Jump to content
Welcome Guest!

Join us now to get access to all our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, and so, so much more. It's also quick and totally free, so what are you waiting for?

jamierk

Top Tier Units Pre and Post GHB2017

70 posts in this topic

Keldaur    18
10 hours ago, Auticus said:

Thats ok I'm done defending it :) it has served me quite well over the years, and took me to top placings in Grand Tournaments back in the day.  I have absolute faith in it.  It is a tool that gives you a great starting baseline.  If you're looking for something that is absolute, I don't feel that that is possible, even with complex game theory calculations, because there is too much abstract in the system that you cannot quantify with calculus.  However... those formulas will tell you what needs babysat and also will help you formulate a power coefficient to determine if your raw without buffs list as as efficient as it could be before you have to start taking buffs and powers into account.

I am not attacking it. I am saying that contextualizing units and player's experience is equally important when it comes results. Would you say that your high placings were solely on decissions based by this data ? No, so that's my point. I am not saying it's not useful, i am saying there are other factors that can be equally or more important than data (using kurnouth example) when it comes to nuance the army list selection or plays.

 

Share this post


Link to post
Share on other sites
Auticus    1,021

I'm not disagreeing, nor was I pushing that the data is the sole factor to consider.

It is akin to using physical records that professional sports use to track their athletes (40 speed, bench press, vertical jump distance, etc).  An athlete running a 4.2 40 yard dash gives coaches a lot of indications on his utility on the field as opposed to an athlete that can only run a 4.8.

The other half of this data which is not for public consumption are the modified stats based on buffs, so that you can compare for example what the output of 3 units of reavers is when next to a bloodsecrator and when not next to a bloodsecrator (among many other things)

I have all of the data that I care about (my builds) and was going to try to do combos outside of my force, but the backlash that it took kind of made it not worth the effort so its been binned and I use it for myself and update my own buffs as I need (or if I know I'm facing a combo I'll put that combo in as well to see where it stacks and where my overall combo sits)

Edited by Auticus

Share this post


Link to post
Share on other sites
WoollyMammoth    366
21 hours ago, swarmofseals said:

a lot of stuff :)

Interesting point about the tarpit size. I always thought of a tarpit as a large thing, because if the tarpit is just a tar-puddle, you can just step out of it.  The examples you give are useful only in regards to limiting the movement of the target. If you charge a unit of screamers, you will have to survive through its attacks in order to pin it down, so only 5 wolves is risky. Screamers can also still retreat and activate their slashing fins, so you are not wasting a ton of their time. A LoC power is mostly in its spells which are not limited by some wolves snapping at it's heels. Shooting as well, is only limited if you prevent it from moving into range of what it wants to shoot. Then if you factor in all the special rules, like the fact that Skaven can just retreat and charge, or Stormcast with a Knight-Heraldor - tarpitting is extremely limited in AoS. The Dire wolves are a great unit for this though with their fast movement, you may be able to pin down a melee only unit for a little bit. Personally I prefer to use their long bases to create a wall instead, forcing my opponent to charge wolves instead of what they really want to attack. Yesterday I used them to screen some Ghouls, so they had to charge the 'wolf wall', and then my ghouls were 2" behind to get a free pile in.

When I think of a true tarpit in AoS, I think of a large unit of at least 20 models, that can physically surround the target and prevent them from retreating. That's a tarpit you aren't walking out of. This can pin through several turns stacking point efficiency. If we are taking ghouls or skeletons, these units are likely to even hurt or kill the target it is pinning. 

Overall, I don't see a lot of tarpits in AoS. Compared to 40k, where a small fast unit could pin down a tank from firing its devastating weapons, -or as you mentioned in WHFB where there were limitations on facing. There is not much you can do with a tarpit in AoS. 

An anvil has many uses, such as the tactic you describe to pin and attack. Technically, the tarpitting I describe above is better for that. Zombies can surround and completely pin a target, leaving a 1/2" gap somewhere for a nasty monster to come in or a unit (preferably with long rage weapons) to charge. If a unit is pinned by a Mourngul and they see a hammer unit coming at them, they can simply retreat. Most of this is moot since everyone primarily focuses on fast-moving self sufficient hammer units in AoS.

Another use is as I described before, using the bulk of an anvil to body block an objective while a weaker horde unit crowds around the objective to keep plenty of bodies on it. An anvil can be a tough monster, or a tough unit. An anvil is simply a unit that can take a punch, therefore you get it to where it wants to be and then you stay there. Generally its smarter to ignore the anvils and focus on killing them last. A Mourngul being an anvil means your opponent needs to try to ignore the -1 bubble. Its easy to stick him in the middle of your army so his bubble is being useful and if your opponent wastes shots on him its usually a good thing.

