Nov 22, 2011

Golden Cheeta Quickstart Guide Part II (stress score)

Earlier this year, I wrote the first part of the quickstart guide to Golden Cheetah, and at the time, vowed to return and write a separate post dedicated to stress scores and other metrics measuring the intensity of workouts. That time has come, and below, I will go over most things I feel you need to know to comfortably work with the performance manager (PM) metrics. However, this is not to be used as coaching advice; I'm simply helping you use the software and make some sense of the numbers and graphs it generates. What those graphs mean in terms of your performance is between you and your coach. At the end of this post, I will link to several references I found helpful in putting this together, and I encourage you to read all of them in their entirety to fully appreciate the numbers/graphs of the PM metrics and begin to understand how to incorporate them into your workout.

Uploading from Garmin

After talking to some people who've read the first part of this guide, I received requests to post instructions on how to upload your ride from a Garmin unit. Personally, I think the GC wiki has an excellent post (with video) on topic, but here are the instructions in brief. If this isn't clear, go to GC wiki and watch the video. 

1. Connect your Garmin unit to the computer and wait for your computer to recognize it as a drive. 

2. In GC, go to Ride -> Import from file. A Finder-type (I work on a Mac) window should pop up.

3. Click on the Garmin unit, which should appear as a drive, and navigate to activities. 

4. Select one or more activities to upload. 

5. Once GC validates the activity(ies), click "Save," and you're set to go. 

Overview of stress management

Why is it important to pay attention to stress scores, or to keep track of them in the first place? The short answer is that knowing this data helps you peak for key races and monitor your recovery, so you know you're actually recovering. However, the caveat to that answer is that the data you receive from GC is only as good as the data you enter. If you are diligent about uploading or manually entering (how-to below) your workouts into GC, you will have a much more accurate picture of your current state of stress. 

In this post, when I refer to "stress," I'm talking about the physical strain placed on the body as a result of cycling workouts, not the stress you may experience as part of life in general. There are also several types metrics referring to stress.

Chronic training load (CTL): I'm going to refer to this as "long term stress" because that's the nomenclature GC uses and I think it's a bit more intuitive. But when you go to do further research on topic (and you should), just know that CTL and LTS are interchangeable. LTS represents the long term effects of training on your body. As you go from base to build, you should see the LTS graph gently rise as the stress on your body increases. As we get into the graphs below, it will become clearer. By default, GC averages your data for the previous 42 days (6 weeks) to get the numbers for LTS.

Acute training load (ATL): For the same reasons listed above, I'll refer to this as "short term stress" (STS). This is a much narrower view of the stress you put on your body and a much better indicator of what you just did to your body with your workout(s). As you go through your first weeks of base, it will rise more acutely than your LTS. As you go through your recovery week, it will flatten out, and as you enter build, it should skyrocket. Again, this will be clearer as we start working with graphs below. 

Training stress balance (TSB): In the simplest terms, this is a number we get when we subtract ATL from CTL. If the number is negative, you're fatigued. If it is zero, you are neither fresh nor fatigued. If it is positive, you are refreshed. The key is to be refreshed, yet not lose fitness. And that's where you and your coach will have to play with numbers and determine how deep of a hole you need to be in short term (STS) and how positive you need to be for your peak (TSB). More on this below. 

One note on TSB is that terms "fatigued" and "refreshed" are relative. After your first rest week following three weeks of base/build, you might still be negative, but it should be a higher (smaller negative number) than when you started your rest week. 

I will also continuously use terms: base, build, peak and rest/recovery. If you are not clear on what they mean, I suggest picking up Joe Friel's The Cyclist's Training Bible, or in the alternative, browse his blog for explanations of periodized training.

Prepping Golden Cheetah

You are likely one of two types of cyclists: the type who has trained with WKO+/TrainingPeaks and now wishes to use Golden Cheetah for one reason or another; or  you've never trained using a performance manager before and you wish to begin monitoring your progress more "scientifically."

