Jump to content


  • Content Count

  • Joined

  • Last visited

About jbrandmeyer

  • Rank

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. There's also a solution that works as a modification to an existing table. Thick insulation foam from a big-box hardware store. Split it down the middle, and add heavy-duty tape along one side of the split so that it can fold up somewhat for storage. Bond black felt to the taped side, and you have a nice wide smooth playing surface. I'm a fan of felt because the models aren't quite so likely to scoot around.
  2. Easier is in the eye of the beholder. My mistakes with math and arithmetic are strongly biased towards bookkeeping errors, so for me any method that cuts down on the bookkeeping gets classified as "easier". One-time re-roll of every die is easy to handle. Since every die can be re-rolled, you can just manage that in the very beginning by assuming that you will re-roll them before rolling them in the first place. For example, black dice go from a distribution of [1, 2, 1]/4 to [1, 10, 5]/16 with Ordnance Experts assuming conservative re-rolling only. Yes, this also implies that OE black dice are mathematically 16-sided. Red dice that you know a priori will be re-rolled are mathematically 64-sided. Harder cases for my method are when you get to re-roll just one die. Or you have two features that allow you to re-roll dice one at a time (say, LTT + an accuracy token). Or you want to evaluate the impact of hard-rolling for a crit result. Or you only needed one accuracy and re-roll the others hoping for more damage. In these cases I switch out to multivariate polynomials and Maxima to manage them. Sloane Phantoms needed `(1/4)*d + (1/8)*d^2 + (1/8)*a + (1/4)*c + 1/4` to describe All The Things (TM) that can happen. I get the appeal of Monte Carlo. Analyzing the new situations just becomes a simple matter of programming. But I get enough of that at work, and wanted to try something new to get closed-form solutions. Octave and Maxima are a little more accessible to non-programmers, so that's also part of why I advocate for them.
  3. Light a man a fire, he'll be warm for the night. Set a man on fire, he'll be warm for a lifetime. Wait, that's not how that goes...
  4. Great shot kid, that was one in a million 136.32! Dice pools in Armada are pretty small. My favorite tool for analyzing their statistics is to use polynomials to describe their probability. For example, the distribution of damage results on one red die is (3/8)*d^0 + (4/8)*d^1 + (1/8)*d^2, where the power of d represents the amount of damage, and each coefficient is the probability of getting that facing. To find out the damage distribution for a pool of dice, just multiply out that polynomial and read off the coefficients. With GNU Octave (a mostly Matlab-compatible free software package), this is equivalent to performing convolution. The `conv` function operates on vectors of coefficients, as it is commonly used for the analysis of signal processing filters. So, lets do that. octave:1> dmg = [3, 4, 1]/8 dmg = 0.37500 0.50000 0.12500 octave:2> dmg2 = conv(dmg, dmg) dmg2 = 0.140625 0.375000 0.343750 0.125000 0.015625 octave:3> dmg4 = conv(dmg2, dmg2) dmg4 = Columns 1 through 6: 0.01977539 0.10546875 0.23730469 0.29296875 0.21630859 0.09765625 Columns 7 through 9: 0.02636719 0.00390625 0.00024414 octave:4> dmg6 = conv(dmg2, dmg4) dmg6 = Columns 1 through 5: 0.0027809143 0.0222473145 0.0797195435 0.1689147949 0.2353477478 Columns 6 through 10: 0.2268676758 0.1550140381 0.0756225586 0.0261497498 0.0062561035 Columns 11 through 13: 0.0009841919 0.0000915527 0.0000038147 So there you have it, the exact distribution of the initial pool of damage on the first throw of 6 red dice. Note that Octave is counting columns on a one-index basis, where we are looking at damage on a zero-index basis. ie, 6 dice give a maximum of 12 damage. There are some sanity checks we can perform. For example, the sum of all those probabilities must equal 1. Also, the chance of getting 12 damage is 1/8^6 and the chance of a complete whiff ought to be (3/8)^6. octave:5> sum(dmg6) ans = 1 octave:6> (1/8)^6 ans = 0.0000038147 octave:7> (3/8)^6 ans = 0.0027809 That checks out. So, what are the odds of getting exactly 9 damage? About one in 160. Better question: What are the odds of getting 9 or more damage? Only marginally better at 1 in 136. If you have a robust tournament of 14 players playing three rounds, every one of which is actually throwing pools that large, every one of which is putting dice on target 3-4 times per game, then you might see one throw that strong the entire day. octave:8> 1/dmg6(10) ans = 159.84 octave:12> 1/sum(dmg6(10:end)) ans = 136.32 This also agrees with Grumbleduke's method, but I find that computing the entire distribution in one shot is a little less error prone. I'll also note that the odds of getting three or less damage from the same initial pool is dishearteningly high at one in 3.6 throws, or roughly once per game depending on engagement and other details. This is why red dice controlling upgrades are so valuable. For more on this topic, search the archives for my posts on the statistics of Sloane Phantoms, and on Blount's Z-95's.