I can't disagree that the LoC is one of the best monsters. A Phoenix is really good and at a bargain but I have never seen it used so I cannot say. A Mourngul just has insane stats and rules which is why its 400 points, but it is a bit overrated since it cant hold objectives or take artifacts. It sets up some incredible synergy though if you stack -to hit with all the death stuff, or in Nighthaunt which has almost nothing to take anyway. I wold say any death list can get great use out of it if used right and/or if your opponent over targets it. I'm interested to see how it holds up in the new environment. 
 

Edited by WoollyMammoth

Share this post


Link to post
Share on other sites

Watching the conversation is fascinating.  We are having a discussion about terminology and reaching a consensus.  Man, I love TGA.

I would also have to separate distinctions that you are describing: Tools versus Unit Types.  To me, you have Tools, or strategies, that your army can use, while the Unit Types get to the specific tactics of a battle.

For example, we an Anvil, a large, heavy unit that can take hits and stick around for a while.  We also have Hammers, that apply lots of brute force to the enemy, and Spears, which use precision and finesse to attack the enemy.  Whereas Unit Types cover Tar Pits, Elites, Heroes/Monsters, Cavalry/Chariots, and Skirmishers.  There can be some overlap with these different distinctions dependent on the army; Cavalry units between armies cover Hammer, Spear, and Anvil, and a Hero/Monster can easily fill any of those roles depending on their specialty.  But in some cases, overlap would be problematic.  Like, would any Tar Pit be able to Spear into the enemy lines, or would a Skirmisher unit want to be an Anvil?  Probably not in both cases.

  • Like 1

Share this post


Link to post
Share on other sites
Sastopher    4
11 hours ago, Auticus said:

I have all of the data that I care about (my builds) and was going to try to do combos outside of my force, but the backlash that it took kind of made it not worth the effort so its been binned and I use it for myself and update my own buffs as I need (or if I know I'm facing a combo I'll put that combo in as well to see where it stacks and where my overall combo sits)

This makes me really sad. I think this data is obviously and intrinsically useful. These sorts of quantitative tools are what allows us to elevate decision making.

Along these lines, would you be willing to open source any of this work for those who want to tinker with it? I personally don't care if it's rough around the edges: being able to compare along the bell curve of other units and synergies is a ton of work to imitate from scratch. I'm also curious how some of the silly combos I've found match up on that spectrum.

Share this post


Link to post
Share on other sites
Auticus    1,021

I have a mostly done beta of the page where users can add buffs to a unit and say where that buff is coming from, which lets you calculate the buff score of the hero or wherever its coming from.  (so that I didn't need to be the person entering all the data)  I used that in my last campaign to gauge actual power levels and give handicaps based on it.  Using that system, any difference in score of about 20 or more was 100% in choosing who the winner was going to be.

The flaw in my score is that I need to figure out how to make ranged scores worth more because an all melee army with a high score will still be less impactful than an all shooty army that has the same score, because the shooty army can reach out with range and the melee army cannot.  

I have to address those flaws first, which I'm hoping to start later by the holidays.  Right now I'm writing the 2nd editions of my 40k grand crusade campaign and my Azyr Empires for AOS for campaign next year.  

I'm not really a fan of open sourcing my code though.  

Share this post


Link to post
Share on other sites
3 hours ago, Auticus said:

I have a mostly done beta of the page where users can add buffs to a unit and say where that buff is coming from, which lets you calculate the buff score of the hero or wherever its coming from.  (so that I didn't need to be the person entering all the data)  I used that in my last campaign to gauge actual power levels and give handicaps based on it.  Using that system, any difference in score of about 20 or more was 100% in choosing who the winner was going to be.

The flaw in my score is that I need to figure out how to make ranged scores worth more because an all melee army with a high score will still be less impactful than an all shooty army that has the same score, because the shooty army can reach out with range and the melee army cannot.  

I have to address those flaws first, which I'm hoping to start later by the holidays.  Right now I'm writing the 2nd editions of my 40k grand crusade campaign and my Azyr Empires for AOS for campaign next year.  

I'm not really a fan of open sourcing my code though.  

The way I did my one was to assume ranged is of equal value to melee up to a certain range then apply a modifier as range increased. 

12" and under x1

12-24 x1.25

24" and over x1.5

reason being is the expected number of attack rounds per game for a ranged unit is about 3-4, 4-5 and 5 respectively. Accounting for deployment and distances etc. 

Most combat units in my experience get a similar amount of overall attack rounds depending on their speed which is accounted for seperately. 

Share this post


Link to post
Share on other sites
Auticus    1,021

Cool.  I had something similar with the ranges.  Little different in the slider values but same concept.

Share this post


Link to post
Share on other sites
Sastopher    4

Interesting. It was my rough perception that melee attacks should be valued about the same, given you're trading less flexibility for being able to swing on both players' turns.

  • Like 1

Share this post


Link to post
Share on other sites
Andreas    221

Yes interesting discussion. My gut feeling is that your ranged modifiers are a bit low. You can get more models to attack when you shoot and you can concentrate the damage. But I have no clue what the values should be.

