News Seagate Ships 20TB HAMR HDDs Commercially, Increases Shipments of Mach.2 Drives

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
I admire how confident you are in your incorrectness while feeling the need to continually try to insult me and accuse me of not knowing what I am talking about....Western Digital has put in writing that MTBF does not mean anything
I've not insulted you; I've simply pointed out you misunderstand statistics. Your last post firmly cements this position. Nothing in that WD statement contradicts me, nor does it imply that MTBF "does not mean anything". If you'll climb down off your high horse for a moment and listen, you'll understand why. Let me start the explanation a bit further back.

You have a mechanical item -- hard drive, automobile, anything. You want to know if it will break down over a specified period of time. Assuming the period is shorter than the "wear-out" period, then you're asking about random faults, which by definition cannot be predicted, no more than you can predict whether the next flip of a shuffled card deck will reveal an ace. All you can predict is probabilities. Not certainties.

With me so far? This is what WD (and all manufacturers of all fault-rated components and equipment) mean when they say their data "does not predict an individual units reliability and does not constitute a warranty". A one-million hour MTBF is not a statement that your drive will run one million hours. It's a statement of probabilities only. That doesn't make it "meaningless". Far from it.

How do you predict experimentally the probability of an ace of spades appearing in a single flip of a deck of cards? (Let's assume the deck composition is unknown, so you can't simply calculate that mathematically). You run experiments, that's how. You flip a card, reshuffle, and flip again. Each of these events is a sample. (Note-- the card flip you're attempting to predict is NOT a sample, nor is the hard drive you own that you wish to know its failure chance. This is the explanation for my earlier statement that you were misinterpreting sampling).

With a large sample size, you can calculate probabilities for the random single-event case you're interested in. The more samples, the better the sample space reflects the entire statistical universe of possible events. But this sample space still applies to single events, but on a probabilistic, rather than a deterministic basis. In the case of MTBFs, even though they're not expressed as percentages, they're still simply probabilities, not warranties of performance. And they are far from meaningless, which is why every manufacturer of critical components goes to such great lengths to calculate those numbers.
 
Then buy another SSD or learn how to uninstall games. With SSD's dropping below $100 per terabyte, no one would recommend a RAID 0 mechanical drive array over SSD's for pretty much any speed sensitive home use case. For regular large scale file storage where speed is largely irrelevant, a mechanical drive is the way to go, though certainly not in a RAID 0 array.

I have
2TB NVME boot drive
2TB Sata SSD
14TB NAS (4 x WD Reds, FreeNAS) over a Gb wired connection

I cannot tell the difference loading GTA5 (or any other large game) over any of them - that current games need massive throughput is a total misnomer - maybe when the PS5 streaming textures tech filters down? I can see a 14TB striped mechanical array being very useful for those of us that like to have all of the games they've actually bought and don't have a huge internet connection to hold replaceable game data. The only reason I have only SSDs in my system is that I'm lucky enough that my business pays to have a decent 'spare' (ahem) NAS at home and I like a PC as quiet as possible. Obviously for a boot device SSD's are way in advance, but gaming? One person calling one app which calls set files at specific points with large lulls in between? It doesn't get ay easier than that for a drive.