Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

So my question is a little bit less concrete than "should I take this, why or why not". And I'm not sure that there even is a good answer out there.. but anyways.

Obviously, luck plays a large part in the outcome of the games we play, and though a truly random set of dice will, in the long run, probably even out statistically. But we've all seen the outliers, the upsets, the just plain (un)lucky.I guess my question is multi-tiered, but here it is: Are certain builds more susceptible to chance? If so, what are they? and, are there cases in which you have to consider chance when building an army list?

They're statistically more susceptible to chance. That's all. Can be better or worse.

Yes, a turn of cold armor saves means nothing to the IG, but hurts meqs badly and murders teqs.

Each model in this game has a cost. Elite armies invest a higher number of points per model in an effort to increase their chances of success when a specific model has to roll dice. A tactical marine is roughly the cost of two and change ork boyz, but gains higher probabilities of success for the investment. With a 3+ save, statistically, it should take 3~ wounds to kill the tactical marine. With a 6+ save, the ork boyz should die to a single wound each. This implies that, if all goes to stats, the marine will be worth more than it's investment, gaining 3 times the survivability for a little over two times the cost.

However, should you have a turn of bad luck and poor rolling your marine could die to a single wound. This would mean that you had lost 2/3rds~ of the invested points in your model for no gain. If, by chance, you have great luck with your 6+ save, you could manage to shrug off 3 or 4 wounds on an ork boy, which means you have roughly doubled the amount of points you spent on that model in terms of it's effectiveness of staying on the table. Conversely, if an ork dies to a single wound, then it was something that was expected given its cost and save. That amount of failed 6+ saves mean far less because their was less value invested in the individual model.

In the end it is up to each individual player whether or not they want to invest more points per model in the hope that statistics will act closer to the average, or whether to have more models with a smaller value in order to counter act random acts of fate.

Elite armies aren't statistically worse, in fact I would think they'd be statistically better. This could be because I define statistical as acting very close to mathematically proven statistics. It's when things start going against the statistical norm that things get dicey for the MEQs out there. They just can't stand up to anomalies the same as armies with higher model counts.

Elite armies aren't worse, but they're more vulnerable to luck for a few reasons:

1) The more dice you roll the less likely you are to see significant deviations from the average. Your death star character with a few very powerful attacks is a lot more likely to do exceptionally well or fail miserably than, say, a unit of 30 boyz which is unlikely to roll very far from the average with that giant pile of dice. Of course exceptional luck, good or bad, can happen to any army, but the low-dice elite army will more frequently see it.

2) Each unit is more important. Fail a morale check with a MSU spam army? No big deal, I've got five more copies of that unit. Fail a morale check with an elite army? It's going to hurt. Of course this works both ways, a terminator squad that gets lucky with its armor saves can wreck half an army, while a Rhino that gets exceptionally lucky with its storm bolter isn't going to have any real impact on the game.

3) High averages leave less room for good luck. Consider that elite death star assault unit. When it rolls average it wipes out the target. When it rolls better than average it wipes out the target even more, but that's just wasted overkill and you don't care. So the only luck you'll ever see have any meaningful impact on the game is bad luck. You're still rolling "fair" dice, it's just that you perceive the bad luck more than the good.

4) Elite armies usually have to fight hard to win, and can't just win by default. Consider an infantry horde IG army: in theory it can win the game without ever firing a shot, simply by putting 200+ guardsmen on the objectives and waiting for the game to end. With an elite army you're much less likely to be in that position of winning by default, so you have to engage in combat, roll the dice, and hope to do at least average. So when the dice do unusual things you really feel it, while the IG player can just shrug it off and be confident that unless the dice are really weird they're not going to lose that objective.

Peregrine wrote:smart stuff

So do you think that this, as you said vulnerability to luck is significant enough to account for in army list building?

The standard deviation of dice results has a greater effect on elite armies than it does on non-elite armies.

It also depends on the game.


Like last Thursday. I played a 1500 game against BAs with my GKs. I had 15 terminators, Grandmaster, Dreadknight, and a Vindicare. He had 10 terminators, 10 assault marines, 2 predators, vindicator, 5 scouts, Librarian, and 2 Priests.

End of the game I had just my Vindicare left. he had a Priest and 5 scouts. And my Vindicare was in combat with the Priest.

It was a kill point game.

My rolling is not that great usually, so I'm a whore for divination and anything that is twin linked

 Lord PoPo wrote:

Peregrine wrote:smart stuff

So do you think that this, as you said vulnerability to luck is significant enough to account for in army list building?

