How to Properly Dyno Test a Modern Porsche
11-15-2007, 12:57 PM
#1
Instructor
Thread Starter
How to Properly Dyno Test a Modern Porsche
I posted this response in another thread on this forum, about our newly released 997 GT3 exhaust, but I feel it deserves its own thread heading for archival purposes:
I'd like to take this chance to educate the forum on the "ins and outs" of properly dyno testing a modern Porsche. There are *many* tricks to ensure valid data, and unless the dyno shop you are using is a hardcore Porsche power tuner, they will most likely not have any idea of what is involved. For example, dyno testing a powerful 'Vette is much different from dyno testing a Porsche, believe it or not.
To successfully dyno test a modern Porsche, the following steps must be taken to ensure valid results:
High VELOCITY fans must be placed in close proximity to the front radiators and also pointing into the open engine bay. Large diameter fans are not high velocity, they are high volume, but move air at ~15mph max. The fans we use have been verified by an anemometer to produce 60mph+ air up to 12" from their outlet. They are small carpet dryer type fans and we use several of them when testing. Do NOT place the engine fan with air blowing directly into the airbox as that will cause turbulence at the MAF, which will corrupt results.
Allow at least 2 minutes cooldown between runs to allow intake and coolant temps to stabilize. We typically run the car at part throttle during cooldown to allow the waterpump to circulate water faster.
The above two steps are CRITICAL in order to maintain stable intake air and coolant temperatures. The fuel injection ECUs on modern Porsches have multiple ignition timing maps, and depending on slight variations in these temperatures, the ECU may jump maps, resulting in useless data.
We have found that it can take up to 8 runs per test session to yield stable and repeatable power results when using the above techniques. Usually, the power climbs as the engine heats up, as the ECU is usually on a lower timing map during warmup. If the power DECREASES during testing, then you have overheated the car and caused the ECU to jump timing maps. Doing this many runs will also allow any adaptations to occur from mods that are being tested. We like to throw in a key cycle or two between runs to accelerate the adaptation routines. Only once we see power results stabilize do we consider our testing to be valid for that session, and all other runs are thrown out.
By using the above test protocols, you ensure valid results. Doing two or three back to back runs with large, slow moving fans pointed at the front or rear of the car is a sure recipe for invalid results.
Further, we always prefer to do before and after testing within the same day to minimize ambient weather fluctuations and the effects they have on results. Even the best SAE correction factors are not 100% accurate when ambient temperature, humidity, and barometric pressure has radically changed from day to day. Our final power product testing is done on the same day by reinstalling all stock components, retesting for a baseline, and then reinstalling the final performance versions and retesting.
To put things into perspective on how difficult it was to produce valid test results during our 997 GT3 exhaust development, we did 159 runs total to accumulate useful data. Power product development on these modern Porsches is not for the faint of heart or casual of approach.
Any questions, let me know.
I'd like to take this chance to educate the forum on the "ins and outs" of properly dyno testing a modern Porsche. There are *many* tricks to ensure valid data, and unless the dyno shop you are using is a hardcore Porsche power tuner, they will most likely not have any idea of what is involved. For example, dyno testing a powerful 'Vette is much different from dyno testing a Porsche, believe it or not.
To successfully dyno test a modern Porsche, the following steps must be taken to ensure valid results:
High VELOCITY fans must be placed in close proximity to the front radiators and also pointing into the open engine bay. Large diameter fans are not high velocity, they are high volume, but move air at ~15mph max. The fans we use have been verified by an anemometer to produce 60mph+ air up to 12" from their outlet. They are small carpet dryer type fans and we use several of them when testing. Do NOT place the engine fan with air blowing directly into the airbox as that will cause turbulence at the MAF, which will corrupt results.
Allow at least 2 minutes cooldown between runs to allow intake and coolant temps to stabilize. We typically run the car at part throttle during cooldown to allow the waterpump to circulate water faster.
The above two steps are CRITICAL in order to maintain stable intake air and coolant temperatures. The fuel injection ECUs on modern Porsches have multiple ignition timing maps, and depending on slight variations in these temperatures, the ECU may jump maps, resulting in useless data.
