PLEXThe main way in which one gets skillpoints is simply to subscribe. You get one character month of training per PLEX, and PLEX have an ISK value.
The formula for skill points is: (primary + secondary/2) SP/minute. Per hour (the time unit used in EVEMon): 60*primary + 30*secondary.
A character with no implants, remapped to the best possible attributes for a skill, gets primary=27 and secondary=21. Thus, 2250 SP/hour. There are 24 hours per day and 30 days per month, so 720 hours per month. One PLEX costs around 525m ISK (May 2013). Thus we can compute the fundamental conversion of ISK into skillpoints: at best, your character pays 324 ISK/SP. At worst, both attributes are 17; in this case you get 1530 SP/hour, or 477 ISK/SP.
ImplantsWith these figures in hand, we can compute the cost effectiveness of implants. However, to do this we must make further assumptions. This is because implants last forever, unless you replace them, or you get pod killed. Because of this, we cannot compute implant values in ISK/SP without some means to bound the lifetime of the implant, and/or discount its gains over time. I am adopting the former approach, and will simply compute implant values assuming they last a finite time, without discounting future skillpoints gained at all.
What are some interesting lengths of time? Well, one thing I'd like to know is: is buying an implant when I get cyber 1 worthwhile, if I plan to get a better implant at Cyber 4? (Cyber 1 to cyber 4 takes around 3 days; 1/10 month.) And similarly, is buying an implant at cyber 1 worth it to get to cyber 5? (17 days: 1/2 month). Many useful low-skilled alts can be created in about 2 to 3 months, so, to stick with a strictly geometrically increasing time sequence, let's adopt multiples of 5, and compute values for 0.1 month, 0.5 month, 2.5 months, 12.5 months, and (for fun) 62.5 months.
Here is a spreadsheet showing this data.
The table shows the skillpoints gained by an implant under the assumption that the character only trains skills whose primary attribute is the one the implant boosts. If the implant boosts a secondary attribute, double the number shown in the table.
I have bolded numbers which are less than the PLEX cost of skillpoints. Thus, we can see that for training many low skilled alts, it is not worthwhile to train to cybernetics IV to get better implants, since these implants are not cost effective in the amount of time that the character will train.
We also can see that for training cyber 4, using a +1 intelligence implant is worth doing, but not a memory implant. If you are going for cyber 5, buying a +1 intelligence implant and a +1 memory implant will be worthwhile. Getting +2 versions or higher is not worth it.
For training for-sale characters of 1 year, +4 implants are a marginal deal in terms of just their skillpoint effect over that year. The most salable skills are combat skills, and these largely divide into two classes: perception/willpower, and intelligence/memory. A typical combat pilot might have half of his skillpoints in each of these categories. So, you'll get about half value (2x the cost shown in the table) for implants boosting the two primary attributes, and a quarter value (4x cost) for implants boosting the secondary attributes. This would suggest buying +3 implants for the primary attributes and +2s for the secondaries.
That said, remember that I assumed that there was zero value to having implants after the end of the time period. This would be true if, for example, the character became a utility alt of some kind and never trained any skill again. However, it is pretty obviously untrue for most for-sale characters: the new owner will value the implants. Most of the cost of the implant can be passed on to the new owner. So, I'd suggest +4 implants to all attributes you train much, which will be all attributes except maybe charisma. Many characters will barely use charisma, and can settle for a +1 or +3 implant.
Cerebral AcceleratorFinally, we discuss the Cerebral Accelerator (CA). CAs are not for sale on the markets; you can only find them in contracts. They typically go for around 300m ISK. They give +3 to all attributes for a starting character but only to the character's 35th day. Is this worth using for a for-sale character? Where is the breakeven point?
I assume you plug the thing in ASAP and train throughout the 35 days. Each hour, the CA gives 270 extra skillpoints, no matter what skill you train. So, the total skillpoints gained is 226800, and the cost is 1323 ISK per skillpoint. This is considerably worse than what you get for PLEX. So in general, I find CAs not worth using.
It is worth noting that the CA can speed you to getting your +5 implants. Assume that you have decided to go for +5 implants for whatever reason. The CA will decrease the time it takes to get them, by about a day and a half. In this case, the CA in effect returns extra skillpoints, namely, the ones that you get because you plug in the +5 implants earlier than you otherwise would. Let's compute that. Normally, assuming you've remapped to Intelligence/Memory, and you use two +1 implants, it will take 13d, 16h, 16m to train Cybernetics V. With a CA boost, it takes 12d, 6h, 18m. So, you save roughly 1d, 10h. That's 34 hours extra time with +5 implants instead of +1, which is worth 360 skillpoints/hour. So, total is 12240 skillpoints gained, increasing the amount the CA is worth by about 5%. Still not worthwhile unless you can find a really cheap CA.
We can also compute when a CA is always superior to PLEX, that is, when it gets you at least 324 ISK/SP. This happens at a cost of 73.5m ISK. (Add 5% if you use it to train Cyber V.) So, if you ever see one for that price, grab it. Actually, what matters here is the proportion of the price of the CA to the price of PLEX. Since we assumed a price of 525m for PLEX, what we see here is that CAs are worth using if you can get one for ~0.14 PLEX.