Farragar.com

Testing web crawler.

• Optimisation implemented. Efficiency improves, but surprisingly inaccurate… O(a((N/a)^2)
• Overhaul of XMLModelFile.xml, including layers, functions, agent names & variables
• Global variables (delta T, min interraction radius, dampening factor, etc…) can now be set at run-time
• Design section writing

• Revelation! Reduce size of timesteps, then stack them such that we only recalculate positions for N/x particles per iteration, with x the number of stacks. Vastly improve framerate for sacrificing accuracy.
• Extended circle distribution to 3D – now can spawn spheres
• Sphere with a 2D plane – looks like spacetime! link
• A bunch of spheres with a plane: youtube
• A run with 3 spheres in an even distribution, vels at t0=0, so dwarf galaxies collapse/expand for a while. They later merge into one…youtube
• Experiments with a velocity dampener, as well as minimum interaction radius

• First half of week preparing for talk/meeting on Thursday
• Thursday – Flame meeting! All day
• Friday -> Monday – Mostly refactoring of script generator
• We now define a Simulation comprised of many particle distributions, each distribution has:
  • Mass,
  • isDark,
  • positions(x,y,z),
  • velocities(x,y,z)
  • each as individual probability distributions
  • An arbitrary number of particle distributions can now be initialised with high complexity, and all stuck into one simulation

• CUDA and FLAMEGPU installed
• Getting used to Visual Studio and working with FLAMEGPU
• Simple Particle XML description implemented
• O(N^2) Gravity resolution working!
• Simple Python program to generate 0.xml files with very basic parameters
• Several experiments, videos:
  • 30 Particles in 2D
  • 3 Particles in 2D
  • 30 Gravitational Collapse of 3000 particles distributed in a cube, with random seeded masses.
  • 99 particles with a mass of 1, 1 particle with a mass of 100

• Feedback received. Report returned
• Edits to report as a result of feedback:
  • Necessary Requirements of software packages – FLAME, FLAMEGPU, SWARM, GADGET.
  • Implications of using FLAMEGPU on the poster session
• Visual Studio setup & Instalation
• FLAMEGPU technical guide
• Partial XML construction of basic particle mass agent

Recently, Heroes of Newerth went Free To Play, adopting the League of Legends inspired payment model. I wrote this post today to outline why it causes irreperable balance issues and why, in effect, I’m very sad to be leaving.

Asymmetry with regards to balance, game-instances and playing to win.

A game is defined to be, ‘A form of play or sport, esp. a competitive one played according to rules and decided by skill, strength, or luck.’

This article assumes that both you and your opponent are playing to win, which I define to be picking (what you believe to be) the optimum strategy, given all available information.

The concept of playing to win is essential – if both players are playing to win, then the outcome of the game is meaningful; we can derive a conclusion that the winner was higher in some quantified measure or combination of skill, strength and luck. This is the great thing about winning games, and especially those with a higher skill component. By playing a game, you are essentially abstracting your skill into some more easily measured form, and then mathematically demonstrating your superiority.

Now, if one person has an unfair advantage over the other player, then any conclusions that could previously have been drawn from the outcome of that game are immediately null and void. To rephrase, if a game is not equal/fair/balanced/whatever-you-want-to-call-it, then it is not a valid measure of whatever it is testing, and if it is not a valid measure of what it is testing, then the results potentially don’t reflect the actual situation.

Asymmetry causes balance issues in games. A game is said to be asymmetric if one player has different knowledge or a different set of options to another player. A single game-instance of Starcraft is a perfect example of an asymmetric game – if one player is Terran, and another is Protoss, then the game is, by its very nature, imbalanced; both players have access to completely different units, options and play-styles. However, Starcraft has been very well balanced, and huge amounts of work have been put into making sure the three races have no real advantages or disadvantages over the others. (And indeed, the ‘mirror’ matchups where both players are the same race, are symmetric and perfectly balanced, with the exception of map spawn positions, which are negligible in well-designed maps). It is important to note that Starcraft as a whole is symmetric – all players have access to all races. It is only the game-instance which is asymetric, since races are now fixed. This is also true for all good asymmetric games; those which are played at a competitive level. While they can never be a perfectly balanced abstraction of skill in the same way a symmetric game is, they are often so close as to be considered so. Examples of asymmetric game-instances would be all RTSs with multiple races and Heroes of Newerth, among others. Obviously, if a game-instance is symmetric then it is balanced by its very nature; some examples of symmetric game-instances could be Quake, Counter-Strike or Chess.

