| | You are not authorized to post a reply. |
| | Author | Messages | |
elder_basilisk
Sergeant
410 Posts
    
 | | 06/26/2006 2:35 AM |
| One player I know has chosen his warband on several occasions so as to minimize variance. His theory is that, by picking units like the helmed horror that have high attack bonuses and strong defenses and immunities to crits, etc, luck should prove to be less of a factor in his games. However, that does not appear to have been the case as he has recently lost several games after a few strings of bad luck. How are we to understand this? Is he wrong about the helmed horror band being low variance? Is he wrong about low variance bands minimizing luck?
I think there may be a couple answers to this question and will toss them out for us all to discuss.
Answer #1. If we grant the premise that a band like helmed horrors that has low but consistent offense and strong defense minimizes the impact of luck in the outcome of any particular match, then that means that, assuming similar player skill, a match featuring lower variance bands should have a more predictable result. This would, however, apply to unfavorable matchups as well as favorable matchups. If, for instance, Marut+Coatl is a bad matchup for helmed horrors, then the predictability of the band works against it--it consistently loses.
If we were to exaggerate the consistency of bands to the maximum extent (the band with the favorable matchup always wins) then the game would come down pairings with the outcome resolved by how accurately the player judged the meta and how his pairings worked out. That is, of course, an exaggeration (GAS can beat helmed horrors occasionally), but it demonstrates that even a perfectly consistent band (which no "low variance" bands are) would not eliminate the role of luck in tournament settings; it would rather exaggerate the effect of pairings. (I think helmed horrors do this already--since they perform much better on some maps (hellspike or drow outpost, for instance) than others (dragonshrine comes to mind).
Thus a low variability warband would simply transfer the effect of luck to a different level: the more reliable your warband is in a match, the more important the parameters of that match (the warband pairings, map, and side inits) become. There is still variance (and quite significant variance if map and side init are significant in the matchup--as they usually are for horrors. That's two very big rolls to wager your match win on).
Answer 2: This observation goes to the effect of luck on low-moderate damage/high attack bonus warbands. A creature pays a lot of points for a high attack bonus (especially in Lawful Good apparently). That means that it has to hit consistently in order to be effective.
To illustrate the point, let's consider a hypothetical matchup between two fiftyish point hitters: the justicator and the frenzied beserker. Depending upon initiative, it's a match that could go either way. However, the high attack bonus/moderate damage justicator HAS to hit every attack in order to win. The frenzied beserker expects to miss at least one attack and still have the potential to win the battle in its third activation.
To illustrate the point:
Round 1:
Justicator move+smite chaos-25 damage on beserker
beserker move+attack-30 damage on beserker, 1 potential hit
Round 2:
Justicator attack twice-60 damage on beserker
beserker attacks twice-65 damage on beserker, 3 potential hits (the battle could be over at this point, but lets assume only 2 hits and that the justicator makes morale).
Round 3:
If the justicator wins, he attacks twice and kills the beserker (he may die due to deathstrike, but that's not really important right now)
If the beserker wins, she attacks twice and with 5 potential hits, may well have the 3 actual hits that she needs to kill the justicator.
The long and the short of it is that, while the justicator is significantly more likely to hit every attack (especially considering the frenzied beserker's armor class), it also needs to hit every attack. Thus while you can take rolling the odd four or eight and still hit (you still hit a marut on the primary attack with a justicator on an 8 actually), you still miss if you roll a 1 (or sometimes a 4 or even an 9 against higher AC creatures and especially on the second attack. Consequently while a series of slightly lower die rolls can be dealt with the occasional outlying number--the one, etc-- has more impact. A creature that makes up for low-moderate damage by doing it reliably really suffers if it doesn't actually deal the reliable damage.
