I've been working on a game mechanic that revolves around sword-combat, and have reached a point at which I feel that I could use an external perspective on it. To that end I have a basic prototype (a link to which should be at the end of this post), and would appreciate feedback on it.

(This is, by the way, the second version of this prototype; I posted an earlier version on other fora, and this incorporates some of the feedback that I received regarding that version.)

Regarding feedback, I'm interested in any impressions, but am primarily focussed on whether this game mechanic is worth pursuing further. I do have a few specific questions, but would appreciate it if anyone inclined to try the prototype read them after playing, simply to reduce any bias that they might introduce if they're in mind during play; to that end, I'm hiding them within a "spoiler" enclosure.

Feedback questions:
[spoiler]How intuitive did you find the control method?
Was the combat fun?
How challenging was it to beat the enemy character?[/spoiler]

Known issues:

• Horrible, horrible graphics and sound. ^^;
• A minor feature not mentioned in the instructions: press "=" and "-" on your keyboard to increase and decrease the game resolution, respectively.
• If the game doesn't start up in full-screen mode, try one of the above resolution-change keys to force it into full-screen.
• The mouse, although hidden, isn't constrained to the window; as a result I recommend sticking to full-screen mode, and even then the cursor may escape on multi-monitor setups.
• It's possible to "dodge" out of the circular area in which the game starts.

The download is a Windows installer about 22MB in size. (It should work on both 7 and XP; I don't know about 8, I'm afraid.) The engine used is Panda3D.

You should be able to download the prototype via this link.

im naturally atracted to swordfighting games so I gave this a try. and I liked it

it feels natural to move the sword with the mouse like that, and since this is a duel fight with the view locked at the enemy you don't have the typical problem of the genre (my own swordfighting game included) of needing the mouse for something else (rotating the view)

I found two flaws:

- attacking doesn't quite work as you described it. the instructions say 'swipe the mouse with the left mouse button held'. this would mean I need to click and hold, then drag. but that didn't quite feel like it because just by moving one pixel counts as a drag, and this sensitivity makes it impossible to get the attack the way I intend it to.

it felt more effective to move the mouse first, and mid-movement just click the mouse button

- you can spam attacks indiscriminately, and the game becomes extremely easy

was it fun? yes, it's a fresh take on swordfighting

how challenging was it? not that much after the first try. the collision detection is way too broad and generic (ie. at any angle will your sword block the enemy's) so it's really easy to turtle. not so easy to hit him, sure, but if you keep parrying eventually you get there (or if you spam attacks). I also felt we both had too much health points, but maybe that's just my own taste of a more 'life or death' scenario in swordfights.

now you mention this is a game mechanic, so this tells me this is only a small part of your game. if you have the time to improve it and add variety I'd say it's worth it.

btw it works in win8, but MS smartscreen filter didn't like the installer at all

I agreed with Chosker,

Although you can swipe a mouse to move a sword once stabbed but miss.

I fee dodgel it isn't useful as parrying, and you will miss when dodge and attack most of time, maybe because of known issue.

Also, if you intently miss an attack (stab left or right side) will make enemy stay in parry mode, and you can drag a sword into body for free hit.

Thank you both for your feedback!

First of all, I'm glad that it seems to have promise!

I find it interesting to note that, across the various fora on which I've posted prototypes of this mechanic, the reported difficulty of the combat has varied significantly, with some people finding it easy and others finding it very difficult. I suspect that most of those who find it easy are--as seems to be the case with you two--discovering and exploiting the AI's weakness to spammed attacks.

- attacking doesn't quite work as you described it. the instructions say 'swipe the mouse with the left mouse button held'. this would mean I need to click and hold, then drag. but that didn't quite feel like it because just by moving one pixel counts as a drag, and this sensitivity makes it impossible to get the attack the way I intend it to.

it felt more effective to move the mouse first, and mid-movement just click the mouse button

Hmm... Funnily enough, I had originally placed a minimum length and speed threshold on drag detection, but feedback on the version that contained that indicated a sense of lag in the input, which I think that I ascribed in part to the delay incurred by those thresholds. However, I also had a visible mouse cursor, which may have been producing a bit of dissonance between the near-one-to-one response of that cursor and the slower, interpolated response of the character.

This calls for more thought and experimentation on my part, I think.

- you can spam attacks indiscriminately, and the game becomes extremely easy

Heh, yes, I've encountered this myself, I believe. I'm not yet quite sure of how best to handle this, but do have some avenues of investigation to pursue, I think.

