2013 NE FIRST District Rankings

[Jessica Boucher]

Originally Posted by jwfoss

I was wondering where we would stand in a district points structure.

You know, there's supposed to be a post by one of the NE folks who created a district points simulator so that you could play with all the values ::coughKylecough::....just haven't seen it yet

[Rosiebotboss]

Originally Posted by Jessica Boucher

You know, there's supposed to be a post by one of the NE folks who created a district points simulator so that you could play with all the values ::coughKylecough::....just haven't seen it yet

I'll remind him....

[Bill_B]

Just so I can tell if I'm understanding this. The points were calculated as if the regionals in N.E. were district events. They would then determine invitation to the proposed district championship, viz. the top 80, right? NE District championship to be played "this" weekend from which 30 or so teams being invited to CMP in St. Louis. Have I got it right?

Now can we discuss why this points summary may, or may not, be slightly misleading? First, there is the number of events attended by many of the teams. That is, not enough opportunity for the one-event teams to get points for their Q-wins in this summary. A supposed advantage to the district model is that teams will get more matches played as a result. Perhaps the points for single regional event teams could be amplified a bit to reflect this. E.g., the scores from 9 Q-matches at CT regional might get a multiplier of 14/9ths to predict performance in two district events with 7 Q-matches each.

Second, the relative size of regionals to district events would change the dynamics of those competitions somewhat. Not sure about what the effect on points would be, but I'm sure it would be there.

Then there's the even more nebulous effect of the possibility of 8 hours "out of bag" preceding 2-day district events. Is this a factor? At present, teams attending 2-day events get to schedule a sort of "virtual Thursday" by logging time with their robot out of its bag before the competition. Will we allow this in the NE District?

[MikeE]

Originally Posted by Bill_B

Just so I can tell if I'm understanding this. The points were calculated as if the regionals in N.E. were district events. They would then determine invitation to the proposed district championship, viz. the top 80, right? NE District championship to be played "this" weekend from which 30 or so teams being invited to CMP in St. Louis. Have I got it right?

Now can we discuss why this points summary may, or may not, be slightly misleading? First, there is the number of events attended by many of the teams. That is, not enough opportunity for the one-event teams to get points for their Q-wins in this summary. A supposed advantage to the district model is that teams will get more matches played as a result. Perhaps the points for single regional event teams could be amplified a bit to reflect this. E.g., the scores from 9 Q-matches at CT regional might get a multiplier of 14/9ths to predict performance in two district events with 7 Q-matches each.

Second, the relative size of regionals to district events would change the dynamics of those competitions somewhat. Not sure about what the effect on points would be, but I'm sure it would be there.

Then there's the even more nebulous effect of the possibility of 8 hours "out of bag" preceding 2-day district events. Is this a factor? At present, teams attending 2-day events get to schedule a sort of "virtual Thursday" by logging time with their robot out of its bag before the competition. Will we allow this in the NE District?

Right - the results are informative but not a good predictive model of the district.

The rankings above from this year's Regionals primarily reflect whether a team attended 1 or 2+ Regionals, but under the district model every team would be guaranteed 2 events.

The other impacts come from having smaller events, which increase the expected points awarded in several ways:

• increasing the number of qualification matches per team,
• increasing the probability of being selected for eliminations (e.g. 24/34 rather than 24/65 for a big regional),
• increasing the probability of winning an award, and
• arguably diluting the strength of elimination alliances, hence increasing the variance of expected elimination points

Assuming district events average 34 teams each (155 teams attending 2 of 9 events), and a relatively gentle schedule of 12 qualification matches, then the expected point total under the proposed model would be 39.9 for competition performance only. Award points would be additional.

The average from the table above is only 23.2 including awards.

[Rosiebotboss]

Its great that this conversation is happening. Please continue it, BUT also realize that this is NOT the final point structure. You are giving us good feedback on your thoughts and expectations of what the advancement criteria should look like. Encourage others to take part on this thread. Several of us are monitoring it daily.

The final rounds of meetings are taking place now among the movers and shakers of FIRST in New England to finalize a LOT of things that need finalizing before we go into negotiations with FIRST HQ. And a LOT of the things we are talking about are exctly the things you all brought up in theTown Hall meetings of last fall. So you are having an impact. Keep it up.