We have found that it can take up to 8 runs per test session to yield stable and repeatable power results when using the above techniques. Usually, the power climbs as the engine heats up, as the ECU is usually on a lower timing map during warmup. If the power DECREASES during testing, then you have overheated the car and caused the ECU to jump timing maps. Doing this many runs will also allow any adaptations to occur from mods that are being tested. We like to throw in a key cycle or two between runs to accelerate the adaptation routines. Only once we see power results stabilize do we consider our testing to be valid for that session, and all other runs are thrown out.
By using the above test protocols, you ensure valid results. Doing two or three back to back runs with large, slow moving fans pointed at the front or rear of the car is a sure recipe for invalid results.
Further, we always prefer to do before and after testing within the same day to minimize ambient weather fluctuations and the effects they have on results. Even the best SAE correction factors are not 100% accurate when ambient temperature, humidity, and barometric pressure has radically changed from day to day. Our final power product testing is done on the same day by reinstalling all stock components, retesting for a baseline, and then reinstalling the final performance versions and retesting.
To put things into perspective on how difficult it was to produce valid test results during our 997 GT3 exhaust development, we did 159 runs total to accumulate useful data. Power product development on these modern Porsches is not for the faint of heart or casual of approach.
Any questions, let me know.
11-15-2007, 02:27 PM
#2
Unfortunately, unless you have a Statstitician on staff you will never provide results that are statitstically significant.
Firstly you would have to calculate how large the sample size would need to be to give you the significant results (+ -5%), and that would partly entail running 100's of dyno runs in identical atmospheres WITH and WITHOUT the modifications.
Right now the fact remains that any GT3 dyno run WITHOUT your modification - or anyone elses for that matter may produce the same results.
Right now the only fact you've proved is that the GT3 is capable of putting out more HP than the manufacturer states - for whatever reason. That's it. It may or may not be attributed to anything you did to the car.
Firstly you would have to calculate how large the sample size would need to be to give you the significant results (+ -5%), and that would partly entail running 100's of dyno runs in identical atmospheres WITH and WITHOUT the modifications.
Right now the fact remains that any GT3 dyno run WITHOUT your modification - or anyone elses for that matter may produce the same results.
Right now the only fact you've proved is that the GT3 is capable of putting out more HP than the manufacturer states - for whatever reason. That's it. It may or may not be attributed to anything you did to the car.
11-15-2007, 02:35 PM
#3
Instructor
Thread Starter
I respectfully disagree 100%. What you suggest is purely theoretical.
What I describe is based 100% on experience. Our methodology has been honed over literally thousands of runs on many, many cars. We are not chasing ghosts.
The procedures I described above has nothing to do with modifications we did to the GT3 per se. It applies to even getting an accurate baseline on a stock vehicle. Without the same controls I describe above, you *will* see wildly varying, and thus invalid, results, as we did early on in our learning curve years ago.
Like any experienced and capable machine operator, we learned to control the variances. It did not take a statistician to determine our success in controlling the variables.
What I describe is based 100% on experience. Our methodology has been honed over literally thousands of runs on many, many cars. We are not chasing ghosts.
The procedures I described above has nothing to do with modifications we did to the GT3 per se. It applies to even getting an accurate baseline on a stock vehicle. Without the same controls I describe above, you *will* see wildly varying, and thus invalid, results, as we did early on in our learning curve years ago.
Like any experienced and capable machine operator, we learned to control the variances. It did not take a statistician to determine our success in controlling the variables.
11-15-2007, 03:03 PM
#4
Here is the question you need to answer.
1. How many runs did you do with the GT3 with the stock muffler?
2. How many runs did you do with the GT3 with your final/best saleable version of your exhaust?
3. On how many of those runs of the STOCK GT3 did you see HP increases - stock GT3 with stock muffler?
4. On how many of those runs of the GT3 with your final saleable exhaust did you see HP increases?
In a rudimentary way - without a statstitician - if your answer for questions 1, 2, and 4 are all the same and a high number AND the answer to question 3 is a LOW number - then you are on your way to making me believe that your exhaust is actually doing something.
1. How many runs did you do with the GT3 with the stock muffler?
2. How many runs did you do with the GT3 with your final/best saleable version of your exhaust?
3. On how many of those runs of the STOCK GT3 did you see HP increases - stock GT3 with stock muffler?
4. On how many of those runs of the GT3 with your final saleable exhaust did you see HP increases?
In a rudimentary way - without a statstitician - if your answer for questions 1, 2, and 4 are all the same and a high number AND the answer to question 3 is a LOW number - then you are on your way to making me believe that your exhaust is actually doing something.