2.5 In the situation where you do roll two ones on your two attacks and whiff completely, it helps to consider what has to happen to pull you out of the hole. Crits should theoretically happen as often as "1s" (don't ask me why it didn't seem like it worked that way for me at the qualifier; I'm just forgetting the matcher where I critted but didn't roll a 1 and ignoring the 20s rolled for morale saves, etc:)). This is another case where reliable but mediocre damage output suffers (especially for helmed horrors and maruts--because elemental damage isn't doubled on a crit). Much like in the bad matchup you want a bit of variability so that it is possible to win, when you're losing (because of mistakes, poor luck, or a poor matchup), you want that chance to get back in the game with a lucky roll or two. When playing it safe means that you're going to lose, you want a mini that will be able to swing for the fence.
The point is not that high variability bands are good here. Rather, I'm arguing that early bad luck that puts you in a losing position is harder to make up for the lower variance your band is. Thus a low variance band doesn't simply minimize the significance of luck in the match itself. Rather, it increases the significance of bad luck early in the game while reducing the significance of good luck later in the game. (To point out one situation where this happened for me in the qualifier, I critted a frenzied beserker with one of my justicators. That was good luck, and it was significant. However, the net effect was that I got the extra damage I needed to kill both beserkers in a pair of activations. Had it been an ogre ravager that I had critted with rather than a justicator, the effect would most likely have been two dead beserkers in a single activation and the second ravager moving to base and kill my opponent's pyromancer. Similarly, a crit against my opponent's coatl in a later round morale checked the coatl but didn't kill it outright as it would have done with a higher variance piece).
Conclusion: A "low variance" warband like helmed horrors or marut+coatl doesn't seem to simply eliminate luck. Rather, it emphasizes the factors that determine the conditions of the match (matchups, pairings, map, and side initiatives), and reduces the significance of mild abberations in the dice. Large abberations of bad luck (rolling 4 1s in a row, etc), however become more significant because the figures' costing is paying for reliability, but large abberations of good luck are less significant (four 20s in a row with a helmed horror will morale check a hill giant, but four 20s in a row with a hill giant will usually win a match in a blowout). Furthermore, it is harder to come back from an actual disadvantage (as opposed to a temporary disadvantage in points or something else that is likley to change during the match) through good luck with a low variance warband. So, I think my fellow here is only partially right. A low variance warband will serve to minimize the significance of small deviations from expected rolling in a game. However, it may emphasize the significance of early bad luck.
Note 1a: Why is it then that, at least for a while, many of the top players prefer low variance warbands? I think that is not as much because of how they effect luck, but rather because it is more possible to plan ahead with a low variance warband. With a marut+coatl you can move six and pretty much count on putting 60 damage on something and that enables you to calculate in advance whether or not that is advantageous. With a pair of ogre ravagers, on the other hand, you can calculate how far you can move and attack, but you can't predict the results of the activations as far in advance. (With marut+coatl, you could say "I move 6, attack, snake's swiftness, and then before next round is over I will have done it again and that leaves me with (for instance) a marut facing one 5hp Duergar Champ. At that point, I magic missile him for the win" On the other hand with the pair of ravagers, you have more trouble predicting what will happen 2-3 rounds into the future.
Note II: low variance and high resilience are not exactly the same things. For instance, at least according to some people, SWarm is a low variance band. Everyone is fearless, the sword archon usually hits (well, he's supposed to anyway, last time I ran SWarm, I don't think he was better than 50/50 in the tournament and that was against hill giants, frenzied beserkers, and duergar champs not opposing sacred watchers). However, it is not particularly resilient. If an opposing kord, for instance, gets hot and passes four consecutive incorporeality checks, it's pretty hard to come back with just two sacred watchers and a sword archon left). On the other hand, triple justicator+CoDA while also a fairly low variance band, is a rather resilient one because the 30 point cures on 80 hp justicators give the band a little bit longer before the effects of a few luckier than expected attack rolls (from most figures) impact the band's fighting capability.
So, what do you all think? | |
| |
|
Zippy
Underboss
1993 Posts
Whitewater, WI
| | 06/26/2006 2:54 AM |
| 1) "High" and "Low" variance are loosely applied relative terms.
