how i am testing (is there a better method)

[solarr8]

Just a short note about how i test. I havn't the time to throughly play every single race nor do i want tooo many variables to be involved when i attempt to isolate what i believe to be a possible issue.

You will see (if you work off of any of my saves) that i will employ the use of codes to more easily and quicky set up the desired cercumstances for testing. Unfortunatly, most of the issues i address are mid to end game issues and thus cannot be tested without either playing through or using codes.

You will also notice that i take away as many veriables as possible to insure no interfearence with the data i collect

For example, The lich i used max power and go far. This enabled me to test the lichs mettle on many low mid and higher level dungens. As i was wandering at night (my night time) last night i started plowing the lich through these dungens (i also used the all spells code)

I noticed that the lich had very circumstancial performance, much like the knight. I like this because it does require the use of more strategy however the main issue i saw in this is that with the right cercumstance he was easily defeated.

So basicly thats how im testing now. Without having to playthrough faiths and do this sort of thing manually are there any suggestions people can give to help make sure there are no "loose ends" in testing?

[johngalt137]

One thing I would suggest about playtesting to balance games, or playtesting games in general is to play the game and keep good notes. The problem with setting up specific scenarios is that it may turn out that in real games these scenarios don't occur or you miss taking into account something that would happen in a real game. Also you miss bugs and balance issues that you aren't looking for. When a person setup-tests they let personal bias creep into the testing, because they only test the things they think might be wrong in the first place how surprising is it when they discover a "problem"? Playing the game forces you to think and exhaust different possibilities, it causes you to stumble onto bugs, exploits, balance issues, and occassionally a good strategy. However, you shouldn't just play the game haphazardly, you should play the game that challenges you, and keep good notes so that when you find that constantly player x is really abusing seagulls you might consider using them yourself, or looking more closely at their statistics. Or if you find that you never ever seem to want to build the archer unit of your faith, there might be an issue there. I have always thought the best challenge is to be found in multiplayer, because humans bring things to the table that the computer just can't. (It also seems to have the most bugs). So I guess the answer is that there is no magic formula. You can only go so far with this type of testing. Just like you can only go so far with looking at the unit statistics. This is due to the simple fact that one of the most important factors in balance and testing is context. Sorry I can't be of more help.

[solarr8]

John. Do appretiate your feedback however there are just a few gaps in your logic.

Every tester has a bias. This includes Boaster as well as myself. And generally you don't as a tester waste time testing anything you don't think is worng. If you look up what quality assurance is you'll know what i'm talking about. Furthermore i actually worked and got paid as a tester as one of my past jobs and we didn't play the entire game over and over we played parts of the game and set up different scenarios. You can't play the game start to finish over and over not only because of time constrainhts but because if you don't isolate variables your data becomes flooded with things that can impact what you test. (kind of hard to explain that one. See quality assurance)

I have already stated that i am interested in teasting multiplayer the problem however is that when playing againsted someone else and you and that person are not evenly matched or playing the same than that variable will corrupt testing. My tests are very black and white because they take as much humen out of the equation as possible. I realize i click the mouse on the battle screen different than others which is why i have encouraged people to do as i have and generate their own assumptions.

I understand the concept of wanting legendaries to be different in role as well as ability however I still see order at a disadvantage because of my most importand unaddressed question. What is the justification (or) What makes up for the fact that a capstone unit that can be summioned only once per game dies so easily right out of the temple when others do not. Same situation as the lich. To have a level 15 wizard turn into a level 6 lich is mind boggling and the reasen why i wouldnt just use a level 6 mage to transform is poorly documented and understood as are many things in the mod. Not complaining that the work isnt being done quick enough [By the way] only stating. Furthermore the status of the knight as "second champion" or "Second lord" without any elaboration as to why this is a justification causes further confusion.

As it stands i still think any once per game creature that has the ability to die so easily while fufilling the role in a game of every unit having no purpose but to fight is mind boggeling. I encourage anyone who can help explain this justification to me to please do so so that i may take those things into consideration when i reevaluate my opinion.

As far as keeping good notes one of my earliest posts has almost a page or so of single spaced documentation and as far as context you'll need to help me with the justification i requested above.

Im certainly not a perfect tester but i think that furthers another one of my points. If spending the amount of time i have with the game still hasnt lead to me being able to figure out the "right" way to play the knight then the knight itself has to be more player friendly or something else in the game has to be apparent enough to the game to allude to how the player should approach the knight.

This i believe is an area that the mod severly lacks. In starcraft the game is designed in such a way that even if you have never played the zerg once you summin the ultralisk and see how he works by moving him and attacking and seeing his limitations you are not left wondering why he works a certain way even if you die by a battle cruiser your not going to say "this is gimp i couldnt even attack the cruiser" because the game SHOWS you why this is allowed to happen. That and you're not limited to 1 per game.

Unfortunatly in a turn based game this is alot harder. Since the legendary creatures in the mod are according to boaster ment to be played differently then there needs to be some kind of cue in the game explaining to the player how to approach them. Anyone new to the change to the knight or the mod or the way the knight is expected to be played etc etc etc is going to look at the knight and go [What The FUCK].

Legendary that takes up the same spot in an army and looks like the warrior class tells me how he should be used in a game but when i think legendary i think strongest and im betting that those unfamiliar with the developers thoughts are going to have the same assumptions. Naturally the ultralisk not only has a powerful attack but can also take a beating just like the cruser and the archon. Seeing the knight not being able to take the same beating out of the temple as the capstone unit of othe races can makes me ask why. Right now i try as i might i just cant figure that out.

[johngalt137]

You have demonstrated the difference between testing for balance and testing for bugs. When testing for bugs you set up specific situations to bring the bugs about. Doing this when balance testing, especially at this late stage (post release), just isn't as useful. The reason is that you drop important context. For example you assume that both, say, Chaos and Order will summon their Legendary Creature at precisely the same turn. It is far more likely, that between players of even skill that either Chaos or Order has an edge in that race. Your method of testing erases this edge. Furthermore your method of testing doesn't take into account the travel time if Chaos wants to summon their legendary creature and send it straight for the knight. You still haven't adequately shown that doing such a thing would even be a good idea for chaos. I've said it before and I'll say it again, it isn't certain that you're wrong, but what is certain is that you don't have the data or the notes to back up your claim. You seem to think that any failure when you control the knight is the games fault and not yours.