Here are the results:
--------- 3/31 -- 4/28
Alex A.---- 6:03 --- 5:38
Alex M.-- 5:45 --- 5:54
Brian ---- 7:00 --- 7:00
Cat----- 6:19 --- 6:11
Celeste -- 6:45
Chris H.-- 7:00
Dave ---- 5:08
Erin ----------------- 8:39
James ---- 6:15
Jennifer --------------7:25
Joselyne ---------- 7:01
Mary ---- 6:39 --- 6:39
Matt -------------7:20
Michele -- 8:05
Mindy ---- 7:58
Molly ---- 8:35 --- 8:20
Nick C. ------------- *
Nick M. --- 5:10 --- 4:58
Pablo ---- 6:29 --- 6:13
Pete ----- 5:22 ---- *
Shawn --- 6:35
Slush ---- 5:45 --- 6:20
Below is a scatter plot of last night's mile times vs. Towers PRs. Each point on the graph represents one person who ran last night with their mile time on the x-axis (horizontal) and their Towers PR on the y-axis (vertical). The line is the trend line that shows the average correlation between these two values for this group of people. Only 12 of last night's runners had Towers times and some of the times from last night are estimated due to lack of watch.
Click on the graph to see it full size
If your point falls above the trend line, it means that your mile time is fast proportional to your Towers Time (or you can read it as your Towers time is slow compared to your mile time). A perfect example of this is Nick M. Nick has amazing leg speed on flats and downhills compared to his time on Towers which is still pretty fast.
If your point falls below the trend line, it means your Towers time is fast compared to your mile time (or your mile time is slow). Slush was the perfect example of this last night. Mr. Slusher is a monster on Towers seemingly running faster and faster on each attempt; his mile time last night however was comparably average (hey, he tied with me).
If you look at the point between Nick M. and Scott you find Mr. Consistent himself, Pete, whose data almost defines the trend. What this means is that in relation to the group, Pete is equally strong and fast on both the Towers ascent and the Pineridge Mile.
With only 12 data points and some timing and weather irregularities, this plot is far from perfect, but it does provide an interesting way to assess your strengths and weaknesses in two very different running events.
The next running of the PDM will be Thursday, April 28. Come out and attempt to improve your mile time or set a new personal mile record.
Hurray for statistics!
ReplyDeleteOh man I was so bummed to miss this! Unfortunately, (as if my health luck could get worse) I developed a bad reaction to the medicine they gave me to help with the bad reaction from the other thing that messed up my face! My life is a farce. Anywho, I'll be there with my A game on the 28th...
ReplyDeleteOh, and Matt wants to know what the r squared of your regression line is.... :)
Cool stats. And that was a new mile PR for me. My run down the beer mile section of Towers last spring was 5:24.
ReplyDeleteJ: we missed you and your sprinter speed last night. I hope everything clears up and we see you back on the trails soon.
ReplyDeleteFor Matt, Mike and all the other math geeks out there, the equation resulting from the linear regression of the data is y = 6.88x - 3.38, which for everyone else out there who didn't run the mile with us means that if you take your Towers time and add 3.38 minutes (3 min 23 sec) and then divide by 6.88, you will get a prediction of your expected PDM time.
The r value or the correlation coefficient for the data is 0.88. The correlation coefficient indicates how well the line approximates the data. If |r| = 1, the line is a perfect fit to the data (all the dots fall exactly on a line; if |r| = 0, the line does not fit the data at all (dots all over the place). So our data with a correlation coefficient of 0.88 shows that our data is pretty closely correlated.
Pete: Awesome Pete and that was with a head wind and having run Towers in the morning. I think you are good for a sub 5 before HR.
Geez! All these statistics!
ReplyDeleteCouldn't we just assume that your Towers time and your Downhill Mile time are completely dependent on how much wine you drank the night before, how much wind was in your face, how much junk food you ate at work that day and whether your shoes were tied.
In other words....statistics are awesome, but mostly it's just luck. Nothing to do with talent or training :)
Kidding! Learned alot from the stats and where I need to focus!! THANKS!
Cat: Yes, there are definitely a lot more variables involved which is why it is interesting that the data is so closely correlated. And according to the data, you are one heck of a miler. Not bad for someone who at first mention of this event said, "I hate downhills"
ReplyDeleteLooking at the equation and some of the faster Towers runners, Mr. October should be able to run the PDM in 4:46 and Sam should be good for a 4:41. Hopefully we can get them out next time to test the equation.
It would be interesting to use something like the Reigel formula to predict Towers times from the PDM (or vice versa). The formula is relatively simple as T2 = T1 * (D2/D1)^1.06. Tn are times, Dn are distances. The exponent would have to be tweaked I'm sure to predict an difficult uphill run from a moderately downhill run. I don't know if we have enough data points to find a good fit though.
ReplyDeleteBunch of comments came in while I got distracted by my kids in the middle of typing mine. I like you're prediction of 4:46 for Mr. October. I'm going to hold him to that. The pressure is on, Nick.
ReplyDeleteThis is awesome! I have some work to do on the PDM. And the VBM for that matter.
ReplyDeleteWhat are you, a Math teacher? Greyrock at 4:30am for anyone interested. Ha!
ReplyDelete