2) Consider that a low variance band can have inherent weaknesses, where matchup plays a significant factor. Helmed horrors led by snig, for example, are going to lose initiative a lot on average, so be careful not assign such a thing to bad luck.
3) It's easy for a small group of rolls to appear "out of whack" - it's over a lot of rolls (many games) that average and reality tend to converge more often
4) Autodamage might be considered even lower variance than than a constructs band. A beholder band might exemplify high variance. His helmed horrors band may be toward the low-var end of the spectrum, but it does not define the extent of possible low variance warbands.
just food for thought... | |
There are 10 kinds of people in the world; those who understand binary, and those who don't.
Reference Thread, H/W List, Champion of the Catoblepas | |
|
Pegasus Knight
Sergeant
896 Posts
| | 06/26/2006 3:06 AM |
| Definitely a good look at the topic. I do have one comment however, regarding resiliency in bands. Specifically, you use the Cleric of Dol Arrah's cure spells as an example of this.
Your point is correct, but there is a second way to look at it. One thought is that curing damage forces more dice rolls to be made (more attacks, specifically, are required to kill the piece that was cured). The net effect of having more die rolled is that, theoretically, the dice 'average out.' The longer the game takes, the more die that are rolled, the more likely it is the rolls will curve back to averaging around the 9-11 range in general.
This isn't quite the same concept as "low variance", you're right. But it interestingly enough arrives at similar, albeit not identical, effects. Healing damage is a way of reducing the impact of dice-flukes, if done in significant amounts. Curing 5 HP might not matter; curing 60 sure will affect the math involved and usually level it out over the next few attack rolls on that creature.
I do agree with your thoughts that most tournament players aim for low variance because it's easy to plan around mathematically. A fascinating example of this is how Maruts qualified more than Aspects of Kord this year. A bad morale roll ruins Kord's game; Marut doesn't care about them. A natural 20 can suddenly kill what was thought to be a relatively safe-for-the-moment Kord; Marut doesn't have this problem. While the 'low variance' issue does increase the effects of early bad luck like you said, it does indeed make it easy to tell what's going to happen. This really isn't a new point, just supporting what you said in Note 1a. | |
- Irrationally Fanatical Champion of Pegasus-mounted cavalry
- Proud member of Team Low Tier Beasting: I play CG as my main faction!
- Garland, TX 2006 Qualifier Champion
My trading thread: http://www.maxminis.com/forums/topic.asp?TOPIC_ID=19725 | |
|
Weaponbreaker
Skirmisher
47 Posts
| | 06/26/2006 6:18 AM |
| I agree with Zippy here, The terms and degrees of variance are subject to many things the single most important of which is the skill of the player. You want a better chance to beat a better player you HAVE to play a low variance band, but you will probly still lose. That does not mean the warband was flawed but that you plan of attack and understanding of the meta-stratagy were not up to snuff. I have seen many 'high variance' warbands tear 'low variance' warbands to shreads on the simple fact that the player was better at playing his warband. Take a look at the numbers at each of the qualifiers, some really good players took the day by simply being better, even though using the numbers they should not have a very high chance of winning. If you run the numbers an IM/AM should never beat Coatal/titan but they have. I am admittly a math retard but I can still build a pretty good warband that can win.
Although I think this is a worthy topic for discussion I tend to be on the side of learn the game, not the math and you stand a better chance of winning. | |
Hunter of CE d20's. Destroyer of false hope and punisher evil dice across the planes. | |
|
XAos
Underboss
2413 Posts
London
| | 06/26/2006 7:03 AM |
| There is also the question of the opposing warbands variance.
e.g. Playing a HH warband reduces the variance. But if your opponent plays dual HGB, the game will still be heavily controlled by the dice.
| | | |
| 
Vrecknidj
Warlord
10493 Posts
United States
| | 06/26/2006 8:33 AM |
| Wow, excellent exposition. That was a very long read, but one worthy of the attention of pretty much all the players of the game. Even if someone disagrees, even if that someone disagrees with some fundamental points, such disageement then requires defense and justification.