(For one, I want to do some research into how an actual combatant might be advised to deal with so aggressive a foe; I might also slow the player down, enforcing a bit of a pause after an attack, making the player a little vulnerable if attacking without thought.)

the collision detection is way too broad and generic (ie. at any angle will your sword block the enemy's) so it's really easy to turtle.

I'm looking into that--the collision system is something that I built myself (in short, I wanted a system designed to handle high-speed collisions), but I'm having some trouble with reliably finding appropriate collision normals for "glancing parries"--that is, situations in which the striking sword glances off of the defending sword (as in a downward strike glancing off of a near-vertical defender's sword). I'm working on it.

I also felt we both had too much health points ...

I agree, actually; I'll likely reduce the health at some point. For now, it at least reduces the amount of time between restarts. I might implement automatic restarting for the next version of the prototype; if I do, I may well reduce the health, too.

now you mention this is a game mechanic, so this tells me this is only a small part of your game.

Well spotted! You are indeed correct: I intend for combat in this game to involve scattered, discrete encounters, rather than the more usual business of wading through mobs.

btw it works in win8, but MS smartscreen filter didn't like the installer at all

I'm glad that it works, at the least.

I don't use Windows 8, and so am not familiar with the "smartscreen filter"--how is it responding? Is it like the warning screen that Windows 7 produces, encouraging the user to install only if the file is from a trusted source?

I fee dodgel it isn't useful as parrying, ...

Hmm... I have ideas for enemies that would somewhat rely on the player being able to dodge. At the moment I agree--successfully dodging is trickier than parrying, and can be tricky to attack from--but for now I think that I'll leave it in.

... and you will miss when dodge and attack most of time, maybe because of known issue.

Heh, true--although I've found that with good timing it can be a way of getting in a hit without the AI managing to parry.

Also, if you intently miss an attack (stab left or right side) will make enemy stay in parry mode, and you can drag a sword into body for free hit.

Ooh, well spotted--I'll look into that, I believe.

New version!

Again, feedback would be appreciated.

Changes:

• Various alterations to the AI; it should hopefully provide a greater challenge now.
• Improvements to the physics system; while still not perfect, it should be rather better than it was.
• Increased the damage done by hits to the head or body.
  • It was suggested that I reduce the health of the characters. I chose instead to increase damage done in order that damage done to the head or torso be significantly more dangerous than to the arm: as it now stands (and if one were attacking just one type of body part), winning takes just three blows to the head, five to the torso and a full fifteen to the arm.
• The mechanism that recognises slashes for attacks should be improved.
• Miscellaneous other changes.

Between various of the above changes:

• Turtling should no longer be as easy as it was.
• Spamming attacks should no longer be as useful as it was.
• A somewhat overpowered side-step-and-attack manoeuvre should no longer work.

Known issues:

• The graphics and sounds remain horrible!
• The full-screen bug remains. Again, use the in-game resolution-change controls to force fullscreen. (And again, I do strongly recommend playing in full-screen.)
• There are still a few "cheap" ways to win, I think.
• The mouse is still not constrained to the window.
• It's till possible to manoeuvre out of the circle of visible floor.

AARGH! I do apologise: it would seem that I've made a very silly mistake: the version posted above isn't the new version, but the same version originally posted, I believe. >_<;;;

In short, what seems to have happened was this: I use a separate directory for building the version that I'm going to post, allowing me to leave out files not relevant to the prototype, but which are in the main directory for other purposes (such as prototypes of other gameplay elements). When I went to build the new version, I simply forgot to copy over the new files, and so simply rebuilt the last version.

I'll hopefully post the actual new version a little later on...

On the plus side, the version that I intend to post tonight should have an options menu that allows one to set the mouse sensitivity.

I very much apologise. ^^;

The real new version!

The list of changes given above should still stand, with the addition that there is now a simple options screen on which you should be able to adjust the mouse sensitivity.


@Zero_Breaker:
Thank you very much!

I'm interested to know how you feel about this version--it should hopefully address some of the issues that you mentioned (including spamming attacks and the "universal parry").

There are still, I think, a few ways to exploit the AI or input mechanism--but I'll not yet mention what they are, I think: it might be helpful to know how easy they are to discover for people who don't know the innards of the AI and input mechanisms, and haven't been facing the AI as often as I have.

Ah, thank you for the feedback!

As to those two exploits, thank you for reporting them: I don't have solutions for them in mind yet (the "universal parry" that you noted above may be in part a physics issue), but have now made note of them as issues to work on.