1 melee knife lost to armored 15 longsword?? bug or astromically lucky?

[Bio[redacted]]

I got a raid earlier and somehow two bandits with no armor and a skill of only 1 with only poor knives somehow beat a flak armored charecter with a blazing passion for melee at 15 skill and a longsword. Shouldn't my pawn have easily been able to dodge all the attacks thrown at them? I checked their injuries and they were barely even bleeding, when in reality they should've been hacked to pieces. Did they just get astronomically lucky and have an Xcom moment, or is something seriously bugged here?

[whatamidoing]

I made a little program for some numbers. I used the melee dodge and hit chances from the wiki. I assumed 5 hits to kill and counted ties (both hitting 0 hp at the same time, no turns here to avoid a bias towards the first-striker) as a win for both sides, then let it run until the level 20 won 1000K times (~1 second). At that point the level 10 had won ~592K times. Of course, that's just a 1v1 with evenly matched "gear."

[glass zebra]

Yeah the melee dodge chance adds a ton. 10% vs 30% is almost a 30% increased survival rate.

[whatamidoing]

Publicado originalmente por glass zebra:

Yeah the melee dodge chance adds a ton. 10% vs 30% is almost a 30% increased survival rate.

Actually, I'm stupid and didn't put parenthesis where I needed them and that was resulting in dodges being ignored. With dodges, it's ~242K:1000K. Also, about 1/2 of those level 10 wins are ties, putting them in their own category now so I can experience the satisfaction of watching the ties number decrease when hp increases.

[MtvYoloSwag]

I think combat is just weird sometimes.
Had my pawn with bionic arms melee attack a prisoner doing a prison break.
Her arm broke off.

[ichifish]

Publicado originalmente por Astasia:

Publicado originalmente por ichifish:

Dodge chance is determined by move speed, so if your pawn had flak pants and/or jacket they would be less likely to dodge.

I thought this was the case at one point and probably mentioned it a few times years ago, and it may have been changed later, but melee dodge is currently based on "Moving" not movement speed, this is the pawn attribute like manipulation that is mostly controlled by health and bionics. Wearing armor that reduces movement speed does not reduce dodge chance. Though your next point about an enemy hitting like a colonist's leg does certainly reduce dodge.

Dodge suffers from massive diminishing returns and has a hard cap of 50%, it requires basically every source of moving, bonus dodge, and melee skill to hit that 50% cap. Go-juice, luci, the nimble trait, fibrous mechanites, archotech eyes and legs, etc. A normal pawn is probably only going to be at around 35-40% dodge with maxed melee skill and bionics. Dodge is not a major factor in combat, it was a fairly late addition to the game and I'm still not entirely sure where it is supposed to fit into the balance.

Thanks for clearing that up. That makes a significant difference in my playstyle as I was avoiding Marine and Cataphract armor because I thought moving and movement were the same thing. I still like a lot of mobility on my melee pawns, but now I don’t have to worry about them putting on the heavy stuff when they’re door blocking.

Maybe I don’t get the math, but a dodge of 30% seems pretty substantial if you’re not using a standard line-up-and-shoot killbox. Considering that blockers often have to withstand 20+ hits even a 10% increase could make a big difference. Sure, there are more efficient setups than melee, but if your scenario or playstyle prevents them, getting as close to 40-50% seems like a good move.

One thing I’ve been curious about is how the to hit chance for non-humans is calculated. Animals seem to hit quite frequently, regardless of their type.

[glass zebra]

Animal hit chance is as if the have a melee skill of 4 (it's in their combat info formula). 30% dodge chance is a lot, but that is still only really important if you are able to survive a few hits, since the chance is too low to reliably save someone from a 1 hit kill or similar. For real tanks that can take a few hits (like the 5 in whatamIdoing's example), 30% dodge is extremely strong.

[ichifish]

Publicado originalmente por glass zebra:

Animal hit chance is as if the have a melee skill of 6.

Thanks! Where did you find that?

Edit: the number seems to have changed from 6 to 4 since I quoted you.

Edit edit: and thanks for the location.

[whatamidoing]

Publicado originalmente por glass zebra:

For real tanks that can take a few hits (like the 5 in whatamIdoing's example), 30% dodge is extremely strong.

Yes, with 1-hit kills the level 10 manages to win ~690K (nice) times, but about half of the total matches (560k/1130k) are ties and therefore would go to whoever attacked first.

