Pace Prediction Spreadsheet

[El Caballo]

Recently I stumbled across an old spreadsheet I had put together about 5 or 6 years ago. I had made it available through the old old C2 forum and was curious if anybody was still using it. After searching both this forum and the UK forum, it is obvious that nobody knows about it--and several people have created similar spreadsheets. I once again am making my spreadsheet available.

My spreadsheet, http://www.wakeley.us/rowing/pace_prediction6.xls, has six prediction formulas.

1. Paul's Law. This a simple formula of "Double the distance, add 5 seconds to the pace." Supposedly, it works best if you have a good balance between strength and cardiovascular fitness. From posts on various forums it seems many people find the 5 second difference to be too large--especially those who focus on the longer distances. An advantage is that you can predict other paces from just a single entry.
2. A best linear fit of pace vs. ln(distance). This is a commonly used but generally tends to predict too slow a pace at short distances (and long distance for those who train for longer distances).
3. An alternative to formula 2 that guarantees a larger slope of the pace vs. ln(distance) plot at short distances and a flatter slope longer distances. I find this works quite well for me (at least when I am in shape).
4. A power formula that is guaranteed to have a higher slope at longer distances. This is seldom accurate, but I threw it in because one forum poster said that a plot of his PB paces vs ln(distance) had a higher slope at longer distances.
5. A variation of formula 4 that adds another term, making it a second order fit. When you have lots of times/distances entered that are good solid PB's, it can be quite accurate, otherwise it can be very poor. This formula requires a minimum of 3 times or distances.
6. Another second order fit, this time a variation of formula 2. Similar to 5, this can be quite accurate (or sometimes very poor) and requires a minimum of 3 times or distances.

If you download the spreadsheet, you will see some of my old times and distances from when I was doing a lot more rowing than I am now. I put them there as an example. You will want to delete those times and put in your own.

I am planning to make some updates to this spreadsheet. Currently, it has all Nonathlon distances (I probably created this spreadsheet about the same time the Nonathlon was created). I plan to add a 4 minute time and a couple of blank times and distances so you can add any other PB time and distance (such as 2500 m or 1 mile). Another update will be to add average Watts to the tables. I usually row with Watts showing on my PM2+ because I find it much easier to control my pace. I also plan to add date and comment columns in the entry table. I expect I'll move the formula listing somewhere else or just get rid of it (or just leave it on the plot).

Any other suggestions for changes? Any comments?

[Citroen]

Try this: http://www.freespiritsrowing.com/conten ... -predictor

[El Caballo]

I've made quite a few updates to my spreadsheet. The new version is at http://www.wakeley.us/rowing/new_pace_prediction.xls. Among the changes:

1) Allows you to plot either pace vs. ln(distance) or power vs. ln(distance). The prediction 4 formula is calculated using either ln(pace) vs. ln(distance) or ln(power) vs. ln(distance). These are equivalent formulas, they just use different constants (easily calculable, one from the other). When plotting power the power vs ln(distance) formula is displayed. All the other predictions are based on pace vs. ln(distance), so that is the only formula ever displayed.

2) 4 min time added

3) Room for 5 non-standard distances or times (put either distance or time to see the prediction in the tables--put both distance and time to use the non-standard distance as input for the predictions and specify time or distance to be fixed in the prediction.

4) Paul's Law is a fixed 5sec/500m slower pace for double the distance. Many people have found that their "doubling factor" is different--older and lightweight rowers especially lean toward a lower factor (based on my experience and what other people have posted in various forums). I've added an option to set a different "doubling factor" when computing the "Paul's Law" prediction (which means it really isn't Paul's Law any more).

5) Date field added to go with your inputs. Currently it isn't used in the prediction, but it does help you easily see which of your inputs is old. A possible future enhancement is to weight the inputs by how recent they are--the more recent times and distances entered affecting the prediction to a greater extent.

6) An average power field added to the tables.

7) To make room for the extra data, I replaced the six prediction tables with 3 tables with selection lists as to which prediction tables to display. Formulas are shown for each prediction below the corresponding table.

Possible future enhancements (if anyone is interested--please let me know if you are):

1) Adding another input as to quality of the row. An outstanding time or distance (for you), could be given a greater weight in the prediction calculations, with a relatively poor row given a low weight.

2) As indicated above, the input time could be weighted by how recent the times or distances were performed.

3) Add a comments field in the input table, giving you room on the spreadsheets to add some notes about the row. To do this, I would probably have to remove another of the prediction tables (leaving two).

4) When power is selected for the plot, also make the predictions using power vs. ln(distance) rather than just plotting the calculated power based on the pace prediction (pace would then be calculated based on the predicted power.

5) Add some more information in the spreadsheet about the prediction formulas. This would include URLs to forum topics or web sites. (I also want to add relevant links to forum topics here in this thread--don't wait for me, post them yourself if you know of any.)

Dougie, I looked at the pace-predictor on freespiritsrowing. Except for the 500 meter fudge factor used there, and for the fact that it only accepts 2 inputs, it is the same as prediction 4 in my spreadsheet. For someone who trains only for 2K it is probably pretty good, but it generally underestimates PB power at longer distances (based on my experience and C2 ranking times and distances others have posted). Also, unlike predictions 3 and possibly 5 and 6 in my spreadsheet, a fudge factor is needed for the 500 meter distance (this could easily be done mentally by just subtracting two seconds from the predicted time in the prediction 4 table).

[michaelb]

Thanks, Bill. I have your original spreadsheet from many years ago saved here somewhere on my computer, and have posted it from time to time. My "issue" with it though is that it gives 5 possible predictions, that seem to contradict each other and don't resolve the question for all of us collectively or for any of us individually which one fits best.