In original D&D, as described on page 34 of Men & Magic, magic-users and clerics can conduct magical research to invent new spells.
The player proposes a new spell, and the referee assigns it a spell level. The character must conduct research for a number of weeks equal to that spell level, and spend a minimum amount of gold pieces based on the spell level.
At the end of the research period, there's a 20% chance of success. The player can spend more money to increase the odds, gaining an additional 20% for each multiple of the minimum cost spent.
What happens if that roll fails? Men & Magic doesn't say, so roll d6:
- Review previously collected research data for another week to correct a flaw in your calculations. It's unnecessary to expend further research funds, but the odds of success will be the same as before.
- This line of research hit a dead end, but you've narrowed the possibilities. After d6 weeks of additional research, your odds of success increase by 20%.
- You nearly found the answer, but your laboratory lacks the proper equipment to conclude the research. Spend an amount equal to the minimum research cost to purchase the new gear, and you will have the answer within a day. Of course, someone could liberate the equipment from the laboratory of a rival spell caster....
- The problem is more obscure than you anticipated. Spend the minimum cost in gold and weeks again. Your chance of success increases by 20%.
- Your hypothesis is flawed. Modify the spell description, and start from scratch.
- You will never succeed researching this spell without further experience. Try again after gaining a level.