If you are in the former category, you have a few extra hoops to jump through to make sure the old data you import will be accurately represented in GC's PM metrics. However, both types of athletes will need to set up their athlete profile.

Once you have GC open, click Golden Cheetah -> Preferences (again, this is for Mac, on a PC I'm guessing it would be Tools -> Options, but send me a note if I'm wrong and I'll correct this). 

Once the preferences window opens, click "Athlete." Then you will navigate through your athlete profile to set up your power zones and your heart rate zones. 




If you are brand new to training, or you haven't tested in a while, you will need to test your fitness to figure out your LTHR and CP(FTP). You can either refer to Joe Friel's book, or if you happen to have Allen and Coggan's Training and Racing with a Power Meter, it too offers testing techniques to help you figure out those numbers. In case there's a question - YOU NEED THOSE NUMBERS, otherwise the data you'll get will be garbage.

To set up your power zones, you'll want to enter your CP on the right hand side. Next, and this is only for those of you who plan on importing previous workouts from WKO+/TrainingPeaks (or anywhere), make sure "From date" predates the first workout you plan to import. GC is set up to use only the workouts after the CP date to create the data in PM metrics. If you fail to change the date and import a ton of rides, when you go to see PM metrics, you'll have a nice flat line across the bottom of the graph for most things. Don't worry if you've imported and didn't change the date, you don't need to reimport your rides, you'll only need to set up your CP anew and delete the old one. Once you do that, GC will recalculate everything and you should see the data represented in the graphs. 

Once you enter your CP and the date, click "Add CP" and you should see a line item with your CP in the window below. Once you click on it, you will see the power zones automatically populate. You can then change them if you like. 



Setting up your HR zones is very similar to setting up your power zones. You'll need to know your resting HR, your max HR and your LTHR. You can get a good sense of your resting HR by taking measurements as you awaken every morning and averaging that number over a week. However, HR varies greatly based on illness, life stress, lack of sleep, a hard workout the day before, etc. So you should account for those things and take your resting HR on a "normal" week, whatever that means for you. To get your max HR and your LTHR, you will need to test. I recommend the same two references as I did above. 

As with your CP, when you enter your LTHR, don't forget to backdate if you'll be importing data. This date should also be set to sometime before the first workout you plan to import. Now just one more thing before we dive into the good stuff.



Manually entering workouts


As I mentioned above, GC only provides you with as good of an indication of your stress balance as the data you put in. Sometimes you go on a ride and don't have your HR monitor, or ride without a powermeter, or you go to an indoor cycling class and can't upload the data to your computer. GC gives you an option to enter a workout manually.

Go to Ride -> Manual ride entry, and the window on the right should pop up. There is apparently a very minor coding glitch in GC, but when you first open this window, the fields next to "hours," "mins" and "secs" are tiny slivers. You need to expand the window by dragging one of the corners to expand the fields.  Then just fill in as many fields as you have information for and click "OK." You notice that the number of fields in this window is rather limited, and you may have way more data at your disposal. Don't fear, you can enter all of that data in the next step. 

Once you manually enter your ride and save it, go to your ride summary from the home screen and select the ride you just entered. Below your ride summary, you will see three tabs: workout, notes and metrics. Clicking on "metrics" gives you the option to fill in those fields that were lacking in the previous window. As an example, below is my 1.5 hour workout at M2. I know my time on the bike, my average watts for the workout, the KJs and my average HR. To enter those values, simply click the checkbox next to them and a field appears. Note that BikeScore and Daniels Points were filled in automatically, as those were the numbers automatically calculated in the previous step. 



Now we're ready to dive into the good stuff!

BikeScore


BikeScore is a metric developed by Dr. Philip Friere Skiba and it gives a numeric value to your workout. A high score means a harder workout, a low score means an easier workout. Here's an explanation of how the number is calculated if you really want to get mathematical. 