You make some very good points and the piece should serve as an excellent learning tool for the newer players.
During the weeks between the Ypsilanti and Niles qualifiers, I was practicing as I usually do--with many different bands, against many different bands, most often simply by playing bands against my friends' bands to help them practice.
I discovered that the Helmed Horror, for all its sturdiness, isn't all that powerful in today's meta. Sure, it's a forgiving piece. You can make some mistakes with a multi-Horror band, and still come back from them. But, you can make some mistakes with a Marut and come back from those too (though it's harder, with the Marut, if those mistakes do 25 damage at once, and you have no way to "heal" that Marut). But, the offensive power isn't great enough to pull you over that hump in the tough matchups.
I agree that, in general, low-variance bands are more likely to climb to the top in a multi-match event. I don't know that anyone played a Beholder band even into the top 8. It's just not feasible right now.
But, there are many ways to achieve this low-variance, and some of them are achieved through combining low-variance pieces (i.e. a triple Helmed Horror band), and some are achieved through very resilient synergies (the SWarm and the Kord/Bodyguard/CoDA come to mind).
Right now, it appears that the low-variance bands that are due to resilient synergies are outperforming those that are due to the inclusion of multiple, low-variance figures.
Dave | |
Knowledge Arcana editor issues 5-9, Phoenix Lore Magazine editor, assistant editor for Rite Publishing;
My Trade Thread and My Reference Thread; Winner of WBC IV, IX and XIII; Rule #0: bshugg is always right! | |
|
doranur
Warrior
180 Posts
| | 06/26/2006 9:10 AM |
| nice article. i agree that lower variance bands are usually much more affected by strings of bad luck than high variance bands. bands with distributed hitting power (4 or more significant sources of damage) are also less effected by strings of bad luck than bands with less distributed hitting power (1 or 2 sources). high damage bands benefit more from positive luck than low damage bands.
if we all agree that the above is true, then the most consistent bands (rolls with bad luck, benefits from good luck) should be high damage quad hitter bands, which is why CE quad was so good for so long (and not bad now either!).
i disagree with the statement that less skilled players should play lower variance bands however. less skilled players typically have less of an idea of how the flow of the game is going to work, they react more on a turn by turn basis rather than specifically planning out 3 to 4 moves ahead. higher variance bands give a specific advantage in this regard; a string of "good" luck in a high variance band dramatically alters the game board, giving an advantage to the seat of the pants player (also the plan ahead player benefits, but less so since his careful planning is no longer as useful). a string of bad luck hurts both players, but once again the plan ahead player is more hurt since his careful plan did not include a really awful string of rolls (4 1's in a row with ulmo for example).
i'm not trying to say we shouldn't plan ahead, but if one player benefits more from a band that is more reliable but less dynamic and the other player benefits more from a band that is highly dynamic and less so from the mathematical probability, those sorts of players should gravitate to the bands that help them the best.
i am a perfect example of the seat of your pants player. i can generally have a good guess of 1 turn ahead what the board will probably look like. i can understand simple traps, and some fairly basic strategies, but often my main strategy is to consistently attempt to do the most damage to the opponents warband while i a) wait for them to make a mistake or b) wait for a crit or other game changing roll to give me an opportunity to put the game away. that's why i played hill giants at the qualifier i attended. hill giants have the speed to take advantage of mistakes, and a good probability of getting a lucky roll that completely changes the game.
quad ravager is another example of a band that generates lots of dice rolls to get those "lucky hits" (although when you're swinging 8 times a round with main hitters, it doesn't take many rounds before a 19 or 20 should show up, not particularly luck just probability).
even the bad side of high variance bands is not as bad for the less skilled player. if your skill is less than your opponents, if you get the same "luck" and you are playing a relatively even matchup, you are probably gonna lose. if you hit a unlucky streak you're gonna lose no matter what, but you were probably gonna lose anyways. a good streak with a band that is low variance (helmed horrors for example) might not win you the game, but a good streak with a band that shines with good luck is much more likely to tip the balance in your favor, allowing you to win the game.
| |
-Doranur
join the revolution! down with round/turn fascism! viva la resistance! | |
|
Orion72
Underboss
1917 Posts
| | 06/26/2006 10:31 AM |
| Another point to explore is just how often dice are rolled in any given game. If I roll a d20 a thousand times, I can expect my rolling average to be around 9-11. If I only roll it ten times, the variance is greater, and I can have an average roll of 7-13, let's say.