11-15-2007, 03:16 PM
#5
Instructor
Thread Starter
As I stated in my original post, this is how we conduct testing:
We will do pull after pull until we see wheel power numbers stabilize. Sometimes this can take 8 pulls until we see the number stop climbing. We typically do at least 3 more pulls to confirm that the numbers have stopped climbing. This is the same procedure regardless if stock or modified. It is how we do baselines and tests.
We did at least 30 runs on two different stock GT3s. They showed wheel power numbers within 2hp/tq of each other, even though the runs *were* done nearly 3 months apart.
We did at least 40 runs on two different GT3s with our final version of the system. In fact, all the development work was completed on the first GT3, and we got *another* one in here to confirm our findings on the first one, as well as to complete some jigging chores.
Unless we could see at *least* 3 runs showing gains or losses during incremental testing, we did not consider the results valid. In reality, when we knew we were heading in the right direction regarding design improvements, *much* more testing was done to really quantify the results. Sometimes it was so obvious we were heading in the wrong direction, like with muffler deletes, that it did not take us 20 runs to verify.
As I stated in the other thread, we did 159 runs between these two cars on multiple hardware configurations. There was no shortage of verification of data here.
Hence, my desire to make this post regarding *our* experience with dyno testing modern Porsches, and how we have learned to eliminate variances and how to obtain valid data.
Let me know if I need to be clearer in answering your questions.
We will do pull after pull until we see wheel power numbers stabilize. Sometimes this can take 8 pulls until we see the number stop climbing. We typically do at least 3 more pulls to confirm that the numbers have stopped climbing. This is the same procedure regardless if stock or modified. It is how we do baselines and tests.
We did at least 30 runs on two different stock GT3s. They showed wheel power numbers within 2hp/tq of each other, even though the runs *were* done nearly 3 months apart.
We did at least 40 runs on two different GT3s with our final version of the system. In fact, all the development work was completed on the first GT3, and we got *another* one in here to confirm our findings on the first one, as well as to complete some jigging chores.
Unless we could see at *least* 3 runs showing gains or losses during incremental testing, we did not consider the results valid. In reality, when we knew we were heading in the right direction regarding design improvements, *much* more testing was done to really quantify the results. Sometimes it was so obvious we were heading in the wrong direction, like with muffler deletes, that it did not take us 20 runs to verify.
As I stated in the other thread, we did 159 runs between these two cars on multiple hardware configurations. There was no shortage of verification of data here.
Hence, my desire to make this post regarding *our* experience with dyno testing modern Porsches, and how we have learned to eliminate variances and how to obtain valid data.
Let me know if I need to be clearer in answering your questions.
11-15-2007, 04:02 PM
#7
Instructor
Thread Starter
I am not sure why you would say it does not matter how many runs we did during development.
Without those incremental runs, we would not know if we were heading in the right direction.
No power product is purely an academic exercise. All the virtual modeling in the world cannot substitute for a real live properly conducted power test.
Our final product would not be what it is unless we had the incremental data to rely on.
The incremental testing showed us the issue with disrupting the resonance tuning going on in the stock exhaust by eliminating mufflers. The incremental testing showed us the difference between header primary lengths. The incremental testing showed us the difference between 2.5" and 3.0" tubing diameters.
What you proposed in your previous post, that unless hundreds of tests were done in identical atmospheres, is just not possible. In fact, if it were, all dyno testing would be a brainless process of strapping down the car and doing the pull. Even the OEMs do not have that luxury. That is where the science of establishing viable test protocols comes into play, which we feel we have succeeded in doing.
Without those incremental runs, we would not know if we were heading in the right direction.
No power product is purely an academic exercise. All the virtual modeling in the world cannot substitute for a real live properly conducted power test.
Our final product would not be what it is unless we had the incremental data to rely on.
The incremental testing showed us the issue with disrupting the resonance tuning going on in the stock exhaust by eliminating mufflers. The incremental testing showed us the difference between header primary lengths. The incremental testing showed us the difference between 2.5" and 3.0" tubing diameters.
What you proposed in your previous post, that unless hundreds of tests were done in identical atmospheres, is just not possible. In fact, if it were, all dyno testing would be a brainless process of strapping down the car and doing the pull. Even the OEMs do not have that luxury. That is where the science of establishing viable test protocols comes into play, which we feel we have succeeded in doing.