To access your BikeScore, as well as everything we'll be looking at below, you want to click on the "Metrics" tab from the home screen. I'm only going to be using my data from "this month" because I had a period of rest between training where I wasn't really diligent about logging stuff in (shame on me), so I just want to eliminate extra noise in the data. Looking at the above graph, you can see the BikeScore for my rides. The longer rides on the weekend have the higher score than the shorter rides during the week. This is a period of base for me, so the score is primarily a function of time because I do my best to stay in zones 1 and 2 when riding. However, a short interval workout can potentially have a higher score than a longer tempo or recovery ride. The values over the bars in the graph are mine, you won't see those on the graph in GC, but if you scroll to over the bar, you will see the exact values. 

BikeScore gives you an indication of how hard you worked that day relative to your other workouts for a given time period. You can control that time period in Settings.

To set up the time period over which your score is calculated, go to Preferences and click on "Settings." In "Basic Settings" you want to change the value BikeScore uses to calculate your score. I have mine set on 28 days (4 weeks). In short, BikeScore numerically answers the question, "how hard did I ride today?"


Relative intensity


Similar to BikeScore, relative intensity numerically answer the question of "how hard did I work?" Unlike, BikeScore, however, RI is a simple mathematical formula I can wrap my mind around without getting the MEGO syndrome (my eyes glaze over). RI = xPower/CP(FTP). So RI is simply a proportion of your normalized power (discussed in Part I) to your FTP, or as I've been referring to it here, CP. As your fitness improves or declines, it's important to continue testing and update your CP to make sure this and other metrics are true representations what you're seeing. 


Performance management metrics

Now, finally, we get to talk about stress scores and stress balance. GC offers three separate PM metrics you can use, Daniels, Skiba and TRIMP. I will go over all three in this blog, but my advice to you, pick one and stick with it. By the end of this section, you should be able to determine which metric will work best for you. All of them show similar data in slightly different ways and use different mathematical formulas to arrive at the numbers, but because all of that data is relative, as long as you don't try plotting Daniels STS vs. TRIPM LTS and making sense of it, you should be fine. I also find it very helpful to look at several metrics plotted on one graph, so I can see the relationship among them.

Looking at Daniels STS, Skiba STS and TRIMP STS plotted on one graph, you can see that the trend lines have a similarly trending slope even when the stress values are different due to the different formulas. 




Let's look at this in detail. While the general similarities are visually obvious, it is more interesting to look at the differences. Note the difference in the first three plot points between TRIPM and Skiba. Namely, the second plot point is missing from TRIMP. Why is this? Well, the missing plot point on Nov. 2 is from my indoor workout and I didn't look at my average watts before I left. For TRIMP, if there are no watts, there is no plot point. Which may not make it the optimal metric for you if you sometimes ride without a power meter. While I try to do all of my training with a power meter, sometimes I race without one, or as above, I forget to record my average watts before leaving the cycling studio. Thanks to a comment by one of my readers, I realized I had this completely backwards. TRIMP uses HR primarily to calculate the intensity of the workout. BikeScore, on the other hand uses your xPower to calculate intensity. So why is that plot point missing? This appears to be a GC issue. I noticed that when I upload manually, there is no zone indicated in my ride summary, even if I enter my average watts and HR manually. I have a suspicion that TRIMP needs HR in zones to run the calculation, while BikeScore simply needs a watts entry. So unless you're uploading your ride from somewhere, you're not going to get a TRIMP calculation in GC. If anyone has any insight on this, I'd greatly appreciate it!

The next thing to pay attention to is the difference in trend lines among the three metrics. Despite some missing points, TRIMP and Skiba trend closely. They look identical, but because I eyeballed it, they are likely slightly different. However, Daniels is clearly a shallower slope. Why is this? Well, unlike TRIMP and Skiba, Daniels assigns difficulty to your workouts in a linear manner. Meaning that two one-hour workouts equal one two-hour workout, assuming the same intensity. Additionally, it was originally developed by Jack Daniels (no joke) for runners and has been adapted in GC to cycling. 

