Saw this yesterday [several times requested] and thought of the one weakness in this. The ever faster move off of validated result batches. You get barely time to blink with for instance Clean Water and a few other sciences, before they get transited to the Master DB, thus removed from the RS pages. Then What?

I'm running pirogue's WCGDAWS twice a day to capture everything. If a result changes status, a new timestamp is affixed and then sorting shows what has validated for the 12 hour period. This so far has not caused me to miss almost any validations. Some that take long to find quorum are moved off even quicker. Then you'll never know... but have full trust that credit was granted... The sideshow to doing science in distributed computing, which always lives center stage [how else do we know the TFL per CPU year contributed ;P ]

--//--

Any 12 hour interval is fine it seems **, no guarentees. Seen on 5400 results in the WCGDAWS db only 3 misses... for ever sitting in PV and Inconclusive. New validations can disappear within a 12 hour period i.e. before the next stats run it thus seem, but it looks to be a rare coincidence and as noted, happens to those that took long to find a wingman. [or were stuck in "try validation" and noone did, then the servers concluding that 10 days was enough and transit them out as no additional wingman was needed ;>]

** Found 1 of the Beta's that completed first on Saturday, validated quickly and already gone.

--//--

It used to be about 72 hours after validation, except then we did not run 700,000 results and more a day. Come 6 months, lest the techs are able to find more creative ways to elongate work as they've been doing to FAAH/HPF2 and seemingly too to GFAM, or combo multiple jobs in 1 task assigned, 1 million could become the norm and that has serious server performance implications, particular the scheduler that is the pivot of the whole operation and what you see on the RS pages. 84 hours is today utopic, compared to any standard BOINC grid that has results on for 24 hours if lucky. Batches completed belong in the scientists hands and not be lingering on the Result Status pages. [Don't ask about DDDT2... clueless why they are given the grand exception... maybe because there's just a few thousand in a run for A/B type :]

At any rate, the various ways of highlighting a validation in the current 24 hour WCG cycle, color coding, tickmarking or adding a 3rd timestamp to the send/received column has been discussed in past and think there was even a way decided on how to show it. If only there was time to do this and not cause a significant performance impact, the always present concern.

The sighted WCGDAWS tool does brilliant, so that's what quite a few trackers do and on top build a detail history and do analysis in many ways. I've got 5400 in there now and the programmer who made the tool has tens of thousands in... ready to filter, sort, average, print copy, slice and dice and export to a spreadsheet... and easily filter out the slackers for a RS page visit to see if it got stuck before it reached the validator.

--//--

Ok cjslman now I better understand what you meant. There is indeed a 12 hours period between statistics update and validated results.
But SekeRob comment is also valid. If Validated WUs are quickly pulled out then depending on the frequency at which you check your results, you would probably miss anyway the total picture.

Due to the huge number of results I have displayed in the Result Status list (around 5'000) I cannot start checking credits. I did once a time consuming statistic across some of my machines and for various projects on "credits requested" and "credits granted". The system is fair in average and that is what counts. Sometimes a WU gets 70% of the credits requested and another WU gets 120%. Understanding why is a specific WU has validated in a particular way is not really very relevant, except when there is a major problem like one project were a majority of WUs start erroring or having incredible runtimes etc. etc.

I have about 1'000 WUs that are validated every day so I use the Results Status in combination with Device Statistics function to check from anywhere if one of my rigs has stalled or has delays in sending WUs that could be a communication problem etc. etc. I can so identify which machine(s) has a trouble and when at home go in my basement and check.