[JackS]

Originally Posted by Rosiebotboss

Encourage others to take part on this thread. Several of us are monitoring it daily.

If this is true I'd like to add my two cents to the discussion.

First, I'd have preferred to see the 5/2 awards system in place in FiM and MAR. Although the point difference is kind of splitting hairs, I think it matters when a bid to CMP is on the line.

Secondly, I'd like to see the DCMP count more than a district. Frankly, a 60-80 team DCMP is way harder than a 30 team district and teams should be rewarded as such. I think the best points system for CMP qualification would be your best (or two best) district events and 2 x DCMP.

And lastly, I really like the points system for eliminations, however I have two things I'd like to see:

1. I think the points should be bumped to 10/15/25/35 because NE does not award points for selection the same way that FiM and MAR do. This rewards teams for making eliminations more than it did before, but not quite as much
2. I think that the points should also be adjusted to be lower for the third robot on an alliance similarly to how selection points are distributed in FiM. Maybe the first two robots on an alliance get 10/15/25/35 while the third robot gets 5/10/20/30.

[Jay O'Donnell]

I was actually planning on doing this during this week. Thanks for saving me a few days of work!

It seems that only the teams attending two regionals have a legitimate chance of doing really well. There are teams ranked below my own who I know are better than us, but we've attended two regionals, giving us a major boost.

[tkell274]

Originally Posted by JackS

1. I think the points should be bumped to 10/15/25/35 because NE does not award points for selection the same way that FiM and MAR do. This rewards teams for making eliminations more than it did before, but not quite as much
2. I think that the points should also be adjusted to be lower for the third robot on an alliance similarly to how selection points are distributed in FiM. Maybe the first two robots on an alliance get 10/15/25/35 while the third robot gets 5/10/20/30.

I agree with your idea that the points should be bumped up for eliminations to really help teams that earned their spot in the eliminations. But I disagree with your point on giving the third robot less points. The randomness of qualifications and the possibility of robots breaking sometimes leads to teams qualification score not matching the true power of their robot and therefore they should not be penalized for how they are picked.

[Kevin Leonard]

Originally Posted by tkell274

I agree with your idea that the points should be bumped up for eliminations to really help teams that earned their spot in the eliminations. But I disagree with your point on giving the third robot less points. The randomness of qualifications and the possibility of robots breaking sometimes leads to teams qualification score not matching the true power of their robot and therefore they should not be penalized for how they are picked.

I agree with tkell. A team that builds a support-based or defense based robot that is still phenomenal in what it is built to do (i.e. 4334 last year and 2789 this year) shouldn't be penalized for pursuing a different strategy than the powerhouse offensive teams. Third picks often make or break alliances, especially at deeper events.

[Chris is me]

I agree with the logic on making the third pick worth just as many points. The serpentine in particular makes this problematic - I would hate to be the 8th seed and have to pull up on my phone which team "needs" the 1st round pick points more and which team doesn't.

[2789_B_Garcia]

Originally Posted by Thunder910

I agree with tkell. A team that builds a support-based or defense based robot that is still phenomenal in what it is built to do (i.e. 4334 last year and 2789 this year) shouldn't be penalized for pursuing a different strategy than the powerhouse offensive teams. Third picks often make or break alliances, especially at deeper events.

Thanks! It's so hard to get love for defensive play...because of our limitations (personnel, funding, etc.) and sponsor issues, it's hard for us to crank out the robots we design and want. We're forced to compensate with scouting and strategy. I've been concerned about district discussions down here in Texas because of how formulas rate defensive teams, in particular because of how the dynamics for eliminations matches are very different than dynamics for quals matches, and our strategy definitely makes a bigger impact in elims. It's my hope that at the end of the month we'll be able to make the point that a creative and smart team can still make a strong impact on the outcome of matches even if you don't have the fanciest robot on the field...but needless to say, we are working on a few surprises for champs that will help us put points on the board

[Kims Robot]

Has anyone run the numbers with the FiM or MAR points system so that its easy to see the difference between the current NE proposal and the FiM or MAR models?

[Kims Robot]

To add to the dicussion I did a quick World OPR Rank VLOOKUP for the NE Teams (Thanks to Ed & Ether's data)

This is by no means enough for ranking into DCMP or WCMP, it just gives a screenshot of offensively which teams should maybe make it into the top NE slots... and is just meant for comparison (ie if there is a team in 10th OPR that doesnt make it into the top 50 of our ranking system, its worth looking at to make sure the NE Ranking structure is a good balance).

