Bracket Racing Wcg Style

[ThE_MaD_ShOt]

Ok folks, we going to have us some down home bracket racing with our crunching rigs. This is not going to start for atleast 1 week due to needing to iron out more details and making sure I haven't missed anything.

How this will work:

One person calls another out. If the person accepts the challenge they quote the call out post with there id of what that are going to use. The person being called out has 48hrs to accept or decline the call out.

Each person chooses which of their rigs they want to use for the contest. It can be as little as one or all you have running. Id's per freedc of said rigs must be posted in this thread in your call out/acceptance thread.

Race/contest runs for 7 days, starts after final update the first night and ends after final update on the last night.

Each contestant estimates what the total points the chosen rig/rigs will accumulate for those 7 days. The estimates must be posted before the start of the contest.

The person that guess the closes without there rig/rigs running over wins. In the case of a double breakout the one that is still closes to there guess wins.

If your shit breaks your still in the contest. You also may still win if the other person breaksouts.

Use whatever strategy you think will get you the win. I know I will be and have ideals on what will pay off for me.

Prizes hmm don't have any. But you are free to wager between yourselves.

If this seems like something that has a good following and lots of participation then we may run it as a season with a grand prize for the one who wins the most contests/races. I case of a tie here we will have to work it out as I have no clue and haven't though that far.

This is open to anyone on our team, also it doesn't matter who calls whom out as the rules for the race evens the field. The one who guesses the closest to what their chosen rigs total point accumulation for the duration of their contest is the winner. No if's and or buts. If you lose put your shit back on the trailer and wait for your next contest whether that be you calling someone else out or someone calling you out.


If this does take off I will come up with a points system to determine who is the seasons winner. Also a season may be up to 6 months in duration with breaks due to team challenges. This could be a yearly thing if we get enough response to it.

If you have further ideas on the way it should be run please pm me a discuss and I will take it up with the Chamber of Team Tpu Wcg Bracket Racing Association of Crunching People Wanting to Have Fun Organization.




ThE_MaD_ShOt

 

[ThE_MaD_ShOt]

Oh and First call out: remeber this is going to happen for atleast a week to make sure all details are worked through.

I am calling out Ion for a picture of a doughnut.

Rig I will use are
3200479 and 3203330

 

[Norton]

Great idea!

Count me in but I will wait for someone to call me out first

 

[twilyth]

I think I'm going to pass on this for now. I barely pay attention to most of my rigs except to see if they're still running. I could use the boincstats host section to see what the ppd pattern is for each rig and make an educated guess but I'm not sure I'm interested enough to do that.

It sounds like it could be fun but I think I'd like to observe for a while first. I do wish you the best of luck though and commend you for having the idea. Anything that stimulates interest in the team is a very good thing.

 

[stinger608]

Something you might want to link to @ThE_MaD_ShOt is where the team members need to go to get the weekly stats in the event a member doesn't know how to gain the information.

 

[[Ion]]

Oh and First call out: remeber this is going to happen for atleast a week to make sure all details are worked through.

I am calling out Ion for a picture of a doughnut.

Rig I will use are
3200479 and 3203330

Count me in!

 

[manofthem]

Bracket racing.... Gotta admit, I feel like Michael Scott here

But seriously I look forward to the little battle between @ThE_MaD_ShOt vs @[Ion], and maybe then I'll even figure this stuff out

 

[t_ski]

Subbed to watch

 

[Arjai]

Hmmmm....interesting!

 

[BarbaricSoul]

I'd rather have it be a call out of a specific system and see who can outscore who.

For instance-
Norton's 4P system VS Ion's 4P system
Norton's 2600k system vs manofthem's 4770k system
My 3930k system vs your (Mad Shot's) 3930k system

 

[[Ion]]

I'd rather have it be a call out of a specific system and see who can outscore who.

For instance-
Norton's 4P system VS Ion's 4P system
Norton's 2600k system vs manofthem's 4770k system
My 3930k system vs your (Mad Shot's) 3930k system

The advantage of matching up two different systems is that can help smooth over differences. For instance, I'm pretty sure that my 12c Xeon would beat MadShot's any day. But his 3930k ought to be faster, so combine the two and it's a fairer race. Similarly, I have the worst of the 4Ps ("only" 32 cores) so Buck's or Norton's would (or should) plow right over it.

