6 replies to this topic

[maimaimi]

Hi

I have run some tests and this is what i came up with:

KEY: (k)1000cbill, (m)minute, (Gm)game minute

A: if you get a fairly fast mech, set it up as an overheat brawler (ie 12LPL-nova) and rush into battle you can quite reliably kill at least 1-2 assault before your own mech goes up in flames from overheat or sustained damage

this takes 6:30 (2:30 game time) per battle and gives around 120k reward (1100k/hr 18k/m 48k/Gm)

B: if you run a well reasoned combat mech (in my case gaus/ERLL/ERML-EBJ) and you fight a well reasoned fight using tactics and stuff you can rack up around 2-3 kills and 600+ damage quite easily

this takes 8:00(4:00 game time) per battle and gives around 180k reward (1350k/hr 22.5k/m 45k/Gm)

C: If you run an assault and spend the whole fight trying your hardest with a good setup all game you can maybe(if you are lucky) get 3-4 kills and 800+ damage (or otherwise go around smashing in faces in your mech)

this takes 11:00(7:00 game time) per battle and gives around 250k reward (1363k/hr 22k/m 35k/Gm)

what can we conclude from this:
TryHard assault player that works for the longest in-game for the benefit of the team gets the LEAST reward while BEZERKER bum rusher with a build specifically designed to burn like magnesium before dying and takes 1-2 targets with it gets better pay for much less effort...

REWARD SYSTEM IS SKEWED : DISCUSS
basically I hypothesize that the reward system is skewed because damage and kills simply do not give that much reward for the effort taken. This is not a shoot em up with infinite enemies, so killing 1 of 12 is not equal to killing 4 of 12 that takes significantly more effort.

TL|DR : dont try too hard to win, there is no actual extra benefit to doing more work once you have 2-3 kills in fact the reward system discourages protracted play

Edited by maimaimi, 25 September 2016 - 01:10 PM.

[Grinster]

Good analysis. Did you collect enough data to split out game mode and win/loss? My Profile stats show Conquest is best for CBills per match. Unfortunately there is only the Base Stats stats for W/L and CB/match and no per-Mech breakdown of CBills/match and W/L.

[Mawcor]

One thing I can't help but notice is that the player in the 'bum-rusher' in the scenario you described spends almost twice as much time waiting in the queue as they do actually playing, while the amount of time the assault player spends in the queue is about half the time they spend playing. Since generally people prefer actually playing the game to sitting in the queue, while in your scenario the bum-rusher will technically gain more per minute while playing, the other player will probably gain more per session because they're a lot more likely to play longer sessions due to actually enjoying themselves.

Now, I must admit I don't have any hard numbers, so this is largely anecdotal, but in my experience the biggest factor in how much you make in a match is whether or not your team won. Not only do you get a bonus for winning, but winning games are almost always ones where your team killed more mechs than the enemy, and more dead enemies means more assist money and generally higher overall damage among your team (and by extension you) as each member is more likely to live longer and be able to shoot more. So, doing what best helps your team or capitalizes the most off of what they're doing is probably going to be more productive than any sort of 'optimal run'.

Finally, I think you've gotten a couple of assumptions wrong.
In my experience, when played right the heaviest mechs tend to die earlier, since their heavier armor and weaponry is needed to spearhead the push or defense and screen for the rest of team. With the heaviest weapons, they'll generally do more damage if they charge the enemy and go out in a blaze of glory than the lighter mechs will.
I also don't think it's possible to reliably get 1-2 kills by 2:30 in a match, particularly in a lighter mech. In my experience, those mechs generally need to be opportunistic: targeting stragglers and enemies who have wandered away from their team, or attacking alongside larger vehicles which can distract enemies who will then leave themselves open for flanking attacks and the like. Basically, you'd need several things to line up: the enemy leaves a mech or two vulnerable for you to exploit, you don't get spotted on the way over or the enemy has no mechs which can respond to your movements, the enemy mech that's left behind or wanders off is not one that can defend itself against that kind of attack. If you turn the corner and the enemy team is together, 4-6 mechs are going to gun you down before you can get off a second shot. If you get spotted on the way in, enemies can either reposition to cover their straggler and you're back in the above scenario, or you'll likely get chewed up from longer ranged attacks on the way in and be too damaged to fight the lone enemy effectively. Granted, sometimes you'll get lucky and neither of these things happen, but I would hardly call that reliable. And then there's the possibility that the enemy mech can defend itself from that sort of rush tactic, either by having a similar build itself (negating the advantage, and making it a toss-up whether you or the enemy gets the kill) or having a build maybe combined with positioning that counters yours. All this to say, I don't believe joining a game with the intent of having finished your contribution will guarantee a kill, and that would have an impact on the earnings-per-game (and thus earnings per minute) over a set of games.

No reward system is perfect, but I'm pretty happy with rewards as they are in this game. Kills and assists are rewarded about equally, encouraging people to save their teammates by finishing off lower health enemies rather than discouraging them for fear of denying their teammates rewards. You get some rewards for spotting targets, though maybe they could be increased a little, and you get rewarded for some damage dealt to enemies if you're providing vision of those enemies. You get some rewards for protecting damaged teammates, and lights and mediums get some rewards for supporting their larger allies rather than running off.
So, to the question of whether or not it is skewed, probably, I don't think there's one that isn't in some way. I'm sure there are ways it could be improved. I am not convinced the problem you described is that much of a concern.

I could always be wrong though. If I may ask, what sort of testing did you do to get the numbers you mentioned?