Purchasable, stateful content with regard to Heroes of Newerth, how this causes extreme-asymmetry and destroys balance, refutations to common arguments and the future competitive viability of this model.

Heroes of Newerth has recently adopted a Free-To-Play model, with a rotating pool of freely available heroes, where players are required to purchase access to others before they can be played. This is a stateful transaction; buying a hero with real money gives you the permanent option to play that hero.

This absolutely and unequivocally ruins any semblance of game balance, by giving those with a real-world advantage (namely money), an in-game advantage, by way of having access to additional heroes. This isn’t the sort of slight asymmetry acceptable in a game of this nature; new accounts only have access to fifteen heroes, of which there are currently seventy eight.

Before this patch the game of Heroes of Newerth with symmetric (as all players have access to all heroes), and the game-instance was asymmetric (but well balanced). Now, not only is the game-instance asymmetric, but the game as a whole is asymmetric – only those people which have bought heroes are able to use them.

Common arguments: The following are refutations to some arguments which I’ve read in defence of this model, usually related to League of Legends.

• The ‘Each hero is individually balanced so I can’t use money to buy an overpowered hero, therefore I can’t buy an advantage‘ argument.

This one is just ridiculous, in Heroes of Newerth there are several heroes which are practically required to play against others. Let’s say someone on the other team picks Tempest. In order to counter Tempest, a hero such as Hellbringer, Andromeda, Pharaoh or Vindicator is generally a good idea; otherwise you’re going to have a really hard time. Imagine you haven’t bought any of these heroes, you’re now royally fucked, directly because you hadn’t purchased one of the appropriate heroes beforehand in the stateful, unrelated, out-of-game system.

Not only that, but imagine you’re playing Quake. The opponent has the choice of three weapons. Gun A does 2 damage per hit, firing every second; gun B does 4 damage per hit, firing every two seconds; and gun C does 6 damage per hit, firing every three seconds. These are all individually balanced: they all do the same damage per second. However, you haven’t bought B or C, so you only have the option of using A. Regardless of the fact that you’re both doing the same amount of damage, your opponent has an inherent advantage in having choices available to him – he has a wider range of tactical options. To claim otherwise is simply incorrect.

• The ‘Every hero can be unlocked just by investing time into the game anyway, so it’s not imbalanced‘ argument.

This argument doesn’t really make any sense. Firstly, it’s not logistically feasible to unlock every hero without paying. At a current total cost of 25620 silver coins, you’d have to play 3200 games winning an average of 8 coins per game. At an average of thirty-five minutes per game, this totals almost 1900 hours of in-game time, not even counting the fact that every fortnight a new hero will be released, setting you back about 400 coins, or another 29 hours.

Not only that, the balance issue is with regards to a single instance of the game, not the stateful progression of your account, so in the minimum 3200 games you’ll be playing before you have every hero, each one will be horribly imbalanced in various directions, and the outcomes of which will be practically meaningless.

• The ‘If you really like the game and you want to play it seriously then you’ll fork out for all (or a large majority of) the heroes’ argument.

Again, this is silly. People should be able to play a balanced game at all levels of play, and to say ‘Well all the good players have all the heroes anyway’ is both condescending and contemptuous. Not only that, but people should, in most circumstances, follow the philosophy of playing to win. I would go as far as to argue that not buying every hero is deliberately putting you at a disadvantage, and runs contrary to that philosophy. Given the choice between not playing and spending over $200 to play optimally, I’ll take not playing, thanks.

I think S2 really need to think about how they wish to push Heroes of Newerth. I do not believe that a game this fundamentally flawed with regards to balance should ever be accepted as a being competitively viable, which is a great shame given the skill, difficulty, intensity and depth of play demonstrated in the current (pre-2.1) tournament scene. I can’t honestly play and support a game which goes quite so against my own personal beliefs of fair-play, equality, and the separation of a game and the outside world (particularly stateful interactions involving money!). I’m extremely disappointed that this has happened, and I hope something is done to remedy it. My suggestion would be a single, one-off payment, no-more than that of buying a new game, which unlocks every current and future hero for play, permanently, similar to the model Stunlock Studios have implemented in Bloodline Champions.