No, because it can go either way. If you list-build accounting for bad-luck, sure. As they say, "Prepare for the worst, hope for the best", but never build a list expecting it to work only under lucky conditions. That's ridiculous.

I think he meant to build taking unlucky conditions into account. I don't know if that's even possible with meqs. I guess don't use terminators? Come to think of it, I don't often use terminators.

There's always the potential for horrible dice rolling (my GK terminator squad having to roll four saves and failing all four), just like there is the potential for amazing dice rolling (my 10 Necron Immortals rolling five sixes to hit with their tesla weapons). I think the point of the elite units compared to the regular units is that they can still function even when your dice are merely "below average."

Even with Vulkan I just about make Average dice which is why my Salamanders pwn hard.
My Nurgle CSM roll badly, I win through tactics and attrition.
My BA are more 'hordey' and have force multipliers (only scattering 1d6, dante precision landing, and of course Sanguinary Priests) which make it about fair for me to play.

Elite, low troop number armies require tactics to make up for the greater potential effect of variability in results. You want to concentrate down so you are making more beneficial rolls, I.e. damaging the enemy, and fewer opposed rolls, I.e. armour saves.


Automatically Appended Next Post:
If you get thiswrong that when "bad luck" can really start to bite.

In a statistical vacuum horde is better than elite army.

Elite armies have a smaller footprint and win games by consistently applying a greater amount of force in points against a lesser amount of points. It's very difficult for hordes to do that to elite armies, and when they try they tend to get clustered.

Before all those elite army players go pat themselves on the back for such brilliant tactics, keep in mind application of force.
It's a lot harder to 100+ guardsmen focus their effect on a single objective than it is for 30 marines.
One of the big advantages of being elite is application of force. It's much easier for elite armies to drop a sledge hammer on a target than it is for scrubs to do the same. The physical size of scrub armies means that often you'll have your own tanks and men in your own way.

I wouldn't say one is better or worse, or that one is easier or harder. They both have advantages. The advantage of being able to cover the whole table is off set by being spread out all over the table.

I think Peregrine sums it up best. Elite armies have advantages of their own (force concentration, mobility, will come out on top in a 1v1 style scenario, etc) but I wouldn't say they're worse than "horde" armies like IG, Orks, and Nids. They're just more likely to notice a batch of odd rolls. I've seen crazy rolls with my IG and Orks, but it's never really hurt me, because I've got so many units that I've got a backup plan (or 3) so it's no big deal. Those same rolls would've meant instant failure for an MEQ army.

You also need to take into account that "elite" armies are usually much easier to play in a tournament scenario. If you're playing draigowing against my Foot IG army in a 2 hour match, I've got to be doing all of my thinking 1-2 to even 3 or 4 turns ahead. I have to figure what I'm doing on your turn, because by the time it rolls to me, I've got to start moving models IMMEDIATELY, or risk running out of time. Meanwhile, your draigowing has a relatively tiny amount of models, giving you more time to figure out exactly what you want to do, since you won't have to do much rolling and movement. Obviously in a non tourney scenario it isn't a big deal, but it's a big killer in a tournament that has come back to bite me many a time.

So yeah, each side has advantages and disadvantages. I think it comes down more to codex type and what units you're bringing rather than how many models you're throwing on the table.

I think elite armies are worse. A concentration of points into less units means I can concentrate fire and more systematically cut off large chunks of your army. You also have fewer scoring units, again making it easier to reduce you below an effective size.

On offense, you'll have less to attack with, and because 40k rarely permits split firing, you have fewer attacks to use and often less dice to roll.

In a multi-game scenario like a tournament or league, your statistical outliers are far more likely to crop up in at least one game, costing you the win. With fewer dice, your potential ceiling of unlikely success is lower, too.

Statistically, no, because even if it were the case that over a broader number of dice rolls in a single game you would have a closer approximation of the average, the MEQs will approach the statistical average across more games.

Now, if you're talking about on an anecdotal or case by case basis, sure, but statistics, forgive me if I'm misunderstanding, but statistics is about taking a broader sample and building predictive data based off of that broad sample, which contrasts to the events of a single game in which you roll poorly. Once you're collecting a wide enough range of statistics that game of bad rolls will just be drowned by the additional games and you will still approach the statistical, or theoretical average. It may take more games, sure.

 Peregrine wrote:

2) Each unit is more important. Fail a morale check with a MSU spam army? No big deal, I've got five more copies of that unit. Fail a morale check with an elite army? It's going to hurt. Of course this works both ways, a terminator squad that gets lucky with its armor saves can wreck half an army, while a Rhino that gets exceptionally lucky with its storm bolter isn't going to have any real impact on the game.