 

[CAPSLOCKSTUCK]

Im with @BarbaricSoul 's opinion, seems a little more scientific, or at least there is the chance to draw better conclusions from the results with matched or closely matched setups. So if anyone has something similar to ThugXeon 2.5ghz - 8 threads i will happily enter the fray.


I am not sure if this goes outside the definition of a bracket race but that whole concept is new to me anyway.

 

[BarbaricSoul]

The advantage of matching up two different systems is that can help smooth over differences. For instance, I'm pretty sure that my 12c Xeon would beat MadShot's any day. But his 3930k ought to be faster, so combine the two and it's a fairer race. Similarly, I have the worst of the 4Ps ("only" 32 cores) so Buck's or Norton's would (or should) plow right over it.

The issued challenge would still need to be accepted. I only used the names I used for the example because I knew you have a 4p system(I'm sorry if your name comes to mind before Buck's does when I think about WCG), Norton has a 4p system, both Norton and manofthem have 4c/8t i7 systems, and Mad Shot 3930k system is very similar to mine (I know, it came from me). The challenge having to be accepted should prevent any unfairly mismatched challenges.

 

[Norton]

Bracket racing is setup to allow direct competition of systems that are of different classes.... the win goes to the one who knows his rig better and can hit their target ppd (I.e Dial in) without exceeding it (Break out)

Example:

Rig A produces a 7 day average of 7,000 ppd and the 28 day average, or RAC, of 6,800 ppd

Rig B produces a 7 day average of 3,200 ppd and a 28 day average, or RAC, of 3,000 ppd

Rig

They opt to race:
- Rig A: Dial's in at his/her RAC- 6,800 ppd
- Rig B: Dial's in at 3,500 ppd

At the end of the race:
- Rig A: produces 7,400 ppd. This is 600 ppd more than his Dial in and loses (called Break out)
- Rig B: produces 3,450 ppd. This is 50 ppd below his Dial in and wins.

*In this example- If Rig a produced 6,500 ppd instead and didn't Break out, he would still lose since Rig B was closest to his Dial in.

You may find the concept a little confusing at first but it is pretty cool once you see it done a few times

 

[Bow]

And I thought I was the only Drag Racing fan here.

 

[manofthem]

Bracket racing is setup to allow direct competition of systems that are of different classes.... the win goes to the one who knows his rig better and can hit their target ppd (I.e Dial in) without exceeding it (Break out)

Example:

Rig A produces a 7 day average of 7,000 ppd and the 28 day average, or RAC, of 6,800 ppd

Rig B produces a 7 day average of 3,200 ppd and a 28 day average, or RAC, of 3,000 ppd

Rig

They opt to race:
- Rig A: Dial's in at his/her RAC- 6,800 ppd
- Rig B: Dial's in at 3,500 ppd

At the end of the race:
- Rig A: produces 7,400 ppd. This is 600 ppd more than his Dial in and loses (called Break out)
- Rig B: produces 3,450 ppd. This is 50 ppd below his Dial in and wins.

*In this example- If Rig a produced 6,500 ppd instead and didn't Break out, he would still lose since Rig B was closest to his Dial in.

You may find the concept a little confusing at first but it is pretty cool once you see it done a few times

Oh ok I see what's going on here. I'm finally understanding this bracket racing thing, and it's sounding pretty interesting, the key being to really know your rigs and their abilities.

 

[twilyth]

Bracket racing is setup to allow direct competition of systems that are of different classes.... the win goes to the one who knows his rig better and can hit their target ppd (I.e Dial in) without exceeding it (Break out)

Example:

Rig A produces a 7 day average of 7,000 ppd and the 28 day average, or RAC, of 6,800 ppd

Rig B produces a 7 day average of 3,200 ppd and a 28 day average, or RAC, of 3,000 ppd

Rig

They opt to race:
- Rig A: Dial's in at his/her RAC- 6,800 ppd
- Rig B: Dial's in at 3,500 ppd

At the end of the race:
- Rig A: produces 7,400 ppd. This is 600 ppd more than his Dial in and loses (called Break out)
- Rig B: produces 3,450 ppd. This is 50 ppd below his Dial in and wins.