[ichifish]

Publicado originalmente por whatamidoing:

Publicado originalmente por glass zebra:

For real tanks that can take a few hits (like the 5 in whatamIdoing's example), 30% dodge is extremely strong.

Yes, with 1-hit kills the level 10 manages to win ~690K (nice) times, but about half of the total matches (560k/1130k) are ties and therefore would go to whoever attacked first.

Could you simulate health (manipulation/moving) by reducing the hit/dodge chance for each hit taken? Maybe that’s getting too far away from the original formula, but the point being that a successful hit translates into more successful hits.

[glass zebra]

I was reading what you wrote there for a bit and was trying to find the reference to the 30% dodge. But in your 5 hit example, not having that 30% dodge would make them win about the same amount of times. So a 30% difference is pretty giant if you can take a few hits.

Publicado originalmente por ichifish:

Could you simulate health (manipulation/moving) by reducing the hit/dodge chance for each hit taken? Maybe that’s getting too far away from the original formula, but the point being that a successful hit translates into more successful hits.

Depends on how much it reduces : P Your tank will likely tank more than 1 fresh raider after one another and so just get worse and worse. Bionic arms and legs can ease that a bit since there is no extra reduction from pain.

[Astasia]

Publicado originalmente por ichifish:

Maybe I don’t get the math, but a dodge of 30% seems pretty substantial if you’re not using a standard line-up-and-shoot killbox. Considering that blockers often have to withstand 20+ hits even a 10% increase could make a big difference. Sure, there are more efficient setups than melee, but if your scenario or playstyle prevents them, getting as close to 40-50% seems like a good move.

It's a matter of perspective. On it's own it sounds decent, but when compared to everything else it's a bit trivial. A colonist in full legendary cataphract armor can take 10-20 times as many attacks as a naked colonist. A colonist in full cover can take 4-5 times as many attacks as a colonist out in the open. A colonist with the tough trait can take 2 times as many attacks as one without. A colonist at max melee skill can take 1.5 times as many melee attacks as a colonist at 0 melee skill from dodge.

It "helps" but it's absolutely not going to be the deciding factor in most combats, especially since most enemies in most raids are shooting at you.

[whatamidoing]

Publicado originalmente por ichifish:

Publicado originalmente por whatamidoing:

Yes, with 1-hit kills the level 10 manages to win ~690K (nice) times, but about half of the total matches (560k/1130k) are ties and therefore would go to whoever attacked first.

Could you simulate health (manipulation/moving) by reducing the hit/dodge chance for each hit taken? Maybe that’s getting too far away from the original formula, but the point being that a successful hit translates into more successful hits.

I could, but we both know that's going to weigh things even more extremely in favor of the level 20.
I tried doing it a lazy way (proportional to the hp) and it ends up in an infinite loop once I try 10 hits to kill and I cba to figure out why. It hits 0:1000k at 4 hits. No guarantee the numbers are accurate.

[gimmethegepgun]

Publicado originalmente por glass zebra:

Bionic arms and legs can ease that a bit since there is no extra reduction from pain.

Not only that, but the lack of bones in the bionics means that hits tend to deal less damage since they don't hit 2 body parts at once. Same for the lack of digits, and the combining of the shoulder, arm, and hand (or leg and foot) into one piece.

[ichifish]

Publicado originalmente por whatamidoing:

Publicado originalmente por ichifish:

Could you simulate health (manipulation/moving) by reducing the hit/dodge chance for each hit taken? Maybe that’s getting too far away from the original formula, but the point being that a successful hit translates into more successful hits.

I could, but we both know that's going to weigh things even more extremely in favor of the level 20.
I tried doing it a lazy way (proportional to the hp) and it ends up in an infinite loop once I try 10 hits to kill and I cba to figure out why. It hits 0:1000k at 4 hits. No guarantee the numbers are accurate.

Thanks!

[The Yeen Queen]

Publicado originalmente por VoiD:

Nobody said anything about dealing 2x damage, I said he's 2x better, as a soldier.

Every time you miss the world doesn't stop to allow you to hit again and try to get your average perfectly after 1000 attacks, the guy who misses twice as much is twice as likely to get killed in many scenarios, as every time they miss a threat isn't erased, every time they miss the enemy is also taking a chance against him.

Nobody's missing twice as much. I don't know where you're getting that idea...