there is no special math that goes in (none to my knowledge at least).

Appreciate your video contents lately which takes real game testing into play. Must have taken lots of time and effort. Real good content to proof the game mechanics!

Good video. Not only does alliance technology not give health, but your own tech gives much more attack and defense than health (troop specific stats are only attack and defense).

Super interesting ! Your conclusion is more of an hypothesis than an actual conclusion, since it couldn't be observed that the lowest stats is actually performing better. It would be very interesting (and expensive :p) to run these same test in open field ! I think the only thing you can conclude at this time, is that health is performing better than defense fora normal attacker. I would suggest running same test with t4 in open field (no need to do t5 since variability seems to no be affected, which makes sense for % boost. Second test that I think could be insightful would be to try a nuker. If you have a nuker with only 1st skill (maybe charlemagne ?) it would be very interesting to see how attack actually impacts it. I'm sure attack has much more value for a nuker than a normal attacker. Thanks a lot for these test, we would not be able to see all that without you

are you in our group chat or something? like this is the 4th or 5th time that my friends discuss something and then either the same day or a day after you upload a video about it 😂 keep up the good content

These tests are very insightful, thanks Chisgule. I would be nice to consider the effects of skills damage since is a very significant combat mechanic in ROK, and how that works when boosting different stats

Just to let you know I picked the Spanish Cavalry and also Forge every weapon towards attack so I did everything the opposite on the advice you just gave us, i started playing the game before I caught onto the videos on KZfaq....lol... and I also love some advice on migration if you could list a couple of Kingdoms down on the on the reply section would be awesome love your videos thanks

keep on doing it chisgule, technical analysis is really required on this game 😁

This video was very helpful. Thank you chisgule.. I was trying to figure out what equipment to forge and one had to do with health and the other I believe was def/att..

Since there is no randomness there is no difference if you use 150k or 10k (the rateo should Be the same). You could test in without ark and have very little res cost (you can even use t4 since there wasn’t that big a difference). I might be wrong

Another test you could perform is taking mid-tier commanders with high health versus meta(-ish?) commanders with either high def/att and (then high health!)...otherwise FASCINATING stuff Sir! Keep it up!

Chis what you all said seems to be true but only for Infantry...from what i learned by playing and testing this game type of stats matters the most by the type of troops you use: Infantry: Defense>health>attack(same as ur result) Cavalry: Health>attack>defense Archers: Attack>health>Defense So yeah health is better because it is never the least needed stat for any troop type including seige. Try running same test with archers you will be astonished to find attack will be best even with all the ark attack buff you get.

I've been loving these recent videos, definitely worth the effort my dude 🤓❤️

wow chisgule your videos are on an absolute next level now bro, thanks for your efforts and I hope I can be with you in the same alliance one day

Awesome Chis! This will really help for all players. Good job! 👍👍

Great video! Very cool to see these testings. Maybe test in reagular kingdom using less troops in each march... like 50k troops each

I figured it varied based on troop type as well. What troops did you use for this test?

Amazing video. Congratulations dude !

Thanks❤ i was thinking about it and found out at last..

Hi Chisgule So the obvious question that arises here and which is possible to test in the aoo is to repeat the test but with both of you having different buffs. Specifically hp vs def. Thanks for your videos, love the search option on the slider.

I think a better word for it is diminished returns where the more you have of a stat the less it gives you the less you have it won’t effect it to a certain point

pioneering in-game science! that's my chisgule!

Nice bro, love these types of videos!

Hi Chis, Great stuff. Boosting the lowest stat seems to be a strong argument and Health does seem to be the priority. But I am also a little surprised between the results of Attack & Defense. I am wondering did Defense do better because it was Infantry that you were testing? I do expect Attack to be better than defense for Cavalry & Archers is you can test it somehow.

Maybe you could test with T1 units in normal kingdom. Less resources and no speedups needed

thanks fo these test. I notice that your tests are done only with infantry and maybe your conclusions would be different if it was with cavs or archers

Great stuff again! Really good to know, while I am beginning to craft my gear 😊

Mind blown!! Great content chis! Keep it up

Not sure if more testing is necessary, but if you were to do anyting else I would say same test on the regular map, no alliance. Hospital bill might be hefty but I think you can afford it ;). Good stuff as usual. Keep it up.

Amazing video, I really like this kind of research you do. I think you should also have tried with different units (ex a unit with more base defense than base attack, etc). Also I don't understand why you say it is better to boost your lowest stat. The more stat you have the bigger should be the bonus right? If your unit has 200 base attack and 190 base defense, 10% attack is +20attack and 10% defense is only 19 (makes sense).

Chis, these kinda videos are soooo freaking helpful! Thanks!

Don't forget that the troop's stats do affect the test, I mean Rome special units have more def than hp (main stats), so taking def % will stack more than hp, but the throwing axemen ( France special units ) have higher hp stats than def, so stacking up hp % will be better :) keep that in mind

I got a theory that def out scales health but theres a graph where health take over the more troops. Not sure if its just in my head but. So rallies/garrison tests would be good. The testing is a huuuuge project! And heard something about diminising returns when you pass 100% meaning 101% is not really a 101% boost but maybe 100,5%

Super helpful as I am newish to the game! Subbed!

Just had to comment to show appreciation to the myth busting content, keep it up!