*In this example- If Rig a produced 6,500 ppd instead and didn't Break out, he would still lose since Rig B was closest to his Dial in.

You may find the concept a little confusing at first but it is pretty cool once you see it done a few times

Thanks for the explanation. That's more or less what I thought happened. What I don't get is why is this interesting? I mean if because of the mix of work units, or lets say because of increased validation speed, I started getting more ppd than my average, why wouldn't I just pause the machine for a few hours to drop my ppd down a little?

There are a lot of variables that go into whatever your total for a given day would be and you don't have control over most of them so I'm trying to understand why I would care about this sort of "competition."

 

[stinger608]

Bracket racing is setup to allow direct competition of systems that are of different classes.... the win goes to the one who knows his rig better and can hit their target ppd (I.e Dial in) without exceeding it (Break out)

Example:

Rig A produces a 7 day average of 7,000 ppd and the 28 day average, or RAC, of 6,800 ppd

Rig B produces a 7 day average of 3,200 ppd and a 28 day average, or RAC, of 3,000 ppd

Rig

They opt to race:
- Rig A: Dial's in at his/her RAC- 6,800 ppd
- Rig B: Dial's in at 3,500 ppd

At the end of the race:
- Rig A: produces 7,400 ppd. This is 600 ppd more than his Dial in and loses (called Break out)
- Rig B: produces 3,450 ppd. This is 50 ppd below his Dial in and wins.

*In this example- If Rig a produced 6,500 ppd instead and didn't Break out, he would still lose since Rig B was closest to his Dial in.

You may find the concept a little confusing at first but it is pretty cool once you see it done a few times

Okay, but where does the light tree come in on this example???? Hmmmmmm, ANSWER ME!

 

[Arjai]

Go!!

NHRA Bracket racing...
It allows two different types of vehicles, horsepower, weight and even class can be different.

The driver states, before the race, his predicted run time, say 11.5sec. His opponent states his time also, say 9.5 sec.

At the line, the Tree is staggered to start driver one, 2 seconds before driver two.

If you run too fast, breakout, you lose! So driver one has a great run at 10.9 except, he broke his time of 11.5. Sometimes when this is about to happen, and the driver is paying attention, you will see them slam the brakes toward the end of a run, before the line!

In a perfect call, on the preset times, both cars would reach the finish line together, and often the races are way too close to call, w/out the electronic sensors!


If that were to happen, perfect runs, the win goes to the best reaction time to the green light. Often times you will see guys who are opposite ranked drivers, try and jump the light and end up redlighting, ending their run as it begins.

The theory here is similar except, we would call out our runtimes as point totals for a week, or whateva. Closest to the point total, without going over, breakout, wins.

So, for instance, @Norton 's 4P vs. my Core2Duo 6600 could be a race to a weekly point total we decide for our boxes OR, we could also say my 6600 can do 2000 points in 26 hours,....on second thought, I'm not so sure how that scenario could possibly be monitored. Scratch the second thought.

Anyway, I hope this makes some sense to those of you who have/ had no idea what Bracket Racing is/was.

I aim to Please.

 

[manofthem]

@Arjai nice write up indeed. Between yours and @Norton's posts, I've a much clearer understanding.

 

[swhite4784]

hmm, I have a IBM X3550 M4 24 cpu system @ 2.10 ghz with 128gig ram to race with.... ok I guess i have 2 of those systems. Vsphere 5.5 hosts count? as you can all tell im a noob when it comes to forums and all this WCG stuff. Im an Engineer by trade VCP/MCSE Engineer to be exact.

 

[t_ski]

Not sure if you can run WCG on VMware hosts. I could see it in a VM, but you'd probably want to do some throttling to prevent it from taking down your cluster.

 

[swhite4784]

Your right you can't run boinc on a host, you would have to build a 2012 R2 VM and apply the cores to the VM.

 

[t_ski]

Your right you can't run boinc on a host, you would have to build a 2012 R2 VM and apply the cores to the VM.

Yep, I know a little bit about VMware