Venous leg ulcers are among the most common chronic wounds. Treatment is commonly with a limb compression bandage. Previous small, often single-center, studies have shown that it is possible to predict which wounds are likely to respond to compression therapy. We designed this cohort study using a dataset of over 20,000 individuals with a venous leg ulcer to investigate the accuracy of several prognostic models. Creating complex models using logistic regression, as well as simply counting prognostic factors, we show that initial measures of wound size and duration accurately predict, as measured by area under the receiver operator curve and Brier score, who will heal by the 24th week of care. For example, a wound that is less than 10 cm2 and less than 12 months old at the first visit has a 29 percent chance of not healing by the 24th week of care, while a wound greater than 10 cm2 and greater than 12 months old has a 78 percent chance of not healing. Ultimately, these models can be applied by a clinician to help determine whom to continue to treat with standard care and perhaps whom to treat with adjuvant therapies. They may also aid in the design of clinical trials.