You are choosing features for a machine learning model to predict seasoned equity offering (SEO) in the next year (year t+1). The prediction is made at the end of each year (year t). Assume that accounting information becomes available three months after the balance sheet date. Which of the following features will not introduce a look-ahead bias? Choose all correct answer(s). To facilitate the interpretation of results, you standardised all accounting ratios in your raw dataset before splitting it into train-dev-test datasets. Predicted cash balance of the firm at the end of year t+1. In making the the prediction, you assume that cash at June 30 in year t changes by the same amount as in year t-1. To test whether managers can time the market, you include abnormal stock returns over the 12-month period after the SEO. To test whether managers try to time the market, you include abnormal stock returns over the 12-month period before the SEO. The firm's growth potential proxied for by analysts' forecasts (in research reports published in year t) of its sales growth over the next three years. Predicted cash balance of the firm at the end of year t+1. In making the the prediction, you assume that cash at Dec 31 in year t changes by the same amount as in year t-1