View Single Post
  #122   Spotlight this post!  
Unread 22-04-2013, 20:47
MikeE's Avatar
MikeE MikeE is offline
Wrecking nice beaches since 1990
no team (Volunteer)
Team Role: Engineer
 
Join Date: Nov 2008
Rookie Year: 2008
Location: New England -> Alaska
Posts: 381
MikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond reputeMikeE has a reputation beyond repute
Re: 2013 Championship Newton Division

Quote:
Originally Posted by JesseK
Anyone feel like doing a Monte Carlo, should the schedule change? Simply swapping 4-6 teams per simulation and recalculating scores may be enough for more accurate predictions overall. I've been able to do one simulation manually, and the point spreads are close enough with the 2 schedules that I have that we could potentially go 6-2 instead of 3-5 (given our expected increased 24-pt teleop disc scoring ... well, that's my expectation...).
OK, so I bit and generated 10 different Newton schedules using an evaluation version of the scheduling software.
Match scores are based on Ed Law's Max OPR data for each team and Ether's world Auto/Climb/Teleop OPR data (linearly scaled to match Ed's Max OPR values).
This gives 10 sets of qualification rankings, one for each of the simulated schedules.

Here's a summary of each team's average, best, worst and "in top 8" rankings across those 10 schedules:

Code:
Team	Rank	Best	Worst	Top8?
25	50.2	35	74	0
68	24.7	2	43	1
79	35.1	18	63	0
88	70.2	36	96	0
93	57.5	25	81	0
116	34.8	15	62	0
122	90.4	83	100	0
128	30.2	8	51	1
131	35.8	12	71	0
141	26.8	11	55	0
155	61.0	24	91	0
175	36.8	1	77	1
180	14.4	1	33	5
190	75.0	54	92	0
191	55.3	19	93	0
195	18.8	2	47	3
217	37.9	2	69	1
225	12.3	2	30	5
296	82.0	68	98	0
303	39.3	7	69	1
341	36.3	3	100	2
353	67.9	30	91	0
492	68.0	38	95	0
503	23.0	8	60	1
537	55.6	28	71	0
604	80.7	22	96	0
771	77.5	39	97	0
829	56.1	33	85	0
1111	83.6	70	98	0
1208	42.6	5	69	1
1251	73.5	40	98	0
1305	36.9	16	80	0
1311	68.2	19	96	0
1410	30.9	1	66	1
1511	75.7	54	98	0
1515	63.7	35	94	0
1523	25.2	3	49	2
1538	8.2	2	24	6
1569	38.5	7	81	1
1640	40.5	21	74	0
1646	67.7	48	95	0
1671	56.9	16	95	0
1676	9.2	1	20	6
1718	17.7	2	38	2
1730	44.2	12	99	0
1739	74.3	35	99	0
1741	20.4	10	42	0
1781	35.1	3	64	1
1885	31.9	5	80	2
1899	33.2	9	67	0
1985	36.7	13	77	0
1986	9.8	1	56	7
2052	21.9	5	79	2
2054	9.8	1	26	5
2080	74.9	37	88	0
2194	48.9	27	71	0
2252	40.7	14	79	0
2283	81.2	49	97	0
2352	45.2	8	82	1
2386	64.2	46	93	0
2389	78.9	44	100	0
2439	41.0	22	70	0
2471	20.4	6	77	3
2612	20.7	1	44	3
2619	19.7	6	59	4
2655	90.0	61	100	0
2783	63.9	28	94	0
2789	62.2	27	94	0
2826	20.2	1	57	2
2851	35.0	2	78	1
2974	40.5	18	79	0
3130	49.7	26	78	0
3137	44.8	19	78	0
3171	83.1	55	100	0
3244	71.6	51	86	0
3280	67.4	29	96	0
3339	60.8	9	94	0
3476	22.5	4	45	3
3574	63.5	43	80	0
3756	47.9	18	81	0
3802	19.5	9	42	0
3880	44.5	8	99	1
3931	37.6	6	68	1
3997	32.0	4	60	2
4276	76.4	31	98	0
4451	51.1	13	93	0
4476	68.0	27	97	0
4487	57.6	23	85	0
4505	69.9	43	96	0
4519	44.1	4	66	1
4531	81.0	26	98	0
4539	61.0	7	90	1
4572	89.0	55	100	0
4586	75.3	47	97	0
4629	79.1	57	99	0
4652	85.8	72	99	0
4656	63.7	38	90	0
4666	87.3	57	99	0
4713	81.5	54	93	0
4828	76.8	36	100	0
(The average rank correlation between schedules is 0.59, but between schedules and max OPR the rank correlation is 0.74)

Picking the most extreme examples, team 341 would rank either 3rd or last, based purely on their randomly chosen alliance partners and opponents.

So I think we can conclude that there is a fairly high degree of scheduling luck, just as we expected.