Search Unity

  1. We are migrating the Unity Forums to Unity Discussions. On July 12, the Unity Forums will become read-only. On July 15, Unity Discussions will become read-only until July 18, when the new design and the migrated forum contents will go live. Read our full announcement for more information and let us know if you have any questions.

RTS mechanics

Discussion in 'Game Design' started by tiggus, Oct 21, 2014.

  1. tiggus

    tiggus

    Joined:
    Sep 2, 2010
    Posts:
    1,240
    I'm curious what people think about the following ideas related to multiplayer RTS.

    1) Static or dynamic placement of resources on maps? Pros/cons? I tend towards static since it lets players build up strategies over time based on the map and removes the "luck" factor as to whether you get a lucky spawn nearby.

    2) Sharing bases as opposed to separate base for each player when playing teams. I don't see too many games that let you share common bases if you are on the same team, not sure if there is a reason for it outside of wanting to spread the players out but I like the concept of a "home base" where all the players on same side group at.

    I have some others but these are the two on my mind lately.
     
  2. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Well, your point 1 may be worth a thread all its own. Luck is a great equalizer, and casual players really like that they can jump in, and maybe even have some slim hope of winning, in games where luck is an important factor. Serious gamers, of course, quickly get bored with games that are dominated by luck. It's the whole chess vs. Monopoly thing.

    The best games, I think, carefully balance both luck and strategy, and it seems to me that random resource placement might do that. Randomness also improves replayability, and offers the reward of discovery (especially if the resources are not immediately obvious, but only revealed through exploration).

    As for point 2, I haven't thought about it much. But off the cuff, what I think would add interesting depth for teams is if each member of the team can somehow specialize. So, let's say everyone on the team can share buildings, but what buildings you can build, or how well you built them, relates to what you've built before. So in this case, good teams would naturally specialize: you build all the combat-related buildings, and I'll build the research-related ones. We have to work together and cooperate for maximum effectiveness.

    Just a thought.
     
    theANMATOR2b and NomadKing like this.
  3. tiggus

    tiggus

    Joined:
    Sep 2, 2010
    Posts:
    1,240
    Good points!

    I think a good compromise is allowing the randomness to be set on the match start screen ahead of time to accommodate both types of gamers.

    I like your idea on sharing the buildings, I hadn't considered that option. So if I built a tech 1 facility and it opens up some advanced research buildings I can focus on those, while my teammate built a barracks 1 and is working on the military unit production.
     
  4. JamesLeeNZ

    JamesLeeNZ

    Joined:
    Nov 15, 2011
    Posts:
    5,616
    1. definitely static. you can throw some dynamic bonuses around though if you wanna mix it up. The main point of static primary resources is fairness. If its dynamic, there's a chance that I could generate a significant advantage early in the game by generating more money (or whatever) quicker, and throwing units against the enemy to stall them further.

    2. I like the idea of a shared base.. however in my game each player will have their own resources to spend. while there is no need for building construction, you could expand to allow it by letting players contribute funds to having a building built/upgraded.
     
  5. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Just for the sake of argument, you could do random placement of resources, and still make it fair, by simply making sure the map is symmetrical in some way (that includes starting positions for the players).

    But of course, this would also mean that as you explore your part of the map, you're learning about the opposite side of the map too, which is weird and takes away some of the fun.

    So, never mind, this was a lousy idea. (Now archived for all eternity.)
     
    theANMATOR2b and Venryx like this.
  6. PJRM

    PJRM

    Joined:
    Mar 4, 2013
    Posts:
    303
    1. I love dynamic resources. doesn't matter what! I like to think of a challenge. You never know what will happen next or what you can find.
    2. You're right. i don't see many games sharing same entities (Building, bases, etc...) but i like the idea, thought it makes a team more vulnerable! Destroy that and the entire team fall! but share base allow the team to evolve faster.
    Best regards,
    PJRM.
     
  7. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Sure, I agree with you on both points. The random luck of initial resource placement is just a leveler — a good player will still beat a newbie, but sometimes he'll have to work a little harder at it, and that's fun for both of them.
     
