Dicewars - how to win

There's this quick, little strategy game, Dicewars. You have stacks of 0-8 dice on each space and when you attack an adjacent space, you roll yours and they roll theirs and whoever has the highest sum, wins. Then you move all but one into the conquered spot and the enemy dice vanish. At the end of each turn, you get a die randomly placed on one of your <8 stacks (or into your bank if all stacks are 8-high) for each territory making up your largest contiguous bloc. You win when you have every space. On that opening page, you choose how many players there are. I play with five most of the time -- enough to have several players, not so many that the starting position is annoyingly chaotic.

The game presents you with the option to accept or reject a map. You might as well cycle through a bunch just to see the kinds of maps it generates, but my personal status is to always accept what it gives me.

You always play the purple side. Your turn order is randomish (I think it's either random or derivative of the starting board which was randomized). In this case, I was the last player; which I prefer. Note how the board has changed between the image above and the one below. In play, you get to watch the changes, but functionally it's irrelevant. You just want to play the state.

So this is our first opportunity to make meaningful decisions. I like having those two armies on that peninsula. But I'm worried about that large pink stack. My winning approach to this game is pretty conservative. I'd normally like to consolidate my holdings more than this, but I'd rather encourage pink to spend his strength to the south. So I'm just going to try and advance both armies south one space. I don't want to do anything with the army to the south because it's three-strong now which is better than anything surrounding it -- I know the AIs won't attack a superior force, and I also don't want to produce extra territories that my reinforcements will be divided by. So I took those two spaces and passed the turn:

Pink went south, just like I wanted. He swept around toward my southern stack, newly reinforced to 4-high. It was an even 4-v-4 attack, but defenders win ties in the game, so I had a slightly better than even odds of standing and that's how it worked out. Just looking at this, the only way I think I can lose this game is by getting greedy -- so I'm not going to do that. But I do want to connect my forces into a single bloc for greater reinforcement and I'm eyeing the best strategy to dominate the left half of the board. I've decided that my goal for this turn is to move the leftmost five-stack down one and the southern four-stack up and to the left one. And stop my turn there.

(In some cases at this point, I'd make a more aggressive sweep of that area. That tactic would usually pay off, but it's riskier. YMMV.)

Both of my attacks worked as planned. (Note, losing a 5-v-1 is possible but really weird. Losing a 4-v-2 is unlikely, but not so weird that you can count on it.) If I'd lost the attack at the north, I would probably have brought the adjacent 5-stack down, opening up my northern belly to enemies, but in this case I left it there.

So now I'm thinking of consolidation opportunities. I'd like to take pink out from behind me and just have a front facing east, but I don't want to gamble on 3-v-3 attacks the way the AI would. So I'm going to continue with fairly safe and sound baby-steps: my 4-stack will take the pink 2-stack, my 6 will take the west-most of the orange 1s, and by southern-most 4-stack will attempt to take out that green peninsula. This move includes things that could go wrong! But nothing did, and none of the enemy moves much mattered to me:

So now I want to clean up my remaining western front, which should be trivial with my 7-stack, and advance to the east, keeping narrow defiles clogged up with my large units. When I get over there, light green will take some working on, but I'm already three territories bigger and with normal luck, that disparity will only continue to grow. So far, I've been taking my turn and letting the AIs all go before grabbing the next image, but this is one after my moves described above all went well for me:

You see that 3-stack I have on the southern front? I could leave it or attack either of two other spaces. Attacking pink isn't bad because it leaves me with two (plus possible reinforcements) against 1s and 2s. Attacking orange is terrible, because it invites that green 4 into my belly. Unless I took that 4 with my 5. But I could lose that attack and then my front would be mostly undefended unless I got really luck reinforcements. In the end, I choose to hold as pictured. Slow and steady wins the race.

My reinforcements went well, and all four other players just passed and collected. I continue to be worried about light green. I really don't want to get one of my armies up against that light green 7 just yet. I could pass my turn and just collect. That wouldn't be stupid. I could attack that dark green 5, and if I win, take the other dark green, the nearby orange and the first of the pink. That's a solid plan. But if I lose that first attack, I'm weakened and give dark green a new lease on life. So I'm going to invite dark green in, continuing a very conservative march. My 5-stack on the south is going to take the first of the pink spaces and my 6-stack is taking orange. And then I'm going to let the others go. Dark green will probably advance two or three spaces into my interior. That'll look like a failure, but I'll just swallow him the next turn.

