I never wanted to be a statistician. I got my wish. I'm a writer. As a result, all of the following will demonstrate I know nothing about numbers.
One runner I know ran last year's Freedom Four 4-miler in 25:47. Then, he ran in last September an 18:38 5K at the same race I had my current PR of 22:00.84. I wondered (and still wonder) if I can take anything from this.
I ran the numbers. In the 5K, he ran each mile at 0.932502033 of each mile of the 4-miler. Some of that relates to the hilly course, the distance, how he felt that day, and so on. Some, I am hoping, can be attributed to training throughout the summer.
Taking my official 4-mile time and multiplying it by 0.932502033, I get what comes down to a 21:18 5K (6:52/mile). Not good enough. 20:00 is the goal.
That runner is very consistent with his training, and has been in good shape for years. Maybe his time will reflect less improvement than mine will. My consistency is new, and so I might still be moving forward in my "leaps and bounds" of improvement. He ran a 26:25.69 last year, while I ran 33:05. His improvement: 0:38. Mine: 3:36.
The sum of it all? The data is not good enough to correlate any likely sub-20 PR by the summer's end. I am expecting to break 22:00 my next 5K if everything keeps going as it has, ideally getting beyond these silly questions.