3) High averages leave less room for good luck. Consider that elite death star assault unit. When it rolls average it wipes out the target. When it rolls better than average it wipes out the target even more, but that's just wasted overkill and you don't care. So the only luck you'll ever see have any meaningful impact on the game is bad luck. You're still rolling "fair" dice, it's just that you perceive the bad luck more than the good.

All of this post was insightful, but these are two really good points I've never seen expressed before. One thing to consider is that when you deal with hordes, you expect very little, so remember when you get good results. The opposite is true for elite units. Combine that with the human minds tendency to remember negative more than positive, and it's tough to work around.

It s a matter of perspective. I play IG. I expect to loose guardsmen every turn. If a chimera blows up and I have half a vet squad left, that is good. I expect to loose troops and a vehicle every turn. I do the most I can to keep this from happening, however, it happens.

Elite army (GK, DA, among others) expect to never fail a save of any type.

Case in point. The other day during a game the SM player drops Magnius Calgary and honor squad right in front of my line, about twelve inches away. They did not last. However, it took three chimeras, forty troops (plasma/Melta vets) a quad gun, CCS, hydra, and two sentinels. Did I mention FRF,SRF to add to,that mix. What is the point, elite armies die to mass fire power. Mass armies die to an elite group,if it gets in right and can do its job.

I would say that both are somewhat balanced in terms of power levels, but you have to factor in different things more with elite armies than with horde armies. With elite armies, you have to know how to apply each unit in your army to its utmost effectiveness, as if you aren't using a unit to its fullest you are simply going to be overwhelmed. With hordes, something I've noticed a lot of better players do is try to use their hordes for board control. Because you can flood the board with units, you can control how your opponent moves on the board, forcing them to either try to go around a unit, or shoot at a less optimal target to achieve their goals. Basically, elite armies are about using your units to their best efficiency, and horde armies are about trying to deny your opponent their chance to do so (as a note, I'm not saying that elite armies can't play board control. Some, like Draigowing could do so efficiently by plopping down a 2+/5++ squad which would only loose every other wound in the middle of the board. It's just that horde armies have an easier time doing so due to their numbers.

 TedNugent wrote:
Statistically, no, because even if it were the case that over a broader number of dice rolls in a single game you would have a closer approximation of the average, the MEQs will approach the statistical average across more games.

Now, if you're talking about on an anecdotal or case by case basis, sure, but statistics, forgive me if I'm misunderstanding, but statistics is about taking a broader sample and building predictive data based off of that broad sample, which contrasts to the events of a single game in which you roll poorly. Once you're collecting a wide enough range of statistics that game of bad rolls will just be drowned by the additional games and you will still approach the statistical, or theoretical average. It may take more games, sure.

The logical flaw you are operating under is that an individual game is a data point when, in fact, each individual roll is a data point. Any given game is, then, a sample of sufficient size to draw conclusions from. A set of games, a tourney for example, is not just a sample, but a set of samples.

The greater the total dice rolls cast by a side in a given situation, the more likely those dice rolls are to approximate an average result. Fewer dice results in a greater incidence of "outlier" phenomena. Because of this, "elite" style armies are more likely to experience abnormal rolls than hordes.

In addition, they experience a statistical "lean" effect as well: When a 5+ save unit experiences a standard roll on a set of twelve wounds it saves 1/3 of its wounds and loses 8 men. When it experiences a "top 1/2" roll, i.e. one in which all results stack in the top 1/2 of available results, it will save 2/3 of its wounds (as 5+ is 2/3 of the set "3-6") and lose 4 men. When it experiences a "bottom 1/2" roll, it will fail all of its rolls and lose 12 men. A high outlier reduces casualties by 50%, and a low outlier increases them by 50%. When a 2+ save unit experiences the same set of rolls under the same conditions they lose 2, 0, and 4 casualties respectively. Thusly, a high outlier reduces casualties by 100%, and a low outlier increases them by 100%. In both incidences the result is "balanced", but in the case of the "elite" style army the magnitude of effect is much higher.

The combination of these two facts means that "elite" armies are more likely to experience statistical outliers and when they occur, for good or for ill, their magnitude will be greater. This can be further extrapolated to games as follows: If multiple significant outliers occur within a turn then the turn itself can become an outlier- in this case a turn with more "good" outliers or "bad" outliers. Such an outlier turn would then be either a "good" turn or a "bad" turn. If it is assumed that your opponent has consistently near average rolls, then having more "good" turns than bad will result in a win, the opposite will result in a loss, and an even distribution will result in a game decided solely by player skill.

