2000 US Ski-O Championships & US Ski-O Team Trials

Garnet Hill Cross Country Ski Center, North River, NY

22-23 January 2000
(postponed from 8-9 January 2000)

hosted by Empire Orienteering Club

Course Maps

• Garnet Hill - Long Blue 1
• Garnet Hill - Long Blue 2
• Garnet Hill - Long Red
• Garnet Hill - Long Orange
• Garnet Hill - Short Blue
• Garnet Hill - Short Red
• Garnet Hill - Short Orange

Road or Trail Analysis

Ernst Linder, US Ski-O Team Executive Steering Committee Chair, offers an analysis of the "road or trail" choice on the Long Blue course:

Since Phil kindly posted maps of the team trials, I decided to do a statistical modeling of the Long course's trail or road route choice question. Those of you familiar with regression analysis and/or Analysis of Variance will be able to make sense of the reported numbers.

In a nutshell: I try to predict the finish time (variable "time-60") from the split time (variable "split-60" ). I subtracted 60 minutes from both times prior to analysis. I then tried to do this separately for each group: Those who took the road (0 - group) and those who took the trail (1 - group). It turns out that the model produced an excellent fit with nearly parallel lines and a shift of about 6.6 minutes. I then pursued the "parallel lines" model which resulted in an extra time of 6.64 minutes for those racers who took the trail. So, the model predicts the following finish times from the split times: Road: Fin. time = 60 + 13 + 1.44 (Split - 60) Trail: Fin. time = 60 + 6.64 + 13 + 1.44 (Split - 60)

The last graph shows the "residuals" from this analysis: There are two racers whose times were not well predicted by the model 1: Gary Brackett (Large positive residual: had a longer time between Split and finish than predicted by model) 2: Carl Fey (large negative residual: had a shorter time between split and finish than predicted by the model)

Hope this is instructive. Happy Skiing!

---

Scatterplot: Time-60 minutes over Split-60 minutes	Whole-Model Test Residual Plot (B: Brackett, F: Fey)

---

Response: Time-60
Summary of Fit			Parameter Estimates
RSquare	0.977389		Term	Estimate	Std Error	t Ratio	Prob>|t|
Root Mean Square Error	2.414419		Intercept	13.052104	2.156966	6.05	<.0001
Mean of Response	54		Split-60	1.4407321	0.066158	21.78	<.0001
Observations	14		Route	6.6457845	1.355999	4.90	0.0005

---

Effect Test
Source	Nparm	DF	Sum of Squares	F Ratio	Prob>F
Split-60	1	1	2764.5847	474.2470	<.0001
Route	1	1	140.0229	24.0200	0.0005

last updated: --Wed May 19 2004 09:35:29 PM EDT--