Just curious have any of you thought about any modifiers if the attacks are less random. An extreme example 60 attacks that causes one damage each on 6+ is better than 1 attack that causes 60 damage on 6+ in most cases. A common example would be how much better is 3 attacks 3+ 3+ damage 1 compared to 1 attack 3+ 3+ damage 3. Or an attack with damage 2 compared to damage D3.

 

Share this post


Link to post
Share on other sites
Auticus    1,021

The models I use use the average.  So for example D3 shots = 2 shots.  D6 shots I have = 4 shots.  Things like that.  

Melee attacks I don't feel are the same at all.  If I take a unit that has 30 shots that after hits and wounds average 5.5 damage a turn and you have a unit that has 30 melee attacks that after hits and wounds average 5.5 damage a turn, but I have a 20" range I get to shoot at you for 1-2 turns and remove models from your unit before you get into contact with me.

By the time you get into contact with me you've taken 11 damage and lost a bunch of attacks and are thus less effective.  

Share this post


Link to post
Share on other sites
daedalus81    689
17 hours ago, Andreas said:

Yes interesting discussion. My gut feeling is that your ranged modifiers are a bit low. You can get more models to attack when you shoot and you can concentrate the damage. But I have no clue what the values should be.

Just curious have any of you thought about any modifiers if the attacks are less random. An extreme example 60 attacks that causes one damage each on 6+ is better than 1 attack that causes 60 damage on 6+ in most cases. A common example would be how much better is 3 attacks 3+ 3+ damage 1 compared to 1 attack 3+ 3+ damage 3. Or an attack with damage 2 compared to damage D3.

 

I have a tool that addresses this and produces a more realistic view of outcomes.  I just don't have all the data loaded and I need to host it.

Essentially it rolls dice across 100,000 iterations and graphs the outcome.

Share this post


Link to post
Share on other sites
Keldaur    18

Not to speak that it's much easier to get to target priority targets easier shooting than on melee.

Edited by Keldaur

Share this post


Link to post
Share on other sites
rokapoke    577
On 9/17/2017 at 11:13 AM, daedalus81 said:

I have a tool that addresses this and produces a more realistic view of outcomes.  I just don't have all the data loaded and I need to host it.

Essentially it rolls dice across 100,000 iterations and graphs the outcome.

I'm curious if that really shows a difference. Ideally a dice roller is a uniform distribution variable, so each value is equally likely. I don't see how you'd find a difference in rolling 100,000 dice versus just calculating mathematical probabilities.

I'm seriously not trying to poo-poo your tool. I'm genuinely curious if there's a difference (as we're all, I imagine, well aware of the perceived differences between the chance to roll a 4+ and the actual result, for instance).

Share this post


Link to post
Share on other sites
daedalus81    689
Just now, rokapoke said:

I'm curious if that really shows a difference. Ideally a dice roller is a uniform distribution variable, so each value is equally likely. I don't see how you'd find a difference in rolling 100,000 dice versus just calculating mathematical probabilities.

I'm seriously not trying to poo-poo your tool. I'm genuinely curious if there's a difference (as we're all, I imagine, well aware of the perceived differences between the chance to roll a 4+ and the actual result, for instance).

No worries.

When you calculate the math you're looking at the average.

What's the mathematical difference between 20 4+/4+ attacks that do 1 damage and one attack that does 20 damage?  Nothing

But you and I both know that there are real in-game consequences to each scenario that don't work out that way.

To put it visually:

image.png.e5d4b7ca5cd49e788130ccd8f506ff41.png

Share this post


Link to post
Share on other sites
rokapoke    577
1 hour ago, daedalus81 said:

To put it visually:

image.png.e5d4b7ca5cd49e788130ccd8f506ff41.png

Good point; however, I wonder how many rolls have such misleading distributions. I wouldn't expect your typical result to have such an extremely long tail, for instance, to drag down the average. 

Either way, I think it would be really interesting to see the output you get!

Share this post


Link to post
Share on other sites
daedalus81    689
16 minutes ago, rokapoke said:

Good point; however, I wonder how many rolls have such misleading distributions. I wouldn't expect your typical result to have such an extremely long tail, for instance, to drag down the average. 

Either way, I think it would be really interesting to see the output you get!

There are insights to be gleaned, but no absolute truths.  That particular tail represents something like 0.1% of all outcomes so the exceedingly rare stuff.  It's more noticeable when you see the whole chart.

Another piece that i've been working on is setting formations.  So, if you have a blob of 40 models on 32mm bases and you specify them to be running 8 wide it calculates their surface area.  The selected opponent will also have a surface area, say a Stardrake.  It then tries to determine how many will get into base and calculates the fight from there.

Edited by daedalus81
  • Like 2

Share this post


Link to post
Share on other sites
Sastopher    4

BTW if people are interested in playing around with numbers on their own and don't want to create their own probability engine, I recommend a public tool I use called AnyDice. I use it quite a bit and it's extremely customizable with functions. Here's a link to one of my workspaces calculating buffs on stormfiends: http://anydice.com/program/d015

Edited by Sastopher

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×