Monday, March 21, 2016

Power Sword of Lightning Claw?



As I continuously tweak my army, I tend to try out different ideas that don't end up on your standard internet lists.  Today I take a look how the standard power sword stacks up against a single lightning claw.  The results are unexpected, but not for the reasons that are obvious.

**Update: The original math had some significant errors that have now been corrected**

There are many arguments that power swords are not a good option, but in a meta that is marine heavy they remain a popular and viable option.  Since this is the exact price of a single lightning claw, these two can be directly compared.

A power sword can be combined with a bolt pistol to gain an extra attack, but the single lightning claw will let you re-roll your failed wounds.  The power sword can potentially kill more models, but the claw is going to be more reliable at wounding.

I started by looking at the average number of wounds (what most people consider MathHammer), and in most cases the lightning claw caused more wounds.  The only times that it didn't were against toughness two and toughness three opponents.

However, average number of wounds isn't a reliable way of looking at this situation, so I ran things again but this time I looked at the chances of causing each distinctive number of wounds (i.e. 0 wounds, 1 wound, etc.)

*Note: All calculations are done with the weapon bearer being an Aspiring Champion*


Fighting Common Opponents:
Against a guardsmen equivalent a power sword and a lightning claw are almost exactly the same on the charge.  Each one will have about a 35% of causing more wounds than the other, and there will be about a 30% chance that they will tie in wounds.  Without the charge, the power sword will do better slightly more often.

Moving to fight a marine equivalent, these two weapons are dead even when we don't get the charge.  Once the charge is factored in, the power sword gains a huge advantage.  With the additional attack from the charge, the chances of the power sword causing more wounds jumps to 55% and the claw drops to 19%.

Fighting against a Daemon Prince equivalent isn't usually recommended for marines, but the results are similar to the previous situation.  Without the charge, the claw has a slight advantage, but they are more or less even.  With a charge the power sword comes away as the clear winner, causing more wounds 47% of the time.  The claw will only cause more wounds 18% of the time.

The significant takeaway isn't that power swords perform better, but that the better performance does not come from the average number of wounds.  Be wary of anyone ONLY using average wounds to justify something mathematically.  In almost all of these situations, lightning claws are going to cause the same or more average wounds than a power sword.  When the chances of causing a certain number of wounds are compared the power sword has a distinct advantage.

Here is a link to my corrected and updated numbers.

A big thanks to Jeremiah Morgan for finding such a massive mistake in my numbers.  I had a significant copy/paste error in my original spreadsheet, and I did not catch it myself.

I also wanted to bump what Jeremiah posted in the comment section, because his analysis is spot on and I couldn't have said things better myself.


The general rule is that swords are better against T3 opponents or when you only have 1 base attack. Otherwise, the claw wins out in nearly every other scenario.

The gap is going to widen in favor of the claw with every additional base attack that you have. So your chaos lords and various equipment or buffs that increase your attacks will ensure the claw outpaces the sword. While the inverse is true as your strength goes up, but less pronounced. I didn't run any simulations with hatred thrown in the mix.

Because of the champion of chaos rule, often you just need to cause one wound to kill the opposing sgt to get a boon. In this case, the claw will whiff far less than the sword. For example against another marine sqt on the charge a sword will whiff 32% while a claw will only whiff 25% of the time. Without the charge a sword will whiff 42% and a claw 39%. In all cases (except against T2) the claw whiffs less and gets a better boost from charging than the sword.


17 comments:

  1. What model were you using as the attacker in these scenarios? Because that definitely affects the relative performance of the weapon setups. For instance, a basic CSM only gets 1 attack so the +1 for the sword/pistol literally doubles your attacks which makes it a lot better than the lightning claw. With a sergeant the +1 is still great has a lesser effect (especially on the charge). The difference between 4 attacks with the sword/pistol and 3 with the lightning claw isn't as apparent. The more attacks a model adds (say a Khorne Chaos Lord on the charge) makes the +1 less and less significant.

    ReplyDelete
    Replies
    1. is still great, but has a lesser effect**

      Delete
    2. I'm using a basic squad leader for a unit of CSM's.

      Delete
  2. My army is tooled to fight Necrons originally, and with their last update the armor save dropping to 4+ on warriors changes the math a bit. According to math hammer, the power maul actually has an advantage over both, but the claw still seems a better option for versatility when you factor in things like Wraiths. Nice writeup!

    ReplyDelete
    Replies
    1. I feel like the power maul is a pretty underrated option for a lot of units. It easily outperforms the LC/p-sword/p-axe for any 4+ or worse armor and outperforms the LC and p-sword against 2+ armor since none of them pen it. It's a jack-of-all-trades weapon that also really shines against vehicles or any high toughness.

      Delete
    2. I absolutely agree. If your meta is anything other than 'All Power Armor', then power mauls are a great option.

      Delete
  3. I have to question these numbers and wonder if you reversed something in the simulation because these are the exact opposite of the results that would be expected.

