is progesterone test really reliable, last year i did progesterone test on 3 different bitches before they had AI, 1 bitch i didnt as she was staying at the stud which i took her on day 10. She bred natural on day 13 and 16, unfortunately 10 days after coming home from the stud i came home and found her dead, she had got uterus infection, i dont know much about it but one of my other females also got it nearly 2 weeks after her returning as she had been in season lucky i caught her in time and she had medication.  Anyway the 3 i had progesterone tests done for and AI none of them got pregnant
Do you not think the vaginal swab is more accurate or is it just pot luck and let nature do its business and have no tests done

Progesterone testing is very reliable as long as you keep testing, every two days until the levels show that the bitch has actually ovulated. The eggs won't be ripe for another two days and survive for about two more.

This can mean three or four tests, depending on how soon ovulation occurs as you generally start testing around day 8 in a maiden, or a couple of days prior to when she was mated last time.

If you have just one test the lab will guestimate when she might be ovulating and ready for mating

This is based on the speed of the hormone level rise in the average bitch, unfortunately few bitches are average.

As for AI this will always be less reliable than a natural service (especially if frozen)and a lot depends on the skill, materials used to get the semen where it needs to be.  Vaginal AI is much less reliable than surgical, which in the UK vets and KC frown on, as it is invasive

Cytology (vaginal swabs) and progesterone are both useful, at different times of the cycle.  Cytology is very useful during proestrus - you will see the cells change from parabasal and intermediate cells during proestrus, to superficial and anuclear cells during estrus.  So:  Cytology will tell you when estrus begins. 

Ovulation can happen any time during estrus, although on average the LH peak is 2 days after the start of estrus and ovulation is 2 days after that.  Most bitches will stand for breeding throughout estrus, regardless of when ovulation is.  And semen lives 3-5 days.  So if you are using natural breedings, identifying the start of estrus using cytology and then breeding a couple of days after that and 48 hours after that, will work well for the majority of bitches.

However, if you are travelling a long way or using frozen semen or doing an AI, you will need to be much more precise - in which case you would still use cytology to determine the start of estrus and start progesterone testing from then onwards.  This eliminates the need for progesterone testing during proestrus and all those annoying and low results which don't seem to be going anywhere - and you save the progesterone tests for when estrus starts.

Progesterone testing doesn't really have anything to do with pyometra - but really you should be on pyo watch any time after a season since there are signs that a bitch needs medical help - she will be generally unwell and lethargic and if you know a bitch, you know something is not right. 

I am not sure why you bred 4 bitches and saw none get pregnant, but I would really be suspecting the presence of herpes in your kennel (which can cause absorbtion early in pregnancy so you never know the bitch is pregnant) and using the CHV vaccine as soon as it is re-available... those don't really sound like good odds at all.  To give any more help, we'd need to know all the progesterone results for the bitches, the days of the season that you got the results, the days they were bred, how they were bred - it's not really possible to give any useful suggestions as to why they didn't get pregnant without more info...

I just looked at your other post (now locked) where you said that you had a progesterone test done on day 9 (I think) which was 2.10 and they told you to go to the stud in 2 days or re-test?!

I've no idea where that advice came from but the table for interpreting progesterone tests is:

0-2  nmol/L           Baseline concentrations too early to estimate ovulation 
3 - 6 nmol /L        Minimum 2 days before ovulation expected however result of 3 - 4 nmol/L may persist for a week or longer before increasing. Earliiest estimating breeding 4 - 6 days but could be longer (re test needed)
7 - 12 nmol/L   Minimum 1 day before ovulation. Estimated window for breeding 3 - 5 days but could be longer  ( re test recommended) 
13 - 18 nmol/L   Ovulation impending or just occurred. Estimated breeding window 2 - 4 days  
19 - 31 nmol/L   Ovulation recently occurred. Estimated breeding window 1 - 3 days
32 - 64 nmol/L   Ova have matured, optimal potential for fertility.  Estimated breeding window  0 - 2 days
65 - 90 nmol/L   Ova have matured but aging, decreased potential for fertility.  Breed  at once
Above 90 nmol/L    Too late

That was not very useful advice - as you can see: "result of 3-4nmol may persist for a week or longer before increasing" - as it has done for you.  Again, you need way more than one test...