Relative efficiencies for the discovery of Earth-crossing asteroids (ECAs) are modelled for various telescopes at Siding Spring. It is found that the narrow-field instruments—the Anglo-Australian Telescope and the 40 in and 2·3m reflectors—are not competitive in this regard for present CCD imaging systems. The UK Schmidt Telescope (UKST), if used to take short-exposure stereo pairs of photographs, would be an effective search tool, outperforming all current systems apart from the Ground-based Electro-Optical Deep Space Surveillance (GEODSS) systems now being implemented by the US Air Force for ECA searches. If a CCD mosaic were fitted to the UKST, its performance would far exceed that of any other device at Siding Spring, and it would produce ECA discoveries at a rate around 3–4 times as high as GEODSS, but at considerable expense. The most sophisticated search instrument currently in use is the University of Arizona’s Spacewatch telescope; a notable result found here is that even with its present CCD, the Automated Patrol Telescope (APT) of the University of New South Wales would be able to match or outperform Spacewatch for all ECA sizes, including ~10m objects, should this modelling be a reasonable representation of its real performance. In terms of cost-effectiveness and telescope availability, the conclusion arrived at herein is that the APT, equipped with small-pixel but large-format CCD chips of high quantum efficiency, would be an extremely effective ECA search instrument: if operated with 12 μm pixel chips covering a 4° × 4° field it might produce ECA discoveries at a rate well in excess of the combined rate for all current search programs.