    An aspiring champ has 2 base attacks, so a sword would have 3 attacks for dual wielding. The value of each power claw attack is more than each power sword attack in all cases because of shred. The value of the sword is the +1 attack. So there is a convergence point when the extra attack is less useful than the value from shred. The marginal utility of going from 3 to 4 attacks (33%) on the sword is less than the value of going from 2 to 3 (50%) so I don't see how the sword gets more useful on the charge. In both cases you gain 1 attack, but the claw attack is more valuable so you would expect that the claw would overtake the utility of the attack even further.

    I'm interested in seeing the distribution plots on these. Obviously a champion could never cause 4 wounds with a LC, but the chances of causing 1 or 2 wounds is much greater with a claw. Where I'm not clear is at 3 wounds. I expect the total amount of wounds caused by LC is greater than the sword in all cases, except T2 models, for an aspiring champ over any reasonable run.

    ReplyDelete
  4. Running my own simulations (100,000 runs), I get the following when I compare the wound totals side by side.

    Against GEQ, without the charge the power sword causes more wounds 37% of the time and the claw 29% with an equal number of wounds 34%. On the charge it's very even with the sword at 34.8%, claw at 35.8% and ties at 29.4%.

    Against MEQs the claw wins out. Without the charge the sword wins 30% of the time, the claw 32%, and ties 38%. On the charge the gap widens with the sword winning 30%, the claw shoots up to 38%, and ties drop to 32%.

    Against a DP, or if you were to fail a fear check against a T5 model, the sword wins 20.7% ,the claw wins 24.5, and ties are up to 55% (mostly because all attacks miss for both weapons). On the charge the claw takes a bigger lead at 31%, with the sword at 23%, and ties at 56%.

    The general rule is that swords are better against T3 opponents or when you only have 1 base attack. Otherwise, the claw wins out in nearly every other scenario.

    The gap is going to widen in favor of the claw with every additional base attack that you have. So your chaos lords and various equipment or buffs that increase your attacks will ensure the claw outpaces the sword. While the inverse is true as your strength goes up, but less pronounced. I didn't run any simulations with hatred thrown in the mix.

    However, causing more wounds in a given round isn't the only statistic we can run. When running the distribution and comparing total number of wounds caused, the lightning claw is ahead, or within 1% in every scenario except T2 enemies.

    Because of the champion of chaos rule, often you just need to cause one wound to kill the opposing sgt to get a boon. In this case, the claw will whiff far less than the sword. For example against another marine sqt on the charge a sword will whiff 32% while a claw will only whiff 25% of the time. Without the charge a sword will whiff 42% and a claw 39%. In all cases (except against T2) the claw whiffs less and gets a better boost from charging than the sword.

    ReplyDelete
    Replies
    1. Re-running my numbers I noticed errors in my calculations in the Charging against a MEQ and Charging against a DP. I am going to update my post to reflect my new numbers (which closely match your)

      Thanks for helping me finding my errors and keeping me honest.

      Delete
    2. I hope you don't mind, but I quoted your analysis in the main post because it was so close to what I wanted to say.

      Delete
    3. I guess it's not too surprising since you ran 100,000 iterations, but those numbers are almost exactly the same as the theoretical percentages. Yay for law of large numbers.

      Delete
    4. Yeah, number simulators can be a good way to check your work. I don't have one (probably should get one), but I usually check my probabilities by making sure they all add up to one. Apparently it is not always reliable.

      Delete
  5. Could you do a breakdown of various strength's chances to actually kill decurion necrons? I fought a destroyer cult last night with a tau S5 spam list and my morale actually broke playing the game at the complete inability to kill necron models. I was curious what the % are to slay models in various circumstances.

    ReplyDelete
    Replies
    1. That's Necrons for ya. One Pulse Rifle shot has only an 8.3% chance to kill a Necron Warrior once you factor in the fact it has to hit (4+), wound (3+), not get saved (4+), and then not get negated by Reanimation Protocols (4+). A 10 man squad of Fire Warriors in Rapid Fire Range only deal an expected 1.67 wounds to Necron Warriors...

      Delete
    2. I haven't played Necrons in a long time, but that is nuts. At that point you are going to need some heavier firepower than your standard pulse rifle.

      Delete
    3. Necrons are insane. They essentially get a 5+ FnP (Reanimation Protocols) that can't be negated by instant death (only suffers -1 to the roll instead). And the Decurion detachment, which everyone uses, improves their protocols to 4+. Combine that with their HQ that gives them re-roll 1s on protocols and you've got a recipe for the most annoying army in the game. Last time I played them I charged a squad of 3 Toom Blades with my Bloodthirster and he couldn't kill them. Reanimation Protocols just keep bringing them back...

      Delete
    4. Damn, that is a low percentage. I actually wanted to try some fire warrior spam lists where you utilize their shield line and the fireblade HQ and ethereal rapid fire boosts to lay down 24-48 S5 AP5 shots per unit. Problem is this just doesn't cut it vs necrons. To kill them its all about stripping off their armor saves. This means AP3 or better (my necron friend doesn't like warriors and rolls destroyers and immortals which have 3+ armor). Tau do have an answer to this but it does invalidate my S5 lists though :(

      Delete