An elite army is more likely, both by frequency and magnitude, to generate enough significant outliers to "swing" a turn one way or the other. They are therefore more likely to have a game that is swung in one direction or the other. Assuming a relatively even spread, about a third of their games will end up in each category (win/loss/skill determined). A player with skill above that of his local meta would thus win about 2/3 of their matches, an average one would win 1/2, and a poor one would win 1/3. The same players, using a less elite army, whether horde or MSU, would have "swung" games much more rarely. Let us assume they have "swung" games at half the rate. They would then have 5/6, 1/2, and 1/6 win proportions respectively.

It can thus be assumed that, for the average player, it makes little difference. For the poor player, playing elites and hoping for good rolls is a viable strategy. For the above average player, armies that perform consistently are less likely to "screw" you out of what would otherwise have been a win.

Please note the following:
Skill comparisons are based on your own local meta, not on objective skill rankings. For example, at a local tourney, I should be well served by MSU. At Adepticon I had best bring a combination of Paladins and prayers.

This does not account for the "Rock, Paper, Scissor" effect, nor does it account for players of such exceptional skill that they can win a game after losing half their army to bad dice on turn one. It also fails to account for exceptionally good or bad army design.

These principals can, however, be extrapolated to things other than "Elite vs. MSU", for example the perennial "Autocannon vs. Lascannon" debate: Autocannons are prized for their consistent efficiency, whilst Lascannons are highly variable, but have impressive peak effectiveness.

I think you're thinking of the question the wrong way. If I use the past as a guideline, traditionally, exceptionally powerful, "cheesy" armies have methods to bypass standard damage (i.e., basic attacks or basic shooting.)

6th edition flyers (immune to hth, 6's to hit)
5th edition mech (immunity to psychic powers, bolters, very little damage to embarked units when vehicles died)
4th edition eldar (skimmer moving fast bypassed penetrate and armor 12 bypassed bolters)
3rd edition, mech assault, like 5th except more assault based instead of shooty

Then there's things like Draigowing, Nob bikers, infiltrating siren prince in 3rd, Tyranid monster army in 4th edition, etc etc. All very very resilient to basic shots and basic attacks. In some cases, like with the infiltrating prince, completely immune to shooting.

It's not a question of elite or not. Many elite armies were very good. Many nonelite armies were also very effective. Your mileage may vary.

And very small addition coming from orks:

If you have less models to move you have more time to think, that's good for elites.

 Tuagh wrote:
(All that great stuff he said.)

Just to reiterate, in organized play your performance is based on multiple games, be it in a league or tournament. Elites deal with outlier games more often, which means your overall performance is more affected. If you play 4 games, it seems likely you'll experience at least one outlier with an elite army, possibly two. If your regular play gets you a win, you're only hurt by your "bad" game. If your regular play can't be relied on, or gets you a loss, you weren't going to win anyway, so just have fun!

I think any statistical advantage they have would probably be drowned out by imbalances in codexes.

 Tuagh wrote:

The logical flaw you are operating under is that an individual game is a data point when, in fact, each individual roll is a data point. Any given game is, then, a sample of sufficient size to draw conclusions from. A set of games, a tourney for example, is not just a sample, but a set of samples.

The greater the total dice rolls cast by a side in a given situation, the more likely those dice rolls are to approximate an average result. Fewer dice results in a greater incidence of "outlier" phenomena. Because of this, "elite" style armies are more likely to experience abnormal rolls than hordes.

In addition, they experience a statistical "lean" effect as well: When a 5+ save unit experiences a standard roll on a set of twelve wounds it saves 1/3 of its wounds and loses 8 men. When it experiences a "top 1/2" roll, i.e. one in which all results stack in the top 1/2 of available results, it will save 2/3 of its wounds (as 5+ is 2/3 of the set "3-6") and lose 4 men. When it experiences a "bottom 1/2" roll, it will fail [...]

No, that's not a "logical flaw," I even said in the passage you quoted outright that within a single game a higher number of dice rolls will better approximate the average. By contrast it takes more games to approximate the average with the elite armies due to fewer dice rolls.

And no, fewer dice rolls do not result in a greater incidence of outlier phenomenon. The chance of something happening is simple probabilities, you take the number of outcomes and the number of "good" outcomes and that's the probability of that good outcome happening. The probability of that occurring is not based on the number of dice rolls, it's based on the probability.

You're conflating the idea that the magnitude of these results has anything to do with the probability, and you're also conflating the idea that the statistical results of a given set of dice rolls has some impact on the actual theoretical probability. The fact that statistical results approximate theoretical probabilities with more data points is merely a confirmation of the theoretical probability.

I can get 14 heads in a row and 1 tails, but the probability of the next toss is still 1/2 and that is not affected in any way whatsoever by the preceding results.