Trending Topics
11-15-2007, 04:53 PM
#8
- funny thing is, i'm quite certain (statistically speaking) that i'm 1 of maybe 2 statisticians on this board
- i'm also pretty certain, that i'm likely the only statistician from the only Mathematics Faculty in North America, that is the University of Waterloo in Ontario Canada, i.e. my degree is from the Department of Mathematics, my major is Statistics
- with all that BS out of the way
- i merely posted my dyno results from testing the Sharwerks exhaust and FVD software for information purposes only, not to start a dyno war, or to show that my car now dynos 388 pink pony wheel horse power
- in my experience modding 3 cars pretty heavily, my 2 cents is to go one step at a time, work with people you trust and that stand behind their products, it is also usually better if they are local but that is a tough one to always achieve
- finally as a statistician, it is going to be very difficult to "prove" real gains with this car as there are so many variables that are constantly changing and uncontrollable, unless the change, positive or negative, is rather large (i.e. chipping / re-mapping a turbo car is pretty obvious to everyone)
- if you take 150 pounds out of this car, re-gear it, put the best headers you can find on the car with high flow 100 / 200 cell cats, run 100 octane with re-mapped software, i think it would be pretty easy (statistically) to show that the car is faster in a straight line doing 5 to 10 straight line pulls against a stock car because if it isn't, then statistically i'm 100% certain that you just blew 10Gs
- finally, there is yet another moving variable, it is that the car itself from brand new to 5,000 miles old can itself likely gain 5+ WHP so this in and of itself virtually makes it impossible to with any certainty at all claim anything if the tests are done over time as the car ages (i actually believe that my car with 12,000 miles has now likely gained 5hp from the very early dyno that was done when the car had 1000 miles on it - it was then showing about 378-379whp - but again that was on a different dyno
over and out
- i'm also pretty certain, that i'm likely the only statistician from the only Mathematics Faculty in North America, that is the University of Waterloo in Ontario Canada, i.e. my degree is from the Department of Mathematics, my major is Statistics
- with all that BS out of the way
- i merely posted my dyno results from testing the Sharwerks exhaust and FVD software for information purposes only, not to start a dyno war, or to show that my car now dynos 388 pink pony wheel horse power
- in my experience modding 3 cars pretty heavily, my 2 cents is to go one step at a time, work with people you trust and that stand behind their products, it is also usually better if they are local but that is a tough one to always achieve
- finally as a statistician, it is going to be very difficult to "prove" real gains with this car as there are so many variables that are constantly changing and uncontrollable, unless the change, positive or negative, is rather large (i.e. chipping / re-mapping a turbo car is pretty obvious to everyone)
- if you take 150 pounds out of this car, re-gear it, put the best headers you can find on the car with high flow 100 / 200 cell cats, run 100 octane with re-mapped software, i think it would be pretty easy (statistically) to show that the car is faster in a straight line doing 5 to 10 straight line pulls against a stock car because if it isn't, then statistically i'm 100% certain that you just blew 10Gs
- finally, there is yet another moving variable, it is that the car itself from brand new to 5,000 miles old can itself likely gain 5+ WHP so this in and of itself virtually makes it impossible to with any certainty at all claim anything if the tests are done over time as the car ages (i actually believe that my car with 12,000 miles has now likely gained 5hp from the very early dyno that was done when the car had 1000 miles on it - it was then showing about 378-379whp - but again that was on a different dyno
over and out
11-15-2007, 05:04 PM
#9
Instructor
Thread Starter
Paul,
My post has nothing to do with your post. You and I did have this same conversation via e-mail, but that only served to reinforce that in fact there is very little standards out there for what is indeed a very difficult vehicle to properly dyno test. That is why I felt a public post was in order.
Like I told you via e-mail, we want to make sure our own products are being tested properly by end users, so it is in our best interest to "spread the word" about how to ensure success in dyno testing these cars.
Yes, there are *a lot* of things to juggle when conducting dyno tests on any car, let alone a car with the ECU mapping challenges that the GT3 has.
However, that does not mean that someone who does the typical 3 runs back to back has the same chance of generating valid data as someone who takes the steps as I outlined above.
We've tested many Porsches with less than ideal dyno protocols and we saw how erratic the results can be.