  5. I've played it only once as P2 so far. Nobody else has chosen to take it. The fleet was an MC80S, MC75O, 2x GR75, and 4x A-wings. Its only 6 drops, but being able to start with a dedicated flanker in the middle or on the psyche-out side of the board gives P2 some great opportunities well above and beyond the whale-watching score through deployment shenanigans. Even just one more free drop turns it from "amazing" to "insane."
  6. You can see the designers are running into trouble with adding new game mechanics that aren't broken in the the presence of previously printed upgrade cards and special effects. Salvo: No crit effects. "friendly non-unique squadrons without strategic." The specific list of upgrades associated with Shriv. "without the printed heavy keyword", "using its printed rear battery armament", and so on.
  7. A Rebel MSU list starts off like this: MC30 w/ ordnance experts 2x CR90A w/TRC 1x GR75 w/comm net The CO starts off with Infiltrator 1 on the MC30 (redeploys the flagship after deployment). It out-activates almost anything else in RitR and strongly rewards the player for controlling the time and pace of engagement. Fleet growth depends greatly on which missions you pick, but there are lots of options. SFO & Intel are 8 points together. APT doesn't even require a win. As the first player, its an ISD assassination squad.
  8. However, you can trade the points you were going to spend on FCT on Rogues instead. Rogues attack at the bottom of one turn, and then double-tap on the top of the next with Yavaris.
  9. This is close to a list I run, except that I put the Exodus Fleet on my MC's, for the benefit of the Abandoned Mining Facility objective. The title kindof acts as a soft counter to XI-7. A comm net can feed a token which regenerates a shield and moves two, or discards a hull damage and moves a shield. The Exodus Fleet never dies with its shields still up on any quadrant. I totally get the benefits of Mon Karron + Ion Crit + WBT, though, so you do you. That said, I'd drop Spinals and aux shield team in favor of Engine Techs. Against skilled players, ET can mean one more full turn of shooting. I'm not sure that the XX-9's are worth it on the Bail ship. In my experience, engagement doesn't really begin until after Bail goes off - folks are too busy running away from him. If the enemy's shields are still up, then the XX-9's don't get a chance to go off. Madine fleets are pretty sensitive to things that muck with command dials, so being able to redistribute the XX-9 points on Leia is a good insurance policy. Alternatively, she gives you much more command dial freedom.
  10. You start with 2 in your initial fleet. But if you play the campaign objective then you can get a unique squadron as the location reward.
  11. If you are running a fleet that has some extra squadron commanding ability (say, an Ackbar assault frigate list), then you can build an SFC out of them that works. I've run 3x X-wings and 3X Z-95's pretty well as a punch-forward alternative to Shara, Tycho and 2x A-wings. Its also more punchy than Biggs and 3x X-Wings. For the low, low cost of only 5 more points, you get +1 deployment, +4 hull, and 9 more attack dice on your turn. Where the A-wing crew works around the periphery to slow the enemy down and gum up the works, the XZ crew try to hit and kill one or two linchpin enemies, while sticking around long enough to slow down a non-Intel blob. I personally find that the next upgrade for this particular SFC should be to upgrade an X-Wing to Biggs, and not a Z-95 to Blount. Blount's cost includes his defense tokens, which aren't going to help. By the time the escorts are dead, the wing had better have already done its job. Also, Blount's second re-roll only applies to his neighbors, not to himself. So while Biggs ends up aiding everyone by making the X-wings better escorts, Blount only really aids 2 of his companions.
  12. Mon Karren, Weapon Battery Techs and a Blue Crit upgrade. That has the potential to force some uncomfortable choices on anything flight with a DCO.
  13. 2400 points of fleet on the table is going to be crowded. One BCC simply isn't going to cut it - its just too hard to keep them near the squadrons. Instead of dropping the jamming field flotilla entirely, I'd replace it with a second BCC, and consider giving both of them boosted comms. That makes it much easier for them to contribute from the periphery, where they can stay out of traffic.
  14. If you also have a spare squadron tree, I've seen folks mount squadrons with their flotillas.
  15. Ordnance MC75? Sold. Assault MC80? Consider the Command MC80+Mines. Mart Mattin might pair well with the Fish Farm, defensively. I feel like one mine for 4 points isn't enough. You're going to need to 2x mines or more to edge out the opportunity costs.
  • Create New...