What this thread is about is the idea that the "magnitude" of dice roll failures in elite armies is more significant than horde armies. It is worth noting however that elite armies typically have better probabilities per dice roll, e.g. a 2+ save versus a 5+ save or a 3+ to hit versus a 5+ to hit. They also tend to have more shots or more attacks. And BTW, in what way is losing an entire squad of 50 p infantry due to a !@#$ leadership roll any better than losing 1 Terminator?

Another thing you people are failing to mention is that Elite armies in this game are typically not subject to the same vagaries of Leadership and morale issues, they're not subject to fear tests, many units like Purifiers do not have to take Pinning tests, they can't get swept; that right there is a huge number of probability checks you don't even have to bother with. A lot of what you pay for with Elite armies, for instance high AP values, AP3 or AP2 which bypass armor save rolls entirely, impact hits which strike at I10 and autohit, or a large number of attacks on expensive pieces of wargear like Power Weapons , is actually consistency. You're paying in a lot of cases for a guaranteed result on a morale check or a sweeping advance that for instance a Guardsmen squad doesn't have. A guaranteed result is always more consistent, because a 1.0 result is a 1.0 result.

Lord PoPo wrote:vulnerability to luck is significant enough to account for in army list building?

It's certainly enough to affect the points cost. Another good example is the deathstrike missile. It has the potential to be WAY more game-changing than a manticore, but the two cost the same because a deathstrike is a lot less likely to get to do any damage at all.

That said, I think it's folly to bring the law of large numbers in here. In a single game of 40k, you're only rolling a tiny number of dice, and only a few of those die rolls really, REALLY matter. 40k itself is a game of luck. Taking units that are only a teensy bit less prone to it doesn't make a difference at all most times.

As for if elite armies are worse or not, I'd say no, now. In 5th ed, hordes were generally better, as it was easier to knock out a few transports/tanks with missile launchers than it was to clear 200 greenskins out of universal 4+/5+ cover. Plus, elite armies tended to get horribly bogged down in close combat, which the best hordes were able to take advantage of.

Now, however, there has been a massive move towards shooting, which helps elite armies relatively better than horde armies (for example, marines always took their 3+ saves against small arms, but for a guardsmen that bolter save went from 4+ cover to 5+, or with focus fire, - ). Also, elite armies got better in close combat relative to horde armies (overwatch, challenges, etc), which really gives elite armies a relative leg up compared to 5th ed.

Are elite armies BETTER than hordes now? I'd hesitate to say, but it's certainly rather close, rather than being pretty lop-sided for hordes.

In 5th ed, I'd table wing armies with just infantry platoons. The last time I did that in 6th, I needed the help of russes and other decidedly elite guard units.

Ted, you're talking about the chance that a particular instance, i.e. a particular dice roll, will have a particular result. You're right that past results do not affect future results.

But this thread isn't really about that. It's not about the average result of a single dice roll in an elite army (As you pointed out, most elite armies actually have a better chance to have a "good" result on any single roll of the dice, because they tend to have higher stats and better gear.). This thread is about the variance of results in an elite army as opposed to a horde army. Now, I'm not a statistician, but the basic idea is that variance measures the average amount of deviation from average of a given data set. Smaller data sets usually have higher variances than larger ones, because there's less opportunity for randomness to get smoothed out by repetition.

Take the example of the coin flip. When you flip a coin twice, the "average" result is one heads and one tails. But there's actually only a 50-50 chance of getting the "average" result: 1/4 of the time you'll get two heads, 1/4 of the time you'll get two tails, and 2/4 of the time you'll get one heads and one tails.

When you flip a coin four times, the "average" result is two heads and two tails. The chance of actually getting the "average" result is 5/8: 1/16 of the time you'll get four heads, 2/16 of the time you'll get three heads and one tail, 2/16 of the time you'll get three tails and one heads, 1/16 of the time you'll get 4 tails, and 10/16 of the time you'll get two heads and two tails.

Thus, you're more likely to get an 'average" result when you flip four coins than when you flip two coins. also, your result is more likely to be closer to the "average."

The point, then, is that smaller armies roll fewer dice over a single game, so the set of dice rolls for a smaller army over a single game has a relatively higher variance.

Exactly, and the longer you play 40k, the more you'll start seeing this random trending thanks to deviance.

I mean, we all know the anecdotal stories of the terminator absolutely refusing to rail a 2+ armor save, for example, but if you start really paying attention to your dice, you'll find that deviation is shockingly common.

To the point, in fact, where knowing statistics is all but useless in an actual game of 40. Really, it's use is limited to other sufficiently abstract domains, like list building.