[maimaimi]

hrm, what i was describing there were more classes of games.

the Bumrush Nova can be done with any noob, its almost guaranteed to get you 1 kill every game
the considerate EBJ is a more rare occurrence, while its posible to get more then 1 kill, it takes more effort and skill
the TryHard Assault it a very rare occurrence, it happens rarely but you still have to put same effort every game

lets say:
Bumrush Nova has 90% success rate
considerate EBJ has 60% success rate
TryHard Assault has 30% success rate

when you adjust for those success rate values simply put there is no incentive to try really hard to win since the reward is not proportionate to the effort put in.

also i do have some tables with individual game results that I keep and i can say i average around 120k/game when you factor in the total flop games and really good runs

[Mawcor]

maimaimi, on 27 September 2016 - 12:25 AM, said:

when you adjust for those success rate values simply put there is no incentive to try really hard to win since the reward is not proportionate to the effort put in.

This is probably the point I most disagree with. Even if everything else you've said is correct, which it might be and I'll discuss that in a moment, this one can't be true. Even if the numbers are a little skewed, such that the highest earnings per game minute come from making a suicidal rush against the enemy every game, it won't take long for that to be a really boring way to play. The numbers aren't so bad that that is the only way to make money, in my experience winning a match always gives significantly more money than losing, so the only people who would want to make use of that are those who don't actually care about the gameplay at all and only play to watch their numbers of C-bills go up. You can't even really say they want to get more mechs, because they clearly wouldn't have any use for them.
While for the most part everyone would probably prefer to make more c-bills than less, I willing to bet that almost everyone who plays this game does so at the end of the day because they get some enjoyment from running around in a giant stompy robot shooting other giant stompy robots. And the suicide rush strategy doesn't really support that.

Aside from that, I'm largely still just wondering where you're getting those numbers from. Are you sure a suicidal rush has a 90% chance of working? What counts as success? Why is the assault always the unfortunate team player, the medium always the one committing suicide, and the heavy always the middle ground? If the assault does as much as the medium does (1-2 kills) and then dies, does that count as a success or a failure? What about other factors, i.e. games where someone does 500 damage but only gets assists?

I've also found that the mech sizes tend to have the opposite survival times to what you've described, with assaults tending to die first due to their large size, heavier weapons complement and inability to escape lead to them being focused down first, while mediums have more speed to get out of trouble and reduced size and weaponry making them typically easier to ignore until later parts of the battle.
Unless one of them is out of position, such as charging straight at the enemy, in which case they need amazing positioning or a lot of luck to not just get gunned down before they can get off a shot. While maybe not the most sophisticated strategy, it requires at least a strong understanding of the map, positioning and the mech's weapons for a suicide rush to be anything close to reliable.

My other thought is, have you tried acting on what you're describing here, that is to say building an overheat brawling and rushing the enemy each game? If you have, did it work? If not, what stopped you? Have you been seeing this particular build-playstyle combination at lot?

[maimaimi]

My numbers are mostly backed by results... I occasionally keep a record of winning, in game time and the rest per mech per setup, it's fairly useful

The numbers and strategy notes come from patterns I noted across games for certain mech builds

The bum rush nova is a bit of an exaggeration, but is my real play style. I usually put a bit more thought into my approaches and have even had games with 5+ kills with most being solo kills from backstabbing/headshot. On average the inflame time is 2:40 for that mech and pays out 105k/game

The considerate heavy is based on the mid band or "good game" average of my EBJ, MDD and various other setups meant for sticking a game out taking a slow and steady approach. these split about even into 3 groupings; really bad(I died), middling(I scored a few kills and earn 120k), and good(assist/kill on almost every enemy mech). These average 7minutes roughly, but some hit closer to the 10 minute mark

The try hard assault is not based on "assaults" per say, although a well set up assault is more likely to hit there but rather a choice selection of those games that went well for whatever reason. these have a few consistent features: they are rare, they are unpredictable, and they were hard games where I somehow won as last man standing or rolled half the enemy team solo by the sheer coincidence that I was lucky and they were not.

TL/DR: I am not saying "all games are as such" or "this gives best reward", but rather commenting on a trend in rewards where 2 quick mediocre/bad games give the same reward as a game where you can with confidence say "I did all the work team, 12 kills 3200 damage, GG all."
(Not to say I have 12 kills or it will take 3200 damage unless you have really poor aim...)

although I do have 2 ERSL nova for rapidly farming c-bills when I want that extra million to buy my next mech. Overheat nova is a pretty power trippy mech, there is a certain rush you get in a pug when you hop a hill, core someone's rear, headshot another heavy and then bum rush the other 6 or so mechs in the area vomiting 12ERSL at their CT/cockpit with no regard to your own survival... real story btw, game time on that was 2:03 with 2 kills and 238 damage, 113k reward.

Edited by maimaimi, 30 September 2016 - 06:41 PM.

[Mawcor]

maimaimi, on 30 September 2016 - 06:26 PM, said:

TL/DR: I am not saying "all games are as such" or "this gives best reward", but rather commenting on a trend in rewards where 2 quick mediocre/bad games give the same reward as a game where you can with confidence say "I did all the work team, 12 kills 3200 damage, GG all."

Ah ha, this is what I needed to see. I wasn't sure if you were arguing that the optimal way to play MWO was to get killed in the first two minutes, or if the argument was that smaller, more maneuverable mechs with short range and hard hitting loadouts received disproportionate rewards.

It's possible. My experience is that unless I end up as the bait generally the team being successful seems to be the biggest factor in my games, but the rewards might not be all that different.
If it's true I don't mind it too much, mostly because my satisfaction at a given match comes from how well I did and how well the team did first and second, with what my rewards for the match were coming in at a distant third. When I think about it, maybe that's not such a bad thing, as when I have a disappointing match now I know I can say to myself, 'hey, at least it didn't limit my rewards for this'.