  8. JamesLeeNZ

    JamesLeeNZ

    Joined:
    Nov 15, 2011
    Posts:
    5,616
    Youll never get the hardcore gamers interested if theres too much random chance to get screwed from the get go.

    hardcore gamers are teh ones you want playing your rts.
     
  9. tiggus

    tiggus

    Joined:
    Sep 2, 2010
    Posts:
    1,240
    Yeah that's why I was thinking having it as an option makes the most sense. It can default to static but hey if you want random resource spawns just set that in the game lobby.
     
  10. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Are they? I'm not so sure.

    Fighting games went through a boom in the 90s, where hardcore players demanded more and more complexity, and game developers delivered, with each new game harder to master than the last. The result? Their market shrunk and shrunk. Now fighting games are pretty much a niche genre, which a small (hard) core continues to play, but they've become almost irrelevant in the mass market... except for Super Smash Bros, which hardcore gamers disparage as a "kiddie brawler" but which has made more money than all the other fighting games in recent years combined.

    So I dunno... maybe this merits its own thread. But I don't think it should be taken as a given that you should cater to the hardcore players in your game design. That's a serious choice with potentially big consequences.

    But of course, on this particular issue, @tiggus is quite right: it's easy to have both fixed and random maps, so why not have it both ways?
     
    theANMATOR2b likes this.
  11. JamesLeeNZ

    JamesLeeNZ

    Joined:
    Nov 15, 2011
    Posts:
    5,616
    if you want your rts to be huge, hardcore gamers are the ones you want.. of course this is really only desktop gamers.. obviously doesnt apply to all genres, but I think rts is one that warrants it. Often rts gamers are looking for a bit of complexity.

    It can go either way like you say though. Too hard and it will put off casual gamers which is obviously a bad thing.

    You never see random resources in the biggest RTS games. Of course the only big rts that even comes to mind is starcraft.
     
    theANMATOR2b likes this.
  12. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Yeah... it sounds a lot like the arguments that were made for more-complex fighting games to me. "Fighting games need complex combos to have appeal." "How do you know?" "Well, look, all the biggest fighting games have complex combos."

    But I'll admit that the RTS genre might be different. It's fundamentally a strategy game, which may put off most casual gamers right off the bat. Fighting games naturally lend themselves to more casual button-mashing, though it took Super Smash Bros to buck the standard control schemes and design one that would really have wide appeal (see this essay for more).

    On the other hand, I was at a neighborhood gathering a couple months ago, and the kids (ages 9-15 or so) were all clustered around their Nintendo DS's playing LEGO Battles. So the idea that an RTS is only going to interest serious gamers may not hold up.
     
  13. JamesLeeNZ

    JamesLeeNZ

    Joined:
    Nov 15, 2011
    Posts:
    5,616
    Wasnt reallly saying rts should only cater to hardcore gamers... of course it can cater to either side... the best games cater to both..
     
  14. IroncladEarl

    IroncladEarl

    Joined:
    Jul 3, 2013
    Posts:
    87
    I'm no expert in game design but I think the luck factor can be great for competitive players as it forces them to adapt their strategies on the fly. To me that would be true skill and not just reflex from playing the same map for hours on end.
     
  15. Zaladur

    Zaladur

    Joined:
    Oct 20, 2012
    Posts:
    392
    RTS generally should cater to
    Small amounts of randomness, maybe. But in very competitive games, where the level of skill of top players is very close, even a little luck can decide a game.

    Starcraft 2 is a good example of an extremely competitive RTS. Aside from spawn locations on some maps, randomness is kept to a minimum.

    I'll agree that competitive isn't necessarily needed for a successful RTS though. Pikmin 3 is a wildly fun example of an RTS that casuals play, without even realizing they are playing an RTS. Spawn locations and cherry roulette often decides games, and yet I keep coming back for more. And I'm a very competitive person.
     
  16. Gigiwoo

    Gigiwoo

    Joined:
    Mar 16, 2011
    Posts:
    2,981
    Are there many examples of well designed mobile RTS's? I've played some good turned based ones. Can't think of many RTS though... Wonder if the design struggles cause of touch, lack of a keyboard/arrow keys, or just the small screen.