The tests you're doing those days are super super nice content ! :)

the hypothesis now accepted is that increasing the lowest status has the most influence. this can be removed by putting everything on the same level. this is difficult to fine-tune, but this would allow you to measure an increase of, for example, 10% in a status, without assuming this hypothesis. once this was in a non-ark environment, it could be fully generalized in 1vs1 situations

Please do more test. This is what I’ve been searching for because it was never really talked about by any youtuber besides the generic more stats more power. Example Cav what stats is best the march and if it’s skill commander vs troop commander

was waiting for this vid hahaha we got some misbehaving governors that need zeroing haha

a very good quality and informative vid. thanks chis.

First thanks for the call out greatly appreciate it. Question for you: So with your recommendation of using defense in AOO over attack I can totally understand, but what about standard fights Open Field/ Rallies / Structures. I would assume you would still recommend boosting your weaker stat. But wouldnt it be different depending on how your using your troops

This video with the video that Dragothien gaming did a while ago both agree that Health is probably the most important of the 3 stats. Thank you

For Osiris I would make the argument that the most important stat to look at is the length of the battle and not sev wounded

Farm accounts on Sunset canyon Stick to City Defender so you and your friends will be at the bottom to fight each other. That way you can hit the revenge button to retest. Plus you don't have to watch the fight you can just check the data.

Love the star wars T-shirt Chisgule

I watched a 4min Lord's Mobile video for you for this. Great vid, keep up the mythbusting.

First off. Thank you chishule! You ROK. Just watched what is best stat and it got me thinking about a theory of mine. You say to increase whatever is your lowest stat. I am wondering if this also goes for troop type. Ex: you have a strong cav commander should you match with higher tier cav or offset the balance by going with higher or more archer or infantry?

I think you're close to the answer, but I might have a reason for what you're observing. Every point of health is enhanced by every point of defense. Every point of defense is more valuable if there's more health for it to apply to. The same is not true of attack, which is a standalone stat in this game. That's the reason that an individual attack point is less valuable. Then for the reason why you see better results for health instead of defense, your reasoning about it being better to increase the lower stat seems correct. Suppose we have X% damage reduction applying to Y health. Then the effective damage applied is BaseDamage*(100-X)/100. To reduce Y health to 0, the attacker needs to apply Y*100/(100-X) base damage (this comes from rearranging the other equation). We want to maximise this, so as both in-game stats are acting as a % enhancement of the base health in this equation, it makes sense that you want defense and health to end up balanced. (This isn't exact maths of course, I'm just providing a rough explanation). And then like you said, as we already get a whole bunch of defense boosts from other places, it's usually going to be that health boosts are what we want. However, for someone like Edward it might be worth checking if you at some point "overtake" your defense with health, and it's instead beneficial to equip a mixture of health and defense items. Sorry this got a bit involved, I have no idea if it's actually true of course, but it attempts to provide some reasoning for your observations.

you need to test the buffs on other troops. The disadvantage of Cavs is exactly that they hav low health, maybe this had influenced on the result. Try to use this method using also INF and ARCS and see if works the same.

Thank you for all the hard work you do for us to to learn , much love ❤️

Another great video! (Yes, I'd like more). Open questions; when you retest a few of the examples outside of AoO, are the results significantly different? Does this mean France needs to be upgraded as a civilization choice? Will you revisit talent builds to take advantage of Medicinal Supplies, Elixor, and Strong of Body?

Drago revealed it month back but still great effort by you

I've proven that, health is the best statistic. & I've dismantle all the attacking equipment, & I lost 50% of the materials. Lol..... Thank you for ur content. Terimakasih. (From indonesia)

Appreciate your hard work chisgule, thanks I learned alot from your videos

Excellent video like always, but your level 20 commander vs his level 45 which im sure has an additional 15 points or so that can be distributed between attack, health, defence or other talents makes his level 45 just much stronger i just wish they were both level 20 or level 45.

Great video very helpful! But you only tested with infantry, could cav and archers have a different reaction to each stat?

The main reason that health is the stat to stack is simply this, there are less health buffs in the game. Research buffs alone boost att and def by 70% but health only 40%. Couple this with a there being no health boost item and you’ve got a 35 percentage point Gap to make up for. Easy math will tell you that until that gap is bridged, adding points to health will give you a higher return on your investment than attack or defense. Adding one more research in the academy solely for health would fix this.

been non stop fighting in kvk 47 564/646 and all over the last 12 hours check it out when ya have time been a blast

have you tried testing the balanced bonus (make atck/hlth/def bonuses as close as possible) and balance stats (make pure stats numbers as equal as possible)? This could be more meaningful than minmaxing

Seeing as the reports show the difference between t4 and t5 are negligible you could res in the open field use maybe 50k marches for reduce the res cost for healing. This would then be able to correlate to t5 marches without the insane res cost.

That is probably the first time City keeper was leading t5 😂

I enjoy and appreciate your tips and guides Chisgule

In a situation where you let the march go very low or you´re doing 1v1 health is very strong. Also for surviving longer so allies can deal dmg in that time. BUT if you´re murder balling, rallying/garrison or just not getting hit a lot or can reinforce/refresh, then i think health is less key. Health helps keeping higher troop count, which is super key for good trades, but when you can get new troops from joining or refreshing i would go defense or atk. For ark when you let troops go to 0, 100% health is meta. In kvk i would still go the lower stat, but if you refresh a lot or reinforce a lot, ignore health. If you go all-in to 0, health is then best, imo. Not sure how you would test. Also atk apprently inflates skill dmg. ect. to an extend, but agian proving all this is hard. Maybe if we test with no atk, and with full atk and see skill dmg difference. Could be done in expedition. Same way test the dmg done to you with defense. As expedition is the same, the first skill atk should prove accurate.