The method of spell research as described in the Dungeon Master Guide and the Wizard's Handbook, is elaborate, but has two distinct disadvantages:

1) It takes to much time. The spell caster must refrain from adventuring for months, not only for his research but also for preparation time and the realization of a breakthrough. The research period is longer and costlier when the rules from the WHB are used, because the chance of success drops from about 60 - 90% to 20 - 30 %.

2) The costs are to high. The spell caster must invest thousands or sometimes tens of thousands of gold pieces in his research. He is not lightly to have this kind of cash, and his comrades won't always be willing to fund his research while they just hang about in the city doing nothing.

All in all, spell research following these rules has a very bad XP to gold and time ratio.

Before I propose a modification of these rules, let me give you a short summary of them:

1) The spell caster must refrain from adventuring during the whole period of research. If the period is interrupted, he must start all over again.

2) The base duration of research is 2 weeks per spell level (-1 for specialist) plus a preparation time of spell level +1 weeks. A check ((modified)spell learn or WIS) is made to determine whether the research succeeded. If not, the spell caster must research another week and check again etc.

3) The spell caster must have access to a library. This library must be:

The assembly of this library costs 1D6 + 4 weeks per 1000 gp. One book is about 1D10 x 100 gp.

4) The spell caster must have access to a laboratory or shrine worth 1d6 x 1000 gp. It can be hired for 2D6 x 10 gp.

5) The operational cost of research is 2D6 x 100 gp per week, half of which is converted into the personal library.

6) The chance of success using DMG is the spell learn chance or a wisdom check, i.e. 60 - 90 %, using the WHB it is 10% + 10% per 1000 gp/week extra spended (base chance, max. 50%) + Intelligence rating (or wisdom for clerics) + level - spell level x2.

7) The XP bonus is 500 XP/spell level.

An example:

A third level wizard wants to create a first level spell. He wants to create his own library and uses the laboratory of some magical university.

Preparation time: 2 weeks.

Research time: 2 weeks.

Creation of library: 15 weeks.

Library: 2000 gp.

Lab. rent: 140 gp.

Operational cost: 1400 gp.

Success chance: 10% + 16 (int) + 3 (lvl) - 2 = 27 %, i.e. he will succeed 1 out of 4 times. This means, he will have to extend his researches one or two times:

1) The spell caster is able to combine research and adventuring. The research rate while adventuring however drops. His research rate depends on his resources. If he has very limited resources of his own the following modifications are used:

When using a bad laboratory (i.e. a room in a small town or village with pots and pans or a small temple for a priest), research is halved.

While traveling, research is one quarter normal.

2) Good resources and professional help increase research rate:

Apprentices and novices (max. 1 per spell level of the spell caster) save 1 day on total base research time.

The participation of spell casters of roughly equal capabilities (lvl - 1 to lvl + 5) save 1 week on total base research time.

Professional guidance (lvl. min. +5) increases research rate by a factor two.

An extensive library and laboratory (which can both be moved in a large cart) increase research with 1% per 1000 gp (max. 50%).

The research period never is shorter than spell lvl weeks.

3) The operational costs are variable, but less than the before stated. They consist of the renting of a lab or the purchasing of lab materials. I would suggest 100 gp per spell level per week base duration.

4) His chance of success is: lvl x 2 + int x 2 (or wis x 2) + apprentices + companion x 2 + teacher x 3 + 10 if specialization or sphere with major access + (lab + library worth)/1000 - spell lvl x 5

Let us now take the same example. Our spell caster now spends a week in town to buy books and minimal lab equipment at a cost of 2000 gp. After this, he follows the group on an adventure which leads him through grasslands littered with small villages and, during the second half, through rough hills. He receives help from a first level apprentice.

The base duration of his research is 14 days. His apprentice's help reduces this to 13 days.

The first part is conducted using bad laboratory circumstances, taking 2 x 6½ = 13 days.

The second part is done under traveling circumstances, taking 4 x 6½ = 26 days.

His small outfit would reduce this by 2%, which is now not significant.

In total, he has now taken 26 days of research and 7 days of preparation, totaling in 33 days, or about a month.

The costs are: 2000 gp investment + 200 gp operational costs = 2.200 gp.

His chance of success is: 6 + 32 + 1 + 0 + 0 + 0 + 2 - 5 = 36 %, or roughly 1/3, which means he probably must research another week or has to look up an essential part in a library in a big city, is one component short or something the like - DM's option. This leaves plenty of role-playing opportunity.

Using this set of rules, our spell caster is able to participate in adventuring (i.e. make money and XP), has to spend less money (7810 - 2200 = 5610 gp less) and it takes him less time (one in stead of five months). This all combines into making the research of new spells more interesting, not only for the wizard himself, but more so for his fellow adventurers who don't have to finance him and don't have to wait for him.

At the DM's option he could require the spell caster to build up a reasonable traveling lab/shrine and a small library before he can research spells of a certain level. One could use the library costs which are stated in the WHB rules as a guideline.