    Gigi
     
  17. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    I played StarFront on the iPad and thought it was well done. It adapted to the touch screen very nicely.

    (Man, what I wouldn't give to have the original Warcraft on the iPad!)
     
    Gigiwoo likes this.
  18. tiggus

    tiggus

    Joined:
    Sep 2, 2010
    Posts:
    1,240
    I would be curious to know this as well, haven't really found much. I think you would really have to limit total number of units significantly compared to your traditional RTS though, too much going on for mobile. One way to help with this would be the Planetary Annihilation approach where you host the simulation in the cloud and clients are basically dumb renderers. But that gets expensive for bandwidth and hosting costs.
     
  19. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Pshaw. Mobile phones today are dramatically more powerful than what Warcraft ran on back in the day. Granted, Warcraft's pathfinding was a bit weak, but otherwise it's not significantly different from modern offerings except in glitz & glitter.

    I do agree that mobile phones are a bit small for it though. I played StarFront on mine, and it was OK, but it was much better on the tablet.
     
  20. tiggus

    tiggus

    Joined:
    Sep 2, 2010
    Posts:
    1,240
    You can get 100+ units moving around without lag on your ipad with Unity? I have tried and failed pretty bad so far. That was with no colliders or physics.
     
  21. DallonF

    DallonF

    Joined:
    Nov 12, 2009
    Posts:
    620
    Ooh boy, I love this subject. Personally, while I love the RTS genre in theory, I think there's a lot wrong with them right now, especially in terms of accessibility.

    I'll share some of my notes:

    So imagine I'm a beginner Starcraft player. (in theory they exist, but I've never gone up against one...) I do everything I've learned from the campaign or from skirmishes: I build up a base, build a bunch of my favorite units, then when I feel I have enough troops, I send them at the enemy base. And then they all get slaughtered because my opponent has more troops than I do. Why? I dunno. I just don't have enough 1337 skillz I guess. My opponent doesn't seem to mind, he keeps doing his thing even though at this point I have zero chance of winning. But I'm a naive noob and I don't know that, so I rebuild my army in hopes of trying again... then eventually he marches into my base with a bigger army than I could ever dream of and I lose.

    This is pretty much every player's first online experience with the game. And while there's nothing wrong with losing your first few online matches, there's a lot of things here that will immediately turn away all but a hardcore gamer who enjoys pain:

    • The losing player has no idea why they lost. He did as well as he could but his opponent just plain did better. The newbie has no idea where to go next, or what to do to increase his chances of winning next time. This, I think, is why MOBAs have pretty much replaced the RTS as the strategic competitive game of choice for so many players: it's easy to see why you lost. Maybe you got outleveled, or maybe you Leeroy Jenkins'd one too many times and got yourself killed at a time when your team needed you most.
    • Arguably, commanding an army is a lot more fun than building one. But skill in commanding an army rarely decides a game - it really acts as a tie breaker for your economic skill. At high levels, both player's armies are roughly the same size (they've maxed out their economic skill), so you have a really interesting tactical battle. At low levels, though, it would be functionally equivalent (actually slightly more fun) to simply end the game after a timer and declare whoever had the biggest army to be the victor. You're basically playing a SimCity speedrun.
    • The game ended long after it was decided, leaving the newbie feeling like a dead horse being kicked repeatedly. (that analogy kind of got away from me) A well-designed game should either cut it off almost as soon as one player gains a runaway lead, or alternatively, throw more obstacles in the leading player's path to make it require more skill to stay in the lead than to take the lead.
    • The real secret to the first point: the newbie's opponent was probably using a memorized build order to build faster than him. To compete, he would have to go memorize one himself and perfect his muscle memory before real strategy becomes a factor in the match. That would almost be OK, except there's nothing in-game indicating that this is the case. He would have to be the sort of person who at least lurks on game forums and wikis - which is only a small fraction of the playerbase.
    Don't get me wrong: when you go through all the hard work to get good at an RTS, it's a blast. But I've come to a time in my own life where I simply don't have time to do homework to get good at a game. (Heck, I play games to avoid doing homework!) And I can't fathom why this must be the case. Surely a gentler learning curve is possible without sacrificing much depth.