The difference is clear if we examine TRIMP and Daniels side-by-side (the bars are TRIMP and lines are Daniels):




Let's look at the last three days on the graph and run some numbers. First let's compare the relative intensities for Nov. 18 and 19 for Daniels and TRIMP. On Nov. 18, I rode for 2:58 at 188 xPower, and on Nov. 19, I rode 5:13 at 191 xPower. The ratio of relative intensity for Daniels of Nov. 18 to Nov. 19 is .48. So my Friday ride was 52 percent less intense than my Saturday ride. However, according to TRIMP, the same ratio is .45, so the Friday ride was 55 percent less intense. That's pretty close. This gets even more interesting, however, as we compare Friday (18th) with Monday (20th). On Monday, I rode for 3:47 at xPower of 172.

The TRIMP score is higher on Monday, but the Daniels score is lower. The Daniels ratio for Monday as compared to Friday is 1.07. So Friday's ride was 7 percent harder than Monday's. The same calculation for TRIMP yields .94, making Friday's ride 6 percent easier. If you're keeping score, that's a 13 percent difference! So when did I work harder? I used BikeScore to settle the argument, and it appears that TRIMP and BikeScore are in agreement, 3:47 at 172 is in fact a harder workout than 2:58 at 188. Remember that BikeScore is the Skiba equivalent of TRIMP points and Daniels points. 

[This paragraph was edited from the original version to coincide with the changes above.] From the above, you can probably guess which system I use primarily to track my stress - Skiba. From personal experience, I know that two hours at 200 watts are more than twice as hard than one hour at 200 watts, so Daniels just doesn't make sense from that standpoint. While I sometimes ride/race without my heavy Powertap wheel, I can use estimated watts to get a relatively close BikeScore. As mentioned above, due to an issue between GC and Garmin, TRIMP points don't calculate if you enter your ride manually, and I haven't yet been able to figure out how to resolve that issue. Again, any insight here would be very helpful.

Now let's check out the PM metrics that help track overall stress in all three systems.

Daniels metrics

Below is a plot of my Daniels LTS, STS and TSB.



As I mentioned above, you can see that STS rises more acutely than LTS, and the TSB line appears to be the exact opposite of STS. TSB is the difference of LTS and STS, so as you can see on the graph, when STS and LTS intersect, TSB equals zero - a point where you are neither fresh nor exhausted. I've selected the view of "This Month," but obviously there are a few days left this month and my last ride was on Nov. 20. GC projects STS, LTS and TSB to the end of the month as if I will do nothing for the remainder of the month. Extending the graph thusly, you can see that a few days off will have my TSB in positive numbers. In practice, after the build period(s), there will be a peak period where my training will be tapered and my STS will drop, but not as smoothly or acutely as in the graph because I will still ride, just at a lesser intensity or lower volume - it will be a jagged plot trending downward. How positive your TSB has to be for you to be refreshed yet maintain maximum fitness is extremely personal.

Skiba metrics

This is how the same data is represented when plotted with Skiba metrics. 




As you can see, the lines of LTS and STS are similarly plotted in that they are lines, but the TSB is plotted as a bar graph. However, the metrics should be interpreted the same way. When LTS and STS intersect, TSB equals zero, and the remainder of my explanation in Daniels (above) applies equally here.

TRIMP metrics

TRIMP metrics, when plotted, look very similar to Skiba.




As you can see, I can just copy+paste the paragraph from the Skiba section, but I trust in your abilities to scroll up and substitute "TRIMP" for "Skiba."

Now what?

That is exactly the question I would be asking if I read the above post and it ended with the above paragraph. I've crunched some numbers, described some metrics and gave you a few how-to tips. But what you really want to know is how all of this can make you a better cyclists. In one paragraph below, I'm going to try to digest everything above into some advice you can take on and off the road as you gear up for season 2012.