So: let's assume a game goes at least 5 rounds. That's 5 initiative rolls. Let's say at least half your warband has to morale check at some point. That's 4 more rolls. You engage in melee by round 2, and have at least 4 combatants. If you lose one each round, that's 4 + 3 + 2 + 1 = 10 attack rolls. Let's say you have to roll Incorp, Conceal, or a DC save at least once in the game. We're looking at 20 different d20 rolls now, and this is easily within the parameters of what you'd expect in a typical game.
So, how much variance is there in 20 different die rolls? Any math guys want to take a shot at this? | | | |
|
Vrecknidj
Warlord
10493 Posts
United States
| | 06/26/2006 12:48 PM |
| Every time you roll a d20, there's a 50% chance you'll roll below an 11. If you roll a 6 on one roll, there's still a 50% chance you'll roll below an 11 on the next roll. If you roll a whole bunch of single digits in a row, there's still a 50% chance you'll roll below an 11 on the next roll.
That's the way I look at it.
On the other hand, I could be a "glass is half full" type and think "the next time I roll I have a 50% chance to get above a 10."
[:)]
Dave | |
Knowledge Arcana editor issues 5-9, Phoenix Lore Magazine editor, assistant editor for Rite Publishing;
My Trade Thread and My Reference Thread; Winner of WBC IV, IX and XIII; Rule #0: bshugg is always right! | |
|
Orion72
Underboss
1917 Posts
| | 06/26/2006 12:55 PM |
| quote:
Originally posted by Vrecknidj
Every time you roll a d20, there's a 50% chance you'll roll below an 11. If you roll a 6 on one roll, there's still a 50% chance you'll roll below an 11 on the next roll. If you roll a whole bunch of single digits in a row, there's still a 50% chance you'll roll below an 11 on the next roll.
That's the way I look at it.
On the other hand, I could be a "glass is half full" type and think "the next time I roll I have a 50% chance to get above a 10."
[:)]
Dave
The law of averages comes into play though, the more rolls you make.
I could roll a 1 on any given roll; I could roll two or even three 1s in a row. But the law of averages makes it unlikely that I'll roll 1s more than once or twice in a 20-roll game, even though I have the same 5% chance of rolling a 1 as I do of rolling any other number. | | | |
|
Blunt Claymore
Sneak
78 Posts
| | 06/26/2006 5:44 PM |
| Nice perspective, Dr. Basilisk. This is the kind of posting that keeps me coming back to the boards.
Your primary hypothesis is that the primium paid for low variance (e.g. high attack bonus, etc.) pieces may not be justified to the extent that this simply pushes the luck factor onto earlier (and as Orion72 implies, more variable since there are fewer of them) factors, such as match-ups (totally out of a players hands) and initiative (which is mostly out of a players hands). That's a totally novel idea and it really made me think about the current assumptions. There are a couple points that challenge the hypothesis, however.
First, most of the winningest bands this qualifier season are low variance. That said, it could be that most of the bands that arrived at the qualifiers were low variance, and in addition that most of the good players are playing low variance bands and what you're seeing is the benefit of being a good player rather than a low variance band. So, these data challenge but do not reputiate your idea.
Second, I don't think it's clear that a high variance band is punished less by bad match-ups or map selection. You assume that they do, but I can image a beholder band getting screwed on Mithril Mines, or dual-Hill Giants getting blanked on Drow Outpost (just some examples that spring to mind -- I don't even know if 2xHGB counts as high variance -- maybe it's just me that hates the morale checks).