    Oh wait, you asked questions. Here are some answers:
    1) This debate tends to gets players all up in arms - with statically placed resources, hardcore players have the opportunity to do lots of calculations (not fun) to optimize their play, which forces other players to either repeat or copy those calculations (also not fun) to compete. But with randomized resources, you run the very real risk of letting a game be decided by luck and spawn location. In fact, it's not even a risk. It's going to happen, because of the heavy economic focus of the RTS genre. But one suggestion I've never seen is symmetrical randomness. In other words, what if each player was guaranteed to get the same resources at the start of the game, but none of them know what those resources are going to be? Food for thought. Maybe there's a reason I've never seen it in a game before.

    2) It's a novel idea and it would certainly help your game stand out. My only concern would be the potential for griefing - one player gets in the base and just starts messing everything up. Players like to have some breathing room.
     
    theANMATOR2b and JoeStrout like this.
  22. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    You raise some really good points there. So, what's the solution? How would you make an RTS game that is (1) more about skill at war and less about minimaxing the economics, and (2) clearer what you need to do differently when you lose?

    Here's one idea: nix the economics entirely. Either you start off with all your units, or a fixed amount of resources with which to build units, or both players get resources at some steady rate that's outside their control. (All three of these lend themselves quite nicely to handicapping: just give the noob more and the veteran less.)

    This certainly breaks the mold Warcraft established, and some would argue that it's no longer an RTS at all. But it might be interesting, especially if you can somehow emphasize the strategy — make it effective to organize your troops in certain formations, attack and defend key points, flanking, all that jazz.

    Or, keep the economics, but somehow prevent minimaxing... random resources might help, but I'm not sure it's enough. What if you also randomized the various costs (within certain ranges), so each game was different, and you'd have no time to bust out the spreadsheet? Probably a lousy idea, but I'm brainstorming here.

    And here's yet another idea: what if you nix the warfare, and really do make it all about the economics? Since, as you point out, that's generally what decides games (except between two experienced minimaxers) anyway, what if you really did just define a game around an economic victory, and make it all about that? Civ has this as one of its victory modes (or you can go for a technological victory, which amounts to the same thing, since tech and economics are closely intertwined). But you'd face the same issue of, how do you prevent someone "solving" the economic game and executing perfect play every time?

    You could do like Civ does, and have a random map. People will object that games are then decided by luck. But I don't think that has to be true. It should be instead decided by your skill in taking advantage of what the map gives you. For this to work, there needs to be more than one economic path you could take; it needs to be a tree with many possible paths.

    Or, perhaps like Chess and other two-player games, the economies of the players need to be intertwined, so that your best move depends on what the other guy is doing. This would make it so you can't simply "solve" the game ahead of time, but must instead counter your opponent's moves.

    Maybe the fundamental problem with RTS games is that (1) the economy is all-important, but (2) the economic system is too simple, and (3) the economy of each player is basically separate. This leads to people solving the economic game (which as you point out, is boring) and then clobbering anyone who hasn't done the same. So to fix this, we need to either change (1) so the economy isn't so crucial, or change (2) and (3) so it's not solvable.

    What do you think?
     
    theANMATOR2b likes this.
  23. DallonF

    DallonF

    Joined:
    Nov 12, 2009
    Posts:
    620
    This typically is known as a "Real Time Tactics" game. There have been a few in the past, but none of them have really caught on for some reason. My theory: even though it's more fun to command an army than to build one, it's much more fun to command an army if you have built it.

    I think it's a good idea to build RTT, though, since the genre isn't inherently broken in its current form. Makes it far easier to design! But I don't think it's the Holy Grail we're looking for.

    I think this is the essence of what you have to do. I don't know about randomizing costs - that seems unintuitive. But perhaps a very, very complicated cost system that depends on factors in the map and your own actions. Hey, that's an idea: maybe instead of getting more money over the course of the game, you make your units cheaper. But a predictable, mathematical system can be calculated, so we're right back to square one.