Upload as much data as possible. If you have no data at all, but are experienced, you can probably estimate. Even if you estimate on the lower end, it will be better than having a rest day recorded if you record nothing at all. You now know how to upload from your Garmin and how to enter data manually, so there is no excuse. Monitor your training. As you go from week to week in base, you should see your STS rise acutely overall and your LTS rise gradually, with your TSB getting more negative. After a rest week, your TSB should be higher (a lower negative number) than it was at the end of your last training week. If it's not, you're working too hard during your recovery weeks and are not letting your body absorb all of the fitness you've gained. As you head into the build phase of your training, STS will rise even more acutely and LTS will continue to rise gradually, with TSB falling. You should see the same trends in your rest weeks in build as you did in base. Once you start peaking/tapering, your STS will gradually drop, and LTS will drop but even more gradually. Eventually, STS will catch LTS and your TSB will start trending into positive numbers. This is the point where you will likely have some of your best performances, but exactly how positive you have to be is up to you and your coach. Lastly, pick a metric that works best with the type of training you do and equipment you have and stick with it. 

References and helpful links

Here are some references and helpful links to help you do further research, or go deeper where I merely gleaned over.





11 comments:

  1. I think you have a misspoken in the following statement: "For TRIMP, if there are no watts, there is no plot point." TRIMP is HR based metric. Perhaps you didn't have your HR strap on that day?

    ReplyDelete
  2. Thank you. You are absolutely correct! I had TRIMP and BikeScore backwards, and your comment also made me go back and look at GC closer to see why there was a difference in what points were plotted and which were not. I've updated the blog accordingly. Wish I could give you credit for spotting the error, but you're Anonymous :).

    ReplyDelete
  3. Man its seems like a rocket science. Going by feel is a way to train as well.

    ReplyDelete
  4. The calculations can get a bit complex, especially when they are used together on one graph, but you're right, feel has a lot to do with it. You can measure you performance on the bike quite scientifically with the right equipment, but it's still important to keep track of things like sleep quality, diet, things happening in your personal is just another useful tool.

    ReplyDelete
  5. Is there a aggregator tool that takes all the data and makes suggestion. I have read somewhere that you can determine your physical ability per day by doing couple simple test (like maximum taps on keyboard per interval, etc) and assign a number (0-10) which represents your ability to perform.

    ReplyDelete
  6. Unfortunately, there is no tool that aggregates all the data (objective and subjective), however, there are a number of systems that use subjective ratings converted into numbers in an effort to manage stress on the system. There is also the heart rate tracking method, where deviations from the norm usually mean exhaustion or illness.

    ReplyDelete
  7. This comment has been removed by the author.

    ReplyDelete
  8. Thanks again Vitaly for this very detailed guide to Golden Cheetah. I uploaded my training for the last 6 weeks into GC, and then checked my stress balances for this month, to see (in theory) how long it would take me to peak for a race. Even though the metrics had different magnitudes, they were consistent about the effects of resting: After 4-5 days I have a non-negative balance, and after 8-9 days I see diminishing returns from resting: https://img.skitch.com/20111216-pp9krj4cp6hr9d2pbwgcgj5u52.jpg.

    I'm planning to track all 3 stress balances on an ongoing basis to see if any notable differences emerge. If nothing else, it makes me feel good about taking more rest days :)

    ReplyDelete
  9. Hey, Jesse,

    Thanks for reading. As I mentioned above, don't confuse positive stress balance with peaking - those things while related are absolutely not the same thing. For example, I could work out for 2 hours a day 5 days a week doing very easy rides, then rest for three days and get into positive stress. But am I peaked? No. Peaking comes as a period after base, build and is part of a proper taper for an A target race. It's up to you (or your coach) to determine how much base and build you need and what types of systems you should be working to improve and peak. In the grand scheme of things, it's better to follow a well-established plan - like Joe Friel's. In the microcosm that is you, it's about trial and error and figuring what works and what doesn't as you continue racing.

    ReplyDelete
  10. We can't avoid stress that's why we need to find things to divert it. That's how my friends from college counseling services explained to me about working out for stress management.

    ReplyDelete
  11. Thanks for posting this information. Preparing to acquire a power meter I found GC a little lacking on how to set it up. You've answered a number of the questions I had. Your blog has been a big help.

    ReplyDelete