I think the main thing is that WotC is tries to cost-out higher attack bonuses, etc., so to the extent they're successful lower cost pieces that roll on the high side of average should win match-ups against low variance pieces, whereas those that roll on the low side of average should lose those match-ups - irrespective of rock-scissor-paper factors. So, if you're high variance player you've got to ask yourself just one question: are you going to roll on the high or the low side of average?
For what it's worth, I would fully support a move in the meta-game toward more high variance bands. I think that would unloose a lot of cool pieces and variety makes it exciting. I'm just going to stick with my Marut for a while longer. [^] | |
A journey of 1,000 miles starts with a good whooping. | |
|
Tried
Sergeant
501 Posts
| | 06/26/2006 7:53 PM |
| Good write-up, Glyn, and right on point I think.
I believe the best bands must have elements of high variance and low/no variance. (Think about the Gith monks).
I think the point about luck averaging out really doesn't address the issue of timing. Once the bad luck hits, sometimes you just can't get enought good luck to recover. Sometimes you just lose due to luck. There is no way to avoid this, so take appropriate steps. Attend multiple qualifiers (if its important to you), or simply play a lot. Dice may even out in the long run, or may not. But one thing is for sure - with good play, your day will come.
Players that win need good dice and good skill. The exact ratio may vary with the tournament, but the general equation holds.
My take on it is this: Keep trying to make better moves, fewer mistakes, and wait for the dice love to hit you.
(Dice love is fickle. sometimes it takes a while.)
| |
Let it be. | |
|
elder_basilisk
Sergeant
410 Posts
| | 06/27/2006 2:50 AM |
| quote:
Originally posted by Blunt Claymore
First, most of the winningest bands this qualifier season are low variance. That said, it could be that most of the bands that arrived at the qualifiers were low variance, and in addition that most of the good players are playing low variance bands and what you're seeing is the benefit of being a good player rather than a low variance band. So, these data challenge but do not reputiate your idea.
I think that Doranur grasped my perspective on this quite well. The big thing about lower variance bands is that they are predictable. As such, you can plan ahead more reliably. In general, I think that better players are players who plan ahead more and are better able to anticipate the course of the game. Thus, the kinds of warbands that got labeled low variance place give more advantage to a skilled player who can better anticipate the results of his moves than a less skilled player. That is, I think why the winningest warbands tended to be lower variance. I think it also has a lot to do with the metagame favoring lower variance warbands at the moment.
quote:
Second, I don't think it's clear that a high variance band is punished less by bad match-ups or map selection. You assume that they do, but I can image a beholder band getting screwed on Mithril Mines, or dual-Hill Giants getting blanked on Drow Outpost (just some examples that spring to mind -- I don't even know if 2xHGB counts as high variance -- maybe it's just me that hates the morale checks).
It's not a matter of higher variance bands being less disadvantaged by bad matchups. Rather, it's a matter of higher variance bands being higher variance. To illustrate, we can imagine the advantage or disadvantage of a matchup being a static modifier: + or - 4 for instance. Now, if we apply that to 2d10, 1d20, or 3d6 (all of which have very similar averages but different distribution curves), you'll find that applying the -4 to the high variance 1d20 roll still leaves you above ten on six out of twenty rolls. That's 30%. On the other hand, if you apply the -4 to 3d6, the percentage of overall results that are still above ten are significantly lower. Bringing it back to DDM, when you have a higher variance band, there are more times that the variance of the band is more significant than the disadvantage of the matchup.
