Top 25 ETCs after Week 3

[mwtidd]

Originally Posted by speedbuggy76

Alright, I've spent some time trying to re-create your calculations using the data from Peachtree. However, I keep getting a different number for our team (2415) than you do.

Some sample data:
In match 5, blue alliance scored 39 points and was composed of 2415, 1683, and 3694
2415's average alliance score: 77.1
1683's average alliance score: 26.2
3694's average alliance score: 7.1

The average score of the teams on the alliance = (77.1+26.2+7.1)/3 = 36.8

2415's contribution = 39*77.1/36.8 = 81.7

Is this correct? I ran this calculation for all of our matches and found us to have an average contribution of 119.2. This disagrees with your value of 38. Any idea what I'm doing wrong/differently?

Also, how did you break up the Peachtree data for our robot into ERC and EMC? I'm just using the data set on the FIRST website which only contains total alliance scores.

Thanks for putting all this together and answering questions! You have my interest piqued.

Sorry, that's my mistake in the representation of the algorithm. You should use the total score for your alliance not the average...

so
(77.1+26.2+7.1)

instead of
(77.1+26.2+7.1)/3

*trying to run through it off the top of my head doesn't always work out

I have also thought about generating highs and lows, by removing a teams highest and lowest matches, to give an ETC window rather than a set value.

every match is tweeted to the FRCFMS twitter feed and it contains one extra piece of information the red and blue bonus, which is the points put up in the race. by knowing this I can use the same algorithm i use for etc, but use it on the bonus score and then the alliance score - the bonus score (robot score)

[techvikesmom]

Just a note: Team 2054 has been renamed "TECHVIKES", not big blue crew. This is the second year under this name.

[mwtidd]

Originally Posted by techvikesmom

Just a note: Team 2054 has been renamed "TECHVIKES", not big blue crew. This is the second year under this name.

thanks! your name's been changed in my database... a lot of the names are a bit dated. at least we got one fixed .

[KrazyCarl92]

So where do the cheesy poofs or high rollers factor into all of this?

[mwtidd]

Originally Posted by KrazyCarl92

So where do the cheesy poofs or high rollers factor into all of this?

see my new signature My system work off the twitter feed, if your regional is 1 of the 4 that didn't post, I don't have numbers for you. I will try to support untweeted regionals but I won't be able to supply ERCs and EMCs

[Dave Scheck]

Maybe it's just me, but I don't get it. You have us with an EMC of 15. We put our minibot up in first place in 15 of our 16 matches, and got second in the 16th to boost our RPs. Our real minibot contribution based on that data is 29.3751 ( (30 * 15 + 20) / 16). I realize that you're just making an estimate, but when the data is so far from the actual value I don't see how it can be useful.

[Paul Copioli]

Originally Posted by Dave Scheck

Maybe it's just me, but I don't get it. You have us with an EMC of 15. We put our minibot up in first place in 15 of our 16 matches, and got second in the 16th to boost our RPs. Our real minibot contribution based on that data is 29.3751 ( (30 * 15 + 20) / 16). I realize that you're just making an estimate, but when the data is so far from the actual value I don't see how it can be useful.

I have the exact same questions. We have similar results from Detroit for our minibot.

[mwtidd]

Originally Posted by Dave Scheck

Maybe it's just me, but I don't get it. You have us with an EMC of 15. We put our minibot up in first place in 15 of our 16 matches, and got second in the 16th to boost our RPs. Our real minibot contribution based on that data is 29.3751 ( (30 * 15 + 20) / 16). I realize that you're just making an estimate, but when the data is so far from the actual value I don't see how it can be useful.

Originally Posted by Paul Copioli

I have the exact same questions. We have similar results from Detroit for our minibot.

Hey guys, thanks for your posts. I'm glad to know how wrong it was. The reason your EMC is low is because of point inflation. Team's that had 2 minibots go up receive a greater emc than those who only have 1 go up.

Also I found a bug in my database, so it changed the values for some of the team's a little bit (111 being one of them). Please note the top minibot score in the nation is a 19, Dave 111 has a 17 so if you look at it in that context, it means you have one of the best minibots in FIRST.

Paul, 217 you also have a minibot with a rank of 17.

for example if you look at the new list below teams like 148, 40, and 1126 although they are top 25 teams, are lacking with regards to minibots.

Also one other thing, these are all calculated based on qualifying match scores only.

Right now minibot contributions are split among all 3 alliance members, I think I'm going to actually try to figure out the minibot finish order by the scores... shouldn't be too hard

[Jim Zondag]

Originally Posted by Dave Scheck

Maybe it's just me, but I don't get it.

Me either, and I'm not convinced yet that I need any new acronyms or methods. Everyone keeps obsessing about who is the best, but this is irrelevant until Einstien. At a real tournament, you don't really need a spreadsheet to tell you who is the best. This is obvious; my mom can tell you who is the best with no data at all. A thousand varied analytical methods will be able to give you the top 8 teams in order.
You do, however, need good data and methods to tell you who is 16th-24th. This is where the real value add of competitve analysis is: not for the first pick, but for the second.

[mwtidd]

Originally Posted by Jim Zondag

Me either, and I'm not convinced yet that I need any new acronyms or methods. Everyone keeps obsessing about who is the best, but this is irrelevant until Einstien. At a real tournament, you don't really need a spreadsheet to tell you who is the best. This is obvious; my mom can tell you who is the best with no data at all. A thousand varied analytical methods will be able to give you the top 8 teams in order.
You do, however, need good data and methods to tell you who is 16th-24th. This is where the real value add of competitve analysis is: not for the first pick, but for the second.