However, we fine tuned our testing procedures as I described, and our ability to get *very* repeatable data improved greatly. And I mean repeatable as in 2-3 whp variance per run (and for the record, our published data is an average, not a best run scenario).
Do not take my post as a jab at you. It is not. It is intended as a public service in order to standardize the data shared on this forum. Given the vast amount of dyno time we have spent with this vehicle in particular, I feel we are qualified to point out the pitfalls.
My post has nothing to do with your post. You and I did have this same conversation via e-mail, but that only served to reinforce that in fact there is very little standards out there for what is indeed a very difficult vehicle to properly dyno test. That is why I felt a public post was in order.
Like I told you via e-mail, we want to make sure our own products are being tested properly by end users, so it is in our best interest to "spread the word" about how to ensure success in dyno testing these cars.
Yes, there are *a lot* of things to juggle when conducting dyno tests on any car, let alone a car with the ECU mapping challenges that the GT3 has.
However, that does not mean that someone who does the typical 3 runs back to back has the same chance of generating valid data as someone who takes the steps as I outlined above.
We've tested many Porsches with less than ideal dyno protocols and we saw how erratic the results can be.
However, we fine tuned our testing procedures as I described, and our ability to get *very* repeatable data improved greatly. And I mean repeatable as in 2-3 whp variance per run (and for the record, our published data is an average, not a best run scenario).
Do not take my post as a jab at you. It is not. It is intended as a public service in order to standardize the data shared on this forum. Given the vast amount of dyno time we have spent with this vehicle in particular, I feel we are qualified to point out the pitfalls.
11-15-2007, 05:47 PM
#10
Paul,
My post has nothing to do with your post. You and I did have this same conversation via e-mail, but that only served to reinforce that in fact there is very little standards out there for what is indeed a very difficult vehicle to properly dyno test. That is why I felt a public post was in order.
Like I told you via e-mail, we want to make sure our own products are being tested properly by end users, so it is in our best interest to "spread the word" about how to ensure success in dyno testing these cars.
Yes, there are *a lot* of things to juggle when conducting dyno tests on any car, let alone a car with the ECU mapping challenges that the GT3 has.
However, that does not mean that someone who does the typical 3 runs back to back has the same chance of generating valid data as someone who takes the steps as I outlined above.
We've tested many Porsches with less than ideal dyno protocols and we saw how erratic the results can be.
However, we fine tuned our testing procedures as I described, and our ability to get *very* repeatable data improved greatly. And I mean repeatable as in 2-3 whp variance per run (and for the record, our published data is an average, not a best run scenario).
Do not take my post as a jab at you. It is not. It is intended as a public service in order to standardize the data shared on this forum. Given the vast amount of dyno time we have spent with this vehicle in particular, I feel we are qualified to point out the pitfalls.
My post has nothing to do with your post. You and I did have this same conversation via e-mail, but that only served to reinforce that in fact there is very little standards out there for what is indeed a very difficult vehicle to properly dyno test. That is why I felt a public post was in order.
Like I told you via e-mail, we want to make sure our own products are being tested properly by end users, so it is in our best interest to "spread the word" about how to ensure success in dyno testing these cars.
Yes, there are *a lot* of things to juggle when conducting dyno tests on any car, let alone a car with the ECU mapping challenges that the GT3 has.
However, that does not mean that someone who does the typical 3 runs back to back has the same chance of generating valid data as someone who takes the steps as I outlined above.
We've tested many Porsches with less than ideal dyno protocols and we saw how erratic the results can be.
However, we fine tuned our testing procedures as I described, and our ability to get *very* repeatable data improved greatly. And I mean repeatable as in 2-3 whp variance per run (and for the record, our published data is an average, not a best run scenario).
Do not take my post as a jab at you. It is not. It is intended as a public service in order to standardize the data shared on this forum. Given the vast amount of dyno time we have spent with this vehicle in particular, I feel we are qualified to point out the pitfalls.
Extreme example. If I build a cement life jacket - just because I divulge the exact details about how I designed, tested and manufactured it still doesn't mean it is going to perform the way a life jacket should.
11-15-2007, 06:01 PM
#11
Instructor
Thread Starter
I still don't follow you.
What if I built a cement life jacket and did not only "tell" you that it floated, but *showed* you that it floated. After all, with proper water displacement, you can even get a cement lifejacket to float.
