This paper was aimed at the analysis of calendar spread trading of STIR (Short Term Interest Rate) contracts in the intraday timeframe. Calendar spread trading consists of simultaneously buying and selling STIR contracts with different expiration dates. Each of the two contracts individually behave in a rather random (hardly predictable) way. Notwithstanding, they may move together in the long run, with any short term deviations being corrected in the near future. If this long run behavior is empirically confirmed, there is room for a profitable trading strategy. When a sufficiently large deviation of price spread is identified, a trade may be opened by simultaneously buying the under-valued contract and selling the over-valued one. When the deviation is reverted, the trade would be closed out by selling the long position and off-setting the short position. To be successful, this strategy depends on the existence of the long term equilibrium of the contracts and the definition of a threshold that would trigger a change in positions. In this paper, we analyze a sample of 1304 observations, collected every 10 minutes during one month, from 5 different spread series. Long term equilibrium between the pairs of contracts is empirically tested by means of cointegration. Four pairs proved to be cointegrated. For each of these, a simulation analysis allowed the estimation of a threshold which would maximize trading profits. We were able to generate a steady and positive cash flow in simulated environment taking into consideration practical matters related to spread trading such as cost of commissions and execution risk.