It turns out that that AI rolled poorly and only took one space. So my goal is to eliminate that player by taking all three spaces and leaving pink as a buffer between me and that light green 8-stack. I'll attack with my interior 6, hopefully taking two spaces with it and then my forward 6 will take the final dark green territory.

That worked like a charm. Pink's going to grow one die each turn and the light green AI isn't willing to attack them because they've decided I'm the only serious enemy. That's dumb, but common. It means that I can just roll a couple turns forward collecting more reinforcements than my enemies.

So I did that four times, until pink filled up to 8:

So, some things to note about the board: each player is at full strength and so reinforcements have started going into our bank. The bank can hold something like 40 units. Also, because pink is at 8, they'll attack me this turn. I could just let them -- the attacker has a slight disadvantage because of losing ties. But if you look at our faction strengths at the bottom, I have 17 to green's 7, and not only that, but I'm earning more than half of all the reinforcements on the whole board. It's possible to lose at this point through bad luck, but not plausible. Even a bad series of rolls comes to an end and I have enough buffer to last through anything in the realm of likelihood. Also, since my "income" is 17, and a full stack is 8, I make enough to completely replace two stacks. So I should be attacking twice per turn any time I can.

So, I'll go ahead and attack pink. I did it, took them out on the first roll, and ended my turn, building up a little bank. Then green took that space from me on their first attack. Many maps have chokepoints like this one. It can be a little bit of a slog if there's a one-wide isthmus several spaces long. But this won't be too bad. I tried and failed. He tried to move into my side and failed. Then I succeeded and only that, but pressed forward another one even though my roll was 7-v-8:

Notice that I had a choice of three spaces to attack. It was an easy choice, though. If I hold that spot (unlikely for now), I reduced green's income by three, not just one. Another consideration is exposure to enemy attacks. If I wanted to minimize that, I could have gone straight north. If I wanted to maximize that, I could have gone straight east. I do want to maximize it, but the chance to take three income was more powerful. Plus, they still have four attacks on me.

So why did I wan green to have more attack opportunities? It's because they have a slightly greater than even chance of losing each one, and each one requires reinforcements to build back up. If they have a fat bank, I'll need to do that for several turns, milking them down before I can get around to really beating them.

So I let things go and they took both spaces back, then I took one, then they took it back, then I took the same two again and they took the one back. But see this:

Their bank ran out, just like I explained, and the couldn't fully reinforce. So starting now, I'll have much softer targets. So I took the soft space and tried to take another, but predictably failed. Then they took that one space back, and clogged up my one attack opportunity with an 8-stack. But their middle is soft and getting softer:

So now a couple turns have passed. On each of my turns, I take two of them and then they take one back.

You can see that both they and I are split , but I can push three 8-v-8 attacks at the plug and if I capture it early on, I can go ahead and hammer their weak-bits. Note, I don't really want to capture the one green territory that's in contact with orange. I'll just let them hold orange away and once I'm done with green, they'll be a trivial cleanup operation.

Sadly, it took three turns to take that one space out and then I decided to go ahead and try to force the issue. If I can hold what used to be their center, limiting their reinforcements to 1 or 2, then I should knock them out right away.

The rest of the game was boring. I captured two greens, they took one back, I captured their entire south with no resistance. Then I knocked out their north. Orange had a big bank at that point so they were able to keep their two spaces at 8-stacks while sallying temporarily into my territory a few times. But after five turns or so of fighting with only 2-3 income, they ran out and I just sopped them up.

I've had the idea to write something like this up for a while now -- talking about the revelations I've had while learning how to beat this. As much as anything to see if I could translate what I do to words that could teach someone else. Hopefully if anyone is interested enough to try it, I will have succeeded.

Please let me know what you think, or if you know the game, or if you try my strategy -- and any problems you have!