By analogy, I am *showing* you that our exhaust testing is valid by not only offering dyno sheets, but also explaining how our chassis dynamometer tests were conducted. How else am I to *show* you power quantifications? Lap times? Quarter mile times?
I guess you are telling me that unless we did hundreds of runs in 100% controlled ambient conditions, no matter how controlled we made our dyno testing, it is inherently flawed?
Does the fact that you cannot achieve these ideal parameters mean that dyno testing is not a valid way to measure power?
Honestly, I am not trying to be a smart *** here, I am just not sure I understand what point you are trying to make.
What if I built a cement life jacket and did not only "tell" you that it floated, but *showed* you that it floated. After all, with proper water displacement, you can even get a cement lifejacket to float.
By analogy, I am *showing* you that our exhaust testing is valid by not only offering dyno sheets, but also explaining how our chassis dynamometer tests were conducted. How else am I to *show* you power quantifications? Lap times? Quarter mile times?
I guess you are telling me that unless we did hundreds of runs in 100% controlled ambient conditions, no matter how controlled we made our dyno testing, it is inherently flawed?
Does the fact that you cannot achieve these ideal parameters mean that dyno testing is not a valid way to measure power?
Honestly, I am not trying to be a smart *** here, I am just not sure I understand what point you are trying to make.
11-15-2007, 06:07 PM
#12
Rennlist Member
I thought you would only need hundreds of dyno runs if the standard deviation is very high in which case you would have to average over more runs so that the standard deviation of the average is much less than the claimed gain. However, if under their test procedures, they are getting very consistent results (say a standard deviation of +-1%), and with the upgrades they are getting a 3% increase of power with the same consistancy on the same day, how is this not valid and statistically significant?
11-15-2007, 08:06 PM
#13
Instructor
Thread Starter
Here is the question you need to answer.
1. How many runs did you do with the GT3 with the stock muffler?
2. How many runs did you do with the GT3 with your final/best saleable version of your exhaust?
3. On how many of those runs of the STOCK GT3 did you see HP increases - stock GT3 with stock muffler?
4. On how many of those runs of the GT3 with your final saleable exhaust did you see HP increases?
1. How many runs did you do with the GT3 with the stock muffler?
2. How many runs did you do with the GT3 with your final/best saleable version of your exhaust?
3. On how many of those runs of the STOCK GT3 did you see HP increases - stock GT3 with stock muffler?
4. On how many of those runs of the GT3 with your final saleable exhaust did you see HP increases?
I reread what you posted. Yargk's post clued me into what you were really questioning.
Here are the answers:
We eliminated all runs that were a factor of warmup or adaptation. We were able to qualify these runs as statistically invalid because they would occur at the beginning of run sessions.
We would not start to include runs as valid until power numbers would stabilize during the run sessions. By stabilize, I mean they would be within 2-3 peak hp/ftlbs of torque from run to run, with very little overall deviances on curve overlay.
We tested this way for both stock and final product (and all variations during development). Approximately the same number of runs were done on the stock and final product, on the same car, and then also on the same day both stock and modified as a final control of the data.
We then selected an average run to be the official representative of both stock and modified in our published dyno sheets.
I hope this clarifies our data management protocols. All runs that showed power gains from run to run were eliminated to account for warmup and adaptation. All runs that were consistent from run to run were included in our data analysis.
I just want to point out that the point of this thread was not to brag about our exhaust results, but to make an argument for stricter dyno testing protocols based upon how sensitive of a vehicle the GT3 is.
11-15-2007, 08:31 PM
#14
Addict
Rennlist Member
Rennlist Member
I was told that I cheated in my statistics class at MIT because 3 students besides myself all got the same test question wrong (all the same incorrect answer) and the statistic professor told us that was statistically impossible. Needless to say, he was forced to take a year of sabbatical. Point being, statistics doesn't always trump common sense.
This is not a peer-review panel here and we are not publishing in Nature or NEJM. No one has the time or resources to run enough dyno tests to get a p<0.05 and still be able to charge a reasonable amount for a product. We all have to decide if manufacturers are doing their homework in providing us products that actually generate gains, vs products that simply generate income.
This is not a peer-review panel here and we are not publishing in Nature or NEJM. No one has the time or resources to run enough dyno tests to get a p<0.05 and still be able to charge a reasonable amount for a product. We all have to decide if manufacturers are doing their homework in providing us products that actually generate gains, vs products that simply generate income.
