I shall talk about some trading experiments we have carried out on Bitcoin markets, to probe the difference between outcomes of HFT trading strategies as inferred from market data and the outcomes of actual trading. This overview will touch on (1) Lead-lag relationships between different exchanges, and the impact of fee structures; (2) The impact of latency on returns (especbially in back-testing); and  (3) The effect of liquidity on fill probabilities for 'maker' orders. All the results indicate a high degree of short-term effficiency in the BTC markets. Joint work with Jakob Albers, Mihai Cucuringu and Alex Shestopaloff. Part (1) in Applied Mathematical Finance 28 (5), 395-448 (2021); part (2) preprint available on SSRN; part (3) work in progress.