I can imagine beholders getting hosed on mithral mines too, but I can also imagine beholders getting a pair of back to back flesh to stones and winning anyway despite mithral mines. On the other hand, it's a lot harder to imagine helmed horrors prevailing against dual hill giants on dragonshrine. They need a lot more than four rolls to go their way in that situation.
| |
| |
| Knight of the
Round Table
Thenameless
Warlord
12507 Posts
The Fortress of Solitude
| | 06/27/2006 3:45 AM |
| | doranur's reasoning for a less-skilled player to use a high-variance warband is why I chose the HGB and a pair of Orc Champions for my qualifier. I knew going in that I would be facing much stronger players than myself - with the bulk of them playing Couatl/Marut. My only chance was to take a high-speed, heavy-hitting warband and hope for three things. One, win map initiative. Two, roll well on the odd attack. Three, not fail any crucial morale checks. It was a lot to hope for, but it was my only chance to do well. | |
Over 270 successful online DDM trades. | |
|
Wayne
Underboss
1371 Posts
| | 06/27/2006 8:01 AM |
| quote:
Originally posted by elder_basilisk
One player I know has chosen his warband on several occasions so as to minimize variance. His theory is that, by picking units like the helmed horror that have high attack bonuses and strong defenses and immunities to crits, etc, luck should prove to be less of a factor in his games. However, that does not appear to have been the case as he has recently lost several games after a few strings of bad luck.
Just to nitpick, since I agree with most of what you wrote, it doesn't logically follow that because "luck should be less of a factor," and he -- whomever could "he" be?! -- has lost several games due to bad luck, the idea behind low variance is automatically called into question.
Part of the meaning behind "luck being less of a factor" is "luck is a factor less often." If bad luck happens, and is bad enough, it doesn't mean you're wrong to choose low variance ... it just means that you're on the very flat, very unfortunate side of the bell curve. As you point out, Glyn, there's no such thing as a no variance warband.
For example, if a Hill Giant Barbarian is swinging at an Eye of Gruumsh and rolls a 1, he's lost 40 points of damage. If a Helmed Horror rolls a 1, he's lost 15. But if the Helmed Horror rolls three 1s, he's lost 45 points of damage. In other words, it takes bad luck for a HGB to lose 40 points of damage, but it takes extremely bad luck for a Helmed Horror to lose 45 points of damage.
The problem is, as we all well know, extremely bad luck happens. But just because I roll three 1s in a row doesn't logically invalidate the idea that luck should be or almost always is less of a factor when playing Helmed Horrors. | |
Jeff "Wayne Laredo" Wilder | Email | Have/Want List | Trade Policies | Are You an Ethical Trader?
| |
|
hardinjmm
Sergeant
573 Posts
Madison, WI
| | 06/27/2006 2:59 PM |
| Great article, elder-basilisk! This has been said before, but the one variance that you cannot control (although you can predict based on the metagame) is what matchups you face. So, theoretically, a true low-variance band would be low variance in its performance within a single game, and also have a very low number of bad matchups in order to minimize the effects of pairings.
Cheers,
Jon | | | |
|
PerpetualStudent
Sneak
173 Posts
| | 06/27/2006 3:53 PM |
| | While I agree with much of what has been said here, I seem to read a great deal of tournament reports that say the match was decided by one or two rolls. If this is the case, then those are really the only rolls that matter, and there is no way to view this other than luck. The 'I rolled a 20 and critted for the win,' or the even more popular 'I only needed a 3 to hit and win and I rolled a 1.' We see this all the time and there is no way to eliminate this as a variable, minimize perhaps, but not eliminate. Luck will always be a major factor in DDM at least for now. | |
| |
|
Orion72
Underboss
1917 Posts
| | 06/27/2006 4:06 PM |
| quote:
Originally posted by PerpetualStudent
While I agree with much of what has been said here, I seem to read a great deal of tournament reports that say the match was decided by one or two rolls. If this is the case, then those are really the only rolls that matter, and there is no way to view this other than luck. The 'I rolled a 20 and critted for the win,' or the even more popular 'I only needed a 3 to hit and win and I rolled a 1.' We see this all the time and there is no way to eliminate this as a variable, minimize perhaps, but not eliminate. Luck will always be a major factor in DDM at least for now.
I feel it's a false perception that matches are won or lost on a single roll. How many rolls did it take you to reach that one all-important hit or miss? How many decisions? Which ones could have been made differently?
