Friday, December 7, 2007

Assists: Are They Important?

Daymonster and I (well pretty much Daymonster and everyone) have been arguing back and forth on the issue of how important assists are and whether or not they are a sign of better teamwork. Monster had a good point with his observation that assists would be more likely to come from a player's weapon choices. People who prefer being in a pack with BRs will likely produce more assists or someone with a sniper can pretty easily get one bodyshot and have a teammate who's nearer the opponent finish them off, while players who use the power weapons more (shotty, rocks, sword) will rarely get assists. This doesn't necessarily mean the player with the power weapons isn't being a team player, he could very well be protecting his teammates, who are zoomed in on long range targets with their sniper or BR, from opponents coming in at close range. And of course there's no recordable statistic for verbal assists which come from good communication. We can always improve on that.

Another assist disagreement came from what Daymonster and I thought counted as an assist, a difference that could've strongly influenced his belief that they weren't important. If a player lowers an enemy to the brink of death and a teammate comes and finishes the kill with point blank shotty blast, does the first player get an assist? Daymonster argued no since the shotgun would've killed the enemy in one shot anyway, whether or not his shields had been gone. I argued yes because I always assumed it was that way.

So last night we did an experiment in a custom game where me and Monster logged in and he had his second controller as a dummy player. We found the player and Daymonster lowered his shields while I finished him with a point blank mauler, point blank shotty, carbine headshot, hammer then sticky grenade. We ended the game and out of these 5 trials Monster was attributed with 4 assists.

Let's please continue to debate this issue and I'll get a poll up about it.

In other news, we had a request from a guy in Texas to join our clan. He complimented this blog and has really good stats, but most importantly he sounds like he'd fit in with us. Daymonster should be posting his Get To Know 'Em very soon and let's all try to play with him this weekend.

21 comments:

Jeff said...

I say the more the merrier to the Texas guy. I also say him and Big Guy need to do a Boot-Off (similar to a Beat-Off in name only.....I hope). There will be three rounds, beer-boot, western-wear, and first one to play a game of doubles with Daymonster before Daymonster gets lag-booted (alternating turns of course, in the probable case of a tie we will revert back to another beer-boot).

Jeff said...

Oh yeah, I think assists are directly proportional to the level of teamwork. Somebody needs to call whoever invented the Quarterback rating system and devise a similar thing for Halo 3 play taking into account kills, deaths, assists, objectives (balltime, flag caps, etc.), time lagged out, time spent on Halo tilt, drunkness equalization factor (how well you would have played had you not been hammered)etc.

Pablito Neal said...

we also argued about a formula to figure out a player's value to the team in a game. i suggested:

kills/team kills + 0.5(assists/team assists) - deaths/team deaths

this is of course totally subjective and depends on how valuable we agree assists are.

Pablito Neal said...

haha and after my formula i wanted to add 1 and multiply by 100 so it resembled a qb rating

Squatting Bear said...

the formula needs to be as long and complex as possible.

Squatting Bear said...

we don't want people to be able to understand it, plus it will only gain widespread acceptance if it is super complicated. finally, we'll look like math geniouses.

Daymonster said...

agreed, should wether your team won or lost effect the rating?

Jeff said...

Win vs. loss should be factored in, but not heavily considering it this would be a personal rating and not a team rating (which could also be an interesting discussion topic). It should be disregarded for free-for-alls.

Daymonster said...

Yeah obv this would only be fore TS or Doubles... not objective.

Pablito Neal said...

looks like the poll is going well. we could end up averaging the results and all agree an assist is worth half a kill.

Pablito Neal said...

haha i don't believe there were 2 people who honestly think assists don't matter. i voted for 2/3, adeclipse said he voted 2/3 in the shoutbox.

i suggested 1/2 in my formula as a compromise.

Daymonster said...

With your forumla ou could still have a high rating even if your K/D is negative and it could be negative by quite a bit? Are assist really worth enough to counter dieing more often than killing?

Jeff said...

They are if it's a vast disproportionment. I think this line should still result in a non-horrible rating:

Kills: 7
Deaths: 13
Assists: 12

Realistically you might be able to take half of the assists and make them Kills, resulting in an average line.

Daymonster said...

I don't know what you guys think but we should almost work backwards.

What is the best you could do 50 kills, 0 assists 0 Deaths?

Worst 0 kills 0 assists 0 Deaths?

And then marks for average or middle? Like 13 kills 6 assists (assuming about 50% of kills are assisted?) 12 deaths? I can explain my reasoning for this if it doesnt make sense. Im just rambling here.

Almost Trial and error? And make a formula fit?

Jeff said...

An educated trial and error I think will work best. If we find an anomaly that shouldn't happen the way it does we can tweak it.

Daymonster said...

I think in that case we need to decide what is important to be in the formula.

I think we can all agree kills are important if not the most important. Assists are obviously a hot topic but most of us agree they are important. Deaths need to be figured into it of course.

What about things like multiple kills (double, triple etc) that could have an impact on a game and could maybe factored in. Just a thought. Like SB said it needs to be complicated.

After we figure out TS and TD we can move on to objectives. But I don't think time lagged out should be included because...well it's obv

Jeff said...

Lol....time lagged would be good for you Daymonster, it's a personal ranking not a team ranking, your time lagged could be used to determine how many more kills/deaths you would have had. It would kill your team efficiency rating though.

1 + .1? if part of winning team + (kills/team kills) + 0.5(assists/team assists) - (deaths/team deaths) + (.01 * double kills) + (.02 * triple kills) + (.03 * overkills) + (.04 * killtaculars) + (.05 * killtrocities) + (.06 * killamanjaros) + (.07 * killtastrophes) + (.08 * killapocolypses) + (.1 * killionaires) + (.01 * killing sprees) + (.02 * killing frenzies) + (.04 * running riots) + (.06 * rampages) + (.08 * untouchables) + (.1 * invincibles) * (1.0 - Blood Alcohol Content) * 100

Daymonster said...

ooooooooh, i thought you were saying it would go against you. Yeah I am fairly confident my k/d would be like 7.84 if I didnt lag out.

Daymonster said...

Dammer, also, not complicated enough.

Jeff said...

Oh don't worry this is a prototype, I had to go before I was satisfied so I posted it anyway. I'm thinking instead of having .1, .2, .3, etc for the various sprees an exponential factor should be used instead. Seriously....it's gotta be so much harder to get a killionaire than it is to get a double kill, much more than .9 .

Jeff said...

I'm thinking there should be subtractions for betrayals and suicides. Thoughts?