I should have put more into that answer.
They are the same value, but not the same tolerance is needed for R20. R5, 16, 17 and 18 are all higher tolerance resistors - their true value is very close to 10K. R20 can be a lower tolerance resistor - a 5% or even 10% resistor would be fine. That's not to say that you can't use a high tolerance resistor for R20. R20 is just a pull-up resistor that holds the reset pin high until someone or something sends a ground to reset the system.
Why the higher tolerance? Well, those resistors are used as voltage dividers in the circuit, and having them very close in value to one another makes setting up the probes in software very easy. Otherwise you would have to measure each one manually and input the values. Using high tolerance resistors eliminates the need for this extra step.
And you can spend quite a bit of cash on high tolerance resistors.
digikey 9,999 - 10,001 for a 0.01% tolerance resistor, so a 1 ohm variance
digikey 9,990 - 10,010 for a 0.1% tolerance resistor, so a 10 ohm variance
digikey 9,900 - 10,100 for a 1% tolerance resistor, so a 100 ohm variance
So a 10% tolerance resistor could be off by 1000 ohms!
The takeaway - buy as sensitive resistors as you can afford for sensitive applications! Best bang for the buck occurs @ 0.1% for HM4 R5, 16, 17 and 18.
Besides, you don't want to waste a dollar on a $0.02 10K resistor.
Personally, I'm using these.
http://www.digikey.com/product-detail/en/CMF5010K000BEEB/CMF10KQBCT-ND/2197140