Power versus playability is a unique problem for cube curators. There are more than enough cards of all types now for cube and so the simple metric of playability is no longer the only sensible thing to consider when building a cube. It very simply used to be the case that you could rank cards by how much play they saw and simply cut from the bottom of that list as new stuff arrived and have a well curated cube as a result. There are now however plenty of cards than you would play a lot of the time that are not worth including in your cubes due to the lack of impact they have. The classic examples here is of the utility lands (to include half land MDFCs) and cheap cantrips (to include cheap cyclers). I did an article on the best do nothing cards should you feel you want a better idea of what they are and why you should look to play more of them than your probably do and it can be found at;
Yes, you might chose to play Sokenzan, Cruicible of Defiance in 95% of your cube red decks making it one of the most played "red" cards in the cube however if it has a negligible impact on game impact then it isn't worth a slot in your cube. The mathematical solution is therefor a product of how playable a card is an how much it changes win percentage. Both are things you either need insanely large data sets for or something you have to estimate at. That is fairly easy I find for the amount a card is played but really tricky for the impact it has on win percentage. Even the most broken cards are unlikely to move it by more than a couple of percent, with most cards having a fraction of a percent effect on things. Card type makes it difficult too as some cards are threats that are supposed to win games while other cards disrupt, and other things increase the consistency of your stuff. I can fairly directly compare how much Baneslayer affects games compared to an Avacyn and be happy in my choice to play the latter in preference. Equally I can play both in my cube and quickly spot an increase in white top end threats going late in packs and being left in sideboards. When you have too much of a thing they are not spotted in decks so much as they are in sideboards and late in draft packs. If they start showing up in those places with any regularity it is time to get trimming.
How then do I compare the win percentage change of Tithe to one of these five drop angels? The only way here is to measure it across thousands of games where I have not only the results but also the deck lists. Obviously this is not something I can do as I simply do not have the data. Only really something like MODO or Arena can hope to harvest the data to give precise win percentages for cards in decks. As such I have to be more intuitive. Luckily the highly playable but low impact stuff does wind up seeing a lot of play and so you can observe how impactful they are going to be. So while you do have to intuit your choices for the more subtle cards you do get given the best opportunity to do so. If I am playing a filler or "do-nothing" card and it literally does nothing positive after a couple of months or say in around twenty in game sightings then it is just not high impact enough and I cut it. Equally if it gets used a couple of times in that sort of window but doesn't have an appreciable effect on the game then again it is too low impact and gets cut. I don't need to know exactly how impactful those cards are to any greater precision but it doesn't matter it just isn't going to be relevant in cube when you are below a certain, and relatively high bar. This means I don't need the insane amount of data you might think as I am not trying to be exact, all you need is not impactful enough and you can move on with your day. The slot is going to be better off used by a narrow and infrequently played card that does offer a big impact, or just entirely empty resulting in a smaller, more consistent, and refined cube.
Another convenience when trying to work out which low impact cards to include is that there isn't all that much in the way of direct competition between them and the action cards. You cannot fill a deck up with cantrips and do nothings, it needs some substance. There is very much an optimal number for such things and it becomes fairly obvious when you are reaching that limit. If your strictly better than a basic land fails to get play then you have too much of that sort of thing. As such I tend to find roughly what balance of things I want in my cube and then work towards that balance . This means my threats compete with each other, generally with the same sort of CMC threats in the same colours. My cantrips and do nothings compete with each other too but not with the threats, just so long as I have my ratios of the two roughly right. There are other groups too such as fixing, removal/disruption, ramp, and that sort of thing. Getting a good balance between these sorts of cards for your cube meta is one of the most important tasks of a cube curator and is one of the main contributors to a good drafting and playing experience of a a cube.
Time for a little tangent (skip to the last paragraph if you are not in the mood for such things) - I like to have somewhere between the perfect number of cards and one card over that perfect number for the various categories in my cube. This way I find you can tell more quickly when a group of cards is getting used less than before but also when it is getting used more. In general you want to keep your card categories near to this "perfect" number, which is just another way of saying you want to keep your ratios consistent. Too few of a card type and they will simply always get played all of the time and reduce choice in your draft and construction while removing variety from everything. Too much over this perfect number and you add inconsistency and bloat to your cube. It is harder to see when you need to adapt to a shift in the meta and the variance of your cube will be increasing on all metrics. This way you can adjust accordingly as the meta shifts by adding or subtracting similar cards. If I add cards to a group rather than replacing them then I will start to need to adjust the other categories as well by adding cards to them so as to keep my ratios accurate. Technically if I had an endlessly large supply of action (disruption and threats) of cube power level then I could add in an unlimited number of the various do nothing cards as well without messing up the ratios. What happens in reality is that you hit a stopping block when trying to expand the size of your cube beyond that of practicality or consistency, and that is simply too many weak or non-existent cards in key groups.