Nearly every game I play has that one deciding moment - a notable one was a recent Epic game. My opponent needed to kill my LSD with his Huge Gold to win on points. He needed 2/3/3 on his attack rolls. He rolled a 2 on the second attack, just missing, and leaving my LSD alive with 15 hp. I won the match on points.
What you aren't reading are the many decisions and attack rolls both of us made to reach that point. There were any number of things he could have done differently that would have turned that one critical roll into just another attack. There were things I could have done differently to keep things from coming to that one critical roll.
I do agree that early bad luck is harder on any band than late bad luck. However, it's also true that late good luck can turn a hopeless situation into an amazing comeback. I've seen it. | | | |
|
Wrackspawn
ChristopherGroves
Warlord
6093 Posts
| | 06/27/2006 4:45 PM |
| I call bands hivh-variance if they rely on a very small set of rolls or overload on one conditional set of circumstances to perform well.
HGBs are high-variance in my mind as they are more susceptible to a small string of poor dice.
Catfolk Wilders would be high variance - only affecting living creatures would allow them to destroy some bands but do very poorly against others.
-------------------------
You can make a HGB band pretty reliable with the Wardrummer. An LSD is pretty reliable as well with excellent saves, etc. This mitigates the high-variance ... two rolls at 10+ is one thing, two rolls needing a 4+ is alot more reliable.
A resilient band is adaptable. The HH with a DMM is more resilient than one without as it has tons of immunities and is able to cope with DR. A Couatl not only adds offense, but also resiliency to a warband.
Anyway, that's what I say | |
Triangle DDM Skirmish Group | My Email | 45-ish trades and counting | Stuff for Trade
* * * Show your brother some love and click here * * * | |
|
Meds
Warrior
286 Posts
Alameda, CA
| | 06/29/2006 1:31 AM |
| Interesting thread. If we define 'low variance' to be minimizing the potential effect of dice rolls in a game (both good and bad luck), then I agree that low variance is beneficial to player with the skill advantage.
I do think we need to consider both sets of dice rolls: both your opponents and your own. I.e. sure, you might have an awesome attack bonus, but if your opponent can kill your titan with a single critical hit from any 1 of 4 hitters, that's a high variance matchup.
The matchup of Frenzied Berserker vs. Justicator raised by elder_basilisk is a good one. Compare it with Frenzied Berserker vs. Blood Ghost Berserker. Both side's attacks hit unless someone rolls a 1. That makes for very predictable play (initiative and critical hits and misses aside) and I think it's lower variance than the Justicator case where the FB has a good chance of missing. Does that mean that a triple BGB band is lower variance than a triple Justicator band? No, I think it just depends on the matchup -- against Marut I suspect the Justicator gives lower variance than the BGB.
In general, there are some attributes that I suspect give you a greater chance of a lower variance matchup. Most have been mentioned already: high attack bonus, low damage, high hit points, fearless, immunities, and low AC.
That last one deserves some comment as 'strong defenses' was one of the attributes listed in the initial post. I think that low AC rather than high AC leads to lower variance matchups. I.e. if you're playing Frenzied Berserkers you can plan ahead as if everything's going to hit. It's the corollary of high attack bonus leading to lower variance matchups. I.e. if increasing your opponent's attack bonus by 8 decreases the effect of dice in the matchup, you can have exactly the same effect by decreasing your own AC by 8. Of course a low AC is not exactly desirable, in the same way that low damage output is not desirable, but I do think it reduces the potential effect of luck in the game.
(On a personal note, the Frenzied Berserker has other attributes that make her fairly high variance against the current metagame: high damage, only medium attack bonus, poor saves. I played FBs in the recent qualifier in part because I knew that luck might save me against the many players far more skillful than me, and I considered HGBs for the same reason.) | |
Have/Want List || Trade References | |
|
| | You are not authorized to post a reply. |
|
| |
ActiveForums 3.7 | You must be signed in to participate in the
games. |