I do believe some of this will balance out by normalizing for a single event...

Note: the numbers here are the OPR Ranks for World OPR Ranking, NE OPR Ranking and Jack/Brian's NE Ranks sorted by OPR

EDIT - HUGE Apologies - I forgot Rhode Island!! Added now!

Team/ World OPR/ NE OPR/ NE Rank
177	42	1	19
125	45	2	1
2648	55	3	2
131	71	4	60
126	73	5	5
3467	75	6	4
230	76	7	9
1519	101	8	3
236	115	9	46
195	137	10	12
885	141	11	40
69	149	12	15
176	162	13	13
4564	166	14	51
58	176	15	30
1100	208	16	7
716	226	17	98
3464	229	18	90
1153	236	19	23
175	246	20	14
558	257	21	21
1991	260	22	18
2170	261	23	92
4473	287	24	43
78	303	25	28
133	309	26	25
172	316	27	11
2067	326	28	16
3525	329	29	29
1474	372	30	119
2871	378	31	95
228	408	32	17
157	416	33	37
1699	459	34	81
2168	460	35	6
4097	469	36	88
2349	473	37	71
2064	502	38	104
246	505	39	73
1784	509	40	76
3780	547	41	49
348	565	42	112
178	573	43	93
3718	577	44	102
1965	579	45	85
1277	585	46	63
181	586	47	47
121	596	48	57
213	601	49	113
1350	608	50	70
4555	612	51	136
4176	623	52	103
4812	626	53	77
138	643	54	54
319	649	55	59
151	658	56	61
3323	694	57	100
61	702	58	33
1071	709	59	32
3104	720	60	97
2876	730	61	82
3205	764	62	45
1073	769	63	35
1517	772	64	39
4041	782	65	122
1512	789	66	62
3609	791	67	8
4055	824	68	27
3466	837	69	68
2423	845	70	69
155	867	71	10
4557	880	72	86
4761	886	73	55
1922	897	74	53
1831	903	75	38
2877	921	76	31
97	923	77	83
3958	935	78	20
1761	940	79	84
2084	966	80	108
88	980	81	34
3930	991	82	24
2785	1045	83	147
23	1050	84	106
3146	1070	85	66
3597	1110	86	116
1768	1146	87	94
571	1159	88	41
4048	1171	89	127
238	1180	90	99
190	1190	91	50
839	1200	92	36
3451	1209	93	135
1735	1212	94	42
1058	1218	95	56
3566	1224	96	64
2370	1254	97	67
1757	1256	98	120
3280	1287	99	22
663	1323	100	101
2713	1329	101	121
4311	1358	102	114
173	1375	103	65
3182	1378	104	52
3634	1410	105	123
3236	1414	106	74
4410	1540	107	115
4546	1543	108	44
4042	1547	109	111
3555	1574	110	152
166	1594	111	140
4793	1680	112	117
3499	1695	113	79
1027	1743	114	58
237	1752	115	91
4628	1814	116	139
2523	1833	117	126
509	1840	118	48
3479	1862	119	96
1687	1871	120	75
95	1882	121	26
1740	1887	122	137
3654	1927	123	78
3585	1928	124	87
3461	1997	125	138
4151	2000	126	130
999	2009	127	72
4474	2034	128	131
4034	2052	129	142
2836	2080	130	105
1289	2090	131	110
2104	2124	132	132
1754	2138	133	143
1124	2158	134	146
2342	2166	135	118
4796	2173	136	144
529	2192	137	145
3927	2266	138	109
1729	2270	139	150
2262	2284	140	129
4609	2292	141	89
467	2296	142	134
3719	2335	143	154
2079	2347	144	107
1973	2371	145	80
1721	2374	146	151
1247	2377	147	124
1307	2396	148	141
2621	2398	149	133
4572	2410	150	148
501	2420	151	125
3623	2463	152	153
811	2470	153	149
1099	2483	154	128

[Nathan Streeter]

Originally Posted by JackS

I think that the points should also be adjusted to be lower for the third robot on an alliance similarly to how selection points are distributed in FiM. Maybe the first two robots on an alliance get 10/15/25/35 while the third robot gets 5/10/20/30.