And now a tangent on that first tangent! - A while back I noticed a sharp decline in people playing Terror effects, basically any black creature removal spell at two mana. The more restrictive one mana stuff saw play as did the many three or more mana cards that had no restrictions and hit planeswalkers. It was not wildly long lived but it might have been had I cut all the various options on the Terror like cards. They were not seeing enough play to merit keeping any of them (I likely had 3 or 4 cards like Go for the Throat in cube at the high point of this type of removals popularity) in cube yet I made sure to keep in one so that I could monitor when it started to pick up again. It has picked up but not to the levels it once held and probably lies somewhere between 1 and 2 copies of Terror like cards in my cube currently being the ideal balance. I couldn't tell you exactly what the cause of this lull was nor indeed why it reverted. I could hazard some educated guesses but there would be no evidence to support them!
With that in mind - Mana costs are so abrupt with very little ability to smooth them out. This means that as a meta shifts there will be a point at which certain card costs either cease to be viable or become viable depending on what they do. Slow a format down and you can start to pay three mana for a counterspell, add in all the broken power and even two mana actual Counterspell is too pricy. A small shift in the meta at just the right tipping point can make all, or most, of a group of similar cards as far as effect and CMC go, viable or unviable.
This is what I think happened to black removal. The two mana stuff just sat at the wrong place where there was not enough consistency advantage over the one mana options to make up for the tempo concessions while the exact reverse was true when comparing it to the three mana alternatives. Decks simply played a mix of one and three mana removal giving them the ability to fight tempo well without risking too much in the way of dead cards unable to do the job they are supposed to. Bone Shards and Bloodchief's Thirst arrived on the scene and took the slots of the remaining best Terror effects. A little while later Infernal Grasp was released and we got one of the most reliable Terror style removal spells and more two mana removal started to get played again. Essentially we have two factors at play, a slow and steady move by the meta as a whole applying a pressure on the kinds of card that are viable thus making the conditions ripe for the more sudden and drastic changes when new and powerful cards doing the right things come along to accommodate the needs of that meta.
So, to get a good amount and selection of do-nothing cards (or quality of life cards if you would prefer a more positive term for this fantastic group of cards) in your cube you first need to work out what sort of number your cube supports by adding in a bunch till they stop seeing play. Keep an eye on the draft side of things and make sure people are still getting good choices in their packs and making good decks. When you are happy with this have a bit of a count and see what your ratios look like and try to stick to that when adding or removing cards. Then simply remove the cards that least frequently do stuff (unless it is obviously game winning) until you are nearing your ideal ratios. Good ratios are far more important than having precise number balance for colours or points on the curve or even types of card. I see plenty of people with exactly X cards of each colour or strict allowances on things like point on the curve or numbers of a card type. While being aware of these things is great limiting yourself with fixed arbitrary rules only hinders your ability to curate with finesse.
Ratios of cards that function the same way in game are the most relevant thing to balance by a long way. Further to that you really don't need to be precise, you have a good amount of wiggle room. Just be aware that although you can add or remove stuff without having to add or remove stuff in the other areas of your cube all the time as well that you will be adding a bit more tension on things when you do. I always notice when I get my first big pile of new cards to test from the new sets. Typically they are only in some of the categories and do not offer a well balanced set of things that keep the ratios in order. They are also typically averaging very much the low end of the power level in cube (made the case by my insistence on testing loads of cards!). All this jank added to a cube makes drafts a lot thinner and acts as a good incentive to get to the chopping block fast! Typically we end up adding an extra card to packs or something of that ilk so as to compensate for the lowered quality. I have even taken to testing the cards in groups so as to not add them all in at once but sadly I have not the play time to make this workable at present.
A cube needs a mix of high impact cards and high playability cards and most cards cannot be both, although unsurprisingly the best cards are powerful and playable in much the same way that many of the best cards afford both tempo and value. To find the best low impact cards for you and your playgroup you need to find out how many of that kind of card your cube supports. Then you can have them compete for space internally by seeing how often they are used and how often it is relevant. This leaves you with the right balance of things overall and the best cards within your various well balanced groups. Castle Ardenvale was a perfect example of a card that had to go despite seeming amazing and arguably just being a great magic card. It almost always saw play in white decks and it fairly frequently got activated as far as utility land go. Castle just failed at the very last hurdle of simply never seeming to change the outcome of a game. It turns out that these days in cube spending five mana to make a 1/1 is effectively doing nothing! In practice what it means is that you are actually out of relevant stuff to do and that is as good as a loss against someone who still has anything relevant to do.