Well hindsight is 20/20 but lets look back at the selection for western Michigan where using EMC could have resulted in a very different story, and it all starts with the very first pick.

2054 picks team 67.
2054: ERC- 25 EMC-17 (top minibot and robot)
67: ERC- 20 EMC-8

it looks like an obvious and smart pick but...
in my opinion they overlook one key player which would have changed everything for them.

2767: ERC- 3 EMC-16

Had 2054 picked 2767 they would have had a monopoly over the best minibots. essentially a guaranteed 50 pts each match.

1918 at R: 16 and M:9 would have picked team 67
but finishing 3rd and 4th in the minibot race would have a very tough time beating 2054.

so i think sometimes we even overlook a potential first pick, because we want to form a powerhouse...

[XaulZan11]

Originally Posted by lineskier

Well hindsight is 20/20 but lets look back at the selection for western Michigan where using EMC could have resulted in a very different story....so i think sometimes we even overlook a potential first pick, because we want to form a powerhouse...

Or maybe teams actually watched matches instead of relying on final scores and fancy formulas to rank teams...

I know I may be sounding like a jerk, but as someone who works on the picking list for my team, I would be pretty upset if I worked hard and scouted every match only to be second guessed by someone who likely didn't see any matches. OPR, ETC and various metrics are nice, but there is NO SUBSTITUTE for actually watching matches and seeing how good teams are.

[waialua359]

Originally Posted by XaulZan11

Or maybe teams actually watched matches instead of relying on final scores and fancy formulas to rank teams...

I know I may be sounding like a jerk, but as someone who works on the picking list for my team, I would be pretty upset if I worked hard and scouted every match only to be second guessed by someone who likely didn't see any matches. OPR, ETC and various metrics are nice, but there is NO SUBSTITUTE for actually watching matches and seeing how good teams are.

This is the absolute truth.
The data is a nice to know and cool to know, but it doesnt change how I go about selecting alliance partners, if I'm in the top 8.
In fact, I know who I'd pick this weekend in HI if I had the #1 pick, even before watching any matches at all.

[ttakashima]

I tend to use both as a resource when helping my team chose its partners, numbers and watching matches. While having your scouter watch matches, interview teams, go over autonomous modes, who scored which object and how many times etc. In the end it comes down to both watching matches and using those numbers, weather it be this algorithm or your own method of analyzing different points. Both could be a useful resource in picking a team to be on your alliance.

[mwtidd]

Originally Posted by XaulZan11

Or maybe teams actually watched matches instead of relying on final scores and fancy formulas to rank teams...

I know I may be sounding like a jerk, but as someone who works on the picking list for my team, I would be pretty upset if I worked hard and scouted every match only to be second guessed by someone who likely didn't see any matches. OPR, ETC and various metrics are nice, but there is NO SUBSTITUTE for actually watching matches and seeing how good teams are.

No no, once again, of course this is the case. Where I can't watch the matches, I was trying to gain insights from the #s. I am trying to make statements, to see people prove me wrong . I can never improve without being corrected. These insights are phenomenal

Originally Posted by Joe Ross

Maybe EMC isn't as good of indicator as OPR.

67 - 64.1978
2054 - 58.5124
2767 - 23.3331

Maybe 2767 didn't really have the second best minibot at the regional, but rather was the luckiest because they always had a 2nd minibot on their alliances. 16.67 is the maximum minibot score (50/3). However, to get that you need help from your alliance partners. If you had the absolute fastest minibot, but never had a 2nd minibot from your partners, you would get 10. As others said, only watching the matches lets you know the full story.

Thanks for these insights, and for the record overall I had 2767 ranked very low too. with an ETC of 19 compared to 2054s 43.

I am working on a new way of trying to calculate EMCs by actually trying to estimate which robots put minibots up, and how much they would be worth. So if you put up first and second you both get 25 points. as opposed to 16.

You are absolutely right that EMC can be very much swayed by luck. but that is why I picked 2767, I wanted to see what you guys had to say.

Maybe I would have been better off saying 2767 had a very reliably minibot, to some teams this is more valuable than the fastest.

Regarding reliability how did 2767's minibot compare to 67, 74 , 1918.

these are the 3 teams ERCs
67 - 20
2054 - 25
2767 - 3

again please disprove these claims but:
2054 and 67 had the 2 best robots or most reliable robots
1918 was very good or very reliable
74 was good and reliable

now these statements may be wrong, I'm looking for insights on them

[Joe Ross]

Originally Posted by lineskier

Well hindsight is 20/20 but lets look back at the selection for western Michigan where using EMC could have resulted in a very different story, and it all starts with the very first pick.

Maybe EMC isn't as good of indicator as OPR.

67 - 64.1978
2054 - 58.5124
2767 - 23.3331

Maybe 2767 didn't really have the second best minibot at the regional, but rather was the luckiest because they always had a 2nd minibot on their alliances. 16.67 is the maximum minibot score (50/3). However, to get that you need help from your alliance partners. If you had the absolute fastest minibot, but never had a 2nd minibot from your partners, you would get 10. As others said, only watching the matches lets you know the full story.