I definitely do agree that the points for the 3rd robot should be adjusted... Primarily because that means that the 3rd robot on the winning alliance gets more points than any other alliance's robots (including the first two robots on all of the other alliances). Additionally, this method of assigning points gives very few points to the quarterfinalist alliances.

Although I suspect these points were removed in NE's proposal because it makes things a bit more complicated, I think assigning points based on alliance selection order as FiM and MAR do (16 to first 2 bots on Alliance 1; 15 to first 2 bots on Alliance 2; 14 to first 2 bots on Alliance 3... and 8 points to 3rd robot on Alliance 8, 7 points to 3rd robot on Alliance 7, 6 points to 3rd robot on Alliance 6) is the best way to assign points for eliminations (in addition to points based on finish).

Using 2013 GSR and 2013 Pine Tree as case studies (I chose these two because they're Week 1 vs Week 6, vary significantly in size, and Pine Tree is interesting because the red alliance won each matchup):

GSR Pick Order & Results:

610-4124-3609... W
138-131-58... QF
230-1991-1153... SF
885-1519-133... SF
151-229-1277... QF
1512-1922-1517... QF
61-175-172... F
2791-3467-78... QF

GSR Points (based on current NE proposal)

30-30-30... W
5-5-5... QF
10-10-10... SF
10-10-10... SF
5-5-5... QF
5-5-5... QF
20-20-20... F
5-5-5... QF

GSR Points (NE proposal + alliance selection points)

46-46-31... W
20-20-7... QF
24-24-13... SF
23-23-14... SF
17-17-10... QF
16-16-11... QF
30-30-27... F
14-14-13... QF

Pine Tree Pick Order & Results:

2648-3467-2386... W
176-125-63... F
1153-172-1831... SF
69-133-4564... SF
4473-58-1058... QF
78-1073-1922... QF
3930-4055-157... QF
3609-1071-181... QF

Pine Tree Points (based on current NE proposal)

30-30-30... W
20-20-20... F
10-10-10... SF
10-10-10... SF
5-5-5... QF
5-5-5... QF
5-5-5... QF
5-5-5... QF

Pine Tree Points (NE proposal + alliance selection points)

46-46-31... W
35-35-22... F
24-24-13... SF
23-23-14... SF
17-17-10... QF
16-16-11... QF
15-15-12... QF
14-14-13... QF

Seems like the current NE Proposal has several weaknesses:
- 1st and 2nd robots of each alliance get same reward as 3rd robot.
- Winners get 6x the points that the quarterfinalists get (3x the semifinalists).
- 1st and 2nd robots of finalist alliance (theoretically 3rd and 4th best teams) get 66% the points of the 3rd robot of the winning alliance (theoretically lower than 20th in ranking of teams).

These particular issues are improved with the inclusion of the alliance selection points. It'd be interesting to also add in the win-loss information... but I don't really have the time for that right now.

[JackS]

Originally Posted by Nathan Streeter

Although I suspect these points were removed in NE's proposal because it makes things a bit more complicated, I think assigning points based on alliance selection order as FiM and MAR do (16 to first 2 bots on Alliance 1; 15 to first 2 bots on Alliance 2; 14 to first 2 bots on Alliance 3... and 8 points to 3rd robot on Alliance 8, 7 points to 3rd robot on Alliance 7, 6 points to 3rd robot on Alliance 6) is the best way to assign points for eliminations (in addition to points based on finish).

The one thing I dislike about the FiM system assigning points based on Alliance Selection is that it gives points for essentially the same thing QF gives points for. The thing I do like is it is a good way to breakdown credit on an alliance.

A proposal I would support would be 8 ranking points for AC1 and First pick, and decreasing from there. This would give 3rd robots 0 extra points compared to the pack, but the difference between the last robot and the first is only 8 rather than 6. Then the NE Eliminations points could be bumped to 10/15/25/35 to make up for the point loss and emphasize results more.

Alliance Points:

8-8-0
7-7-0
6-6-0
5-5-0
4-4-0
3-3-0
2-2-0
1-1-0

Using BAE as an example: (10/15/25/35 + Alliance Points)

43-43-35... W
17-17-10... QF
21-21-15... SF
20-20-15... SF
14-14-10... QF
13-13-10... QF
27-27-25... F
11-11-10... QF