    The trick here is also to avoid alienating hardcore players. Because without the hardcore audience, the community falls apart. Newbies feel there's no goal to shoot for, like there's no point playing the game. And hardcore players despise randomness - at least that's what they'll tell you. I think what they mean is that they despise the possibility for randomness to ever decide a match rather than skill (i.e. a player of less skill wins against a player of greater skill).

    I'll return to this topic in a moment...

    You're coming close to one of my conclusions: you don't need a random number generator to make a game random. Players themselves produce randomness in a game, especially when you force them to make uninformed decisions. Think about Rock Paper Scissors - it's essentially a random game, even though no dice are rolled and no cards are drawn.

    So what you really need to do is allow players to just mess with each other. Nothing ruins (in a good way) a solved game faster than the other player doing something you don't expect. But here's the thing with Chess: it's a mostly solved game. There are memorized starts and - amazingly - memorized endings. Players don't usually behave randomly because they know what the optimal move is in that situation! And because the other player knows that optimal move, they know the optimal counter to that move as well. The solutions I've seen are:
    • Add so many game elements that it's impossible to memorize all the situations that could occur. This is pretty tricky to design - usually what you'll wind up doing is creating imbalanced elements that just confuse the game even more for new players, while the hardcore players know that the game really boils down to to a smaller subset of elements that are practical and optimal to use.
    • Hide information from players so they don't have all the facts. In theory, RTS games already do this with fog of war, but it really doesn't change anything because there's very little player interaction in the first few minutes of the game, which is when it really matters.

    Yep, I think we're on the same page here. I think my specific goals would be something like this:
    • Make the economy grow linearly, not exponentially. This keeps the tactical part of the game relevant. (because the difference between 10 and 11 troops still allow for tactical skill to decide the match, as opposed to the usual 10 v 30 type of situations you see)
    • Randomize the starting conditions of the game, but make sure it is still symmetrical.
    • Add an element to the game that behaves essentially randomly, but discriminates predictably: it always makes life difficult for the player that's currently in the lead. Basically like a Blue Shell from Mario Kart (except it should be much more frequent and much more avoidable)
    • A few months ago, I would have said "Encourage players to harass each other in the early game", but I don't know if that's possible now. New players, for some reason, seem to have an implicit "don't attack until I'm ready" rule and I'm not sure if you can actually do anything about that.
    Also, I highly recommend reading these two articles, both by the same guy (who designed some Street Fighter games, and, more interestingly to me, a "sequel" to Chess which addresses its solvability):
     
    theANMATOR2b and Gigiwoo like this.
  24. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Those seem like good ideas. I would still encourage thinking of ways to link the players' economies, so that right from the get-go, your best move depends on what the other players are doing. Maybe it's like a trading game, where the cost of stuff you need (or value of stuff you want to sell) depends on supply and demand. So if the other guy is buying up all the Whoozits, the price of Whoozits goes up, but Whatsits are still cheap, so maybe you should build Whatsit plants instead. Or something like that.

    In standard RTS, the only linking between the economies is that there might be some practical limit on certain resource types (a limited amount of trees for lumber for example). But this generally doesn't matter till the end game, by which point it's too late.

    Oh, and this addresses your "harass each other in the early game" idea. I agree that noobs (in particular) are not likely to attack right away, but I bet they'd very quickly learn to harass the other guys economically. If anybody here is old enough to remember M.U.L.E., they'll know exactly what I mean!

    As for ways to make it harder for the lead player to hold onto the lead, in a game of 3 or more players, you might do some sort of king-of-the-hill mechanic, where it's in everyone's best interest to gang up on the lead guy. For example, the board game King of Tokyo has this; only one player can be in Tokyo, and there are reasons to be there (victory points, plus you get to attack all other players at once), but everybody else in the game is attacking you until you yield.

    So, maybe you could do something similar in an RTS game... while you control the Orb of Power, then you get some important benefit towards winning, but all other units attack only the Orb of Power player's units, and not each other.

    I'm going to read those articles this weekend, they sound really interesting!
     
  25. BFGames

    BFGames

    Joined:
    Oct 2, 2012
    Posts:
    1,543
    You guys might find this paper interesting: https://www.itu.dk/~yannakakis/cig10StarCraft.pdf

    It is about procedural generated maps in star-craft. It is written by one of my teachers, and he mentioned in class that even though it was implemented to create maps with a fair set up of resources, players mostly did not like it, because it is such an important part of the game for many that they know exactly how the map is set up.
     
