The optimizer doesn't always needs stats to give estimates. If full stats are collected and available, then the optimzer will give you realistic estimates, else if the full stats are not collected then the optimizer looks for the sample stats, if collected then optimizer uses them and the estimates will be close to actual one. If no stats are collected then the optimizer does the random amp sampling and gives out the estimates which could be close or way unrealistic. This could be ther reason your estimates are very unrealistic.
Collect the stats and then see the explain plan, which will give you a better picture.
The optimizer doesn't always needs stats to give estimates. If full stats are collected and available, then the optimzer will give you realistic estimates, else if the full stats are not collected then the optimizer looks for the sample stats, if collected then optimizer uses them and the estimates will be close to actual one. If no stats are collected then the optimizer does the random amp sampling and gives out the estimates which could be close or way unrealistic. This could be ther reason your estimates are very unrealistic.
Collect the stats and then see the explain plan, which will give you a better picture.