  26. Gigiwoo

    Gigiwoo

    Joined:
    Mar 16, 2011
    Posts:
    2,981
  27. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Agreed, those were a great read. I do think there is another pitfall not addressed in the slippery-slope article, though: that's a game that drags on and on and never seems to end. This can be a real danger when you have negative feedback (what he calls "perpetual comeback"), unless there is some limited resource (turns, health, whatever) that gets steadily used up over the course of the game, and can't be replenished.
     
    Gigiwoo likes this.
  28. Gigiwoo

    Gigiwoo

    Joined:
    Mar 16, 2011
    Posts:
    2,981
    League of Legends is a pretty good case study. It's a positive feedback loop, and yet sometimes, teams can come back from MASSIVE leads (10k+ gold deficits), if they capture strategic objectives or stall long enough. Around the 50+ minute mark, the death timers grow so massive that a long, drawn out stalemate will be decided by a single battle. Very clever design!

    Gigi
     
    theANMATOR2b likes this.
  29. DallonF

    DallonF

    Joined:
    Nov 12, 2009
    Posts:
    620
    One thing that I think helps with the "perpetual comeback, perpetual game" problem is a non-reversible win condition.

    Example: in Mario Kart, you're always moving forward. Even if you get hit with a blue shell (the classic "perpetual
    comeback" solution, although personally I would prefer it to be more avoidable and more frequent than its current incarnation), you're further along the track than you were a minute ago. The game is going to end when the second-to-last player crosses the finish line, and there's nothing you can do to stop them from doing so - only get there first.

    Another example: many modern board games have a "victory point" system, where the first player to a set number of victory points wins, and once you earn a victory point, you can never lose it. (Another common variant is that you can lose a victory point to another player, and the game ends after a set turn limit)

    Actually, the non-reversible win condition helps for "slippery slopes" as well, if only to make sure that the game ends quickly once one player pulls ahead.

    Compare to a typical RTS, where the game ends when one player's base is completely destroyed. Except a player can rebuild their base! The only reason this works is because of the insane "slippery slope" effect at play, where the first player to lose a unit (or worse, an economic structure) is at an irrecoverable economic disadvantage unless the other player does something stupid. If you were to fix that and make it possible to effectively rebuild your base while fending off attacks, games might go on forever!

    I wonder what a different win condition might look like in an RTS?
     
    theANMATOR2b likes this.
  30. JoeStrout

    JoeStrout

    Joined:
    Jan 14, 2011
    Posts:
    9,860
    Yes, exactly, this is the "limited resource" I was (poorly) referring to above. Whether it's distance to the finish, or turns until the end, or victory points remaining, there's something that steadily dwindles (or goes up) until a clear finish line.

    So, that's a good question: what might that be in an RTS? Just brainstorming here:
    • victory points, gained by killing units or capturing (non-recapturable) flags or some such
    • limited time (equivalent to turns), at which point you just see who's got the most units/money/widgets/whatever
    • key items that get destroyed; first to destroy all the enemy's thingies wins (but I guess this is really the same as capturing flags)
    Hmm... I seem to be stuck in a rut here. Anybody have any other ideas?
     
  31. DallonF

    DallonF

    Joined:
    Nov 12, 2009
    Posts:
    620
    Yeah, that's pretty much it. Unfortunately, a non-reversible win condition tends to be a bit anticlimactic. Then again, so is destroying the opponent's last random building that you couldn't find. The best climactic ending I've seen in an RTS is Supreme Commander: every game ends with a nuclear explosion! (You lose when your special "Commander" unit is destroyed, and he explodes spectacularly) The downside to that system is that it leads to lots of sniping tactics (ignore everything, just focus all firepower on the Commander, who's very vulnerable in the late game), to which the only counter is turtling (a classic tactic that ruins the game because its aim is to reduce player interaction).

    It's a tough problem to be sure. Although, what if you had to destroy three Commanders? And what if each player had five to start with? That would be interesting indeed.