I am starting to test strategies going short only using fundamental data from balance sheets as well as earnings forecasts, ratings by agencies and price itself.
Most important single variable except price will be most likely short interest. Now, I haven't found historic data on a daily basis and I am not even sure if that data is available somewhere. Does anyone know sources?
The problem with short selling is the timing and the money management. I think it is useless to try to forecast banruptcies. Trying that is tempting but I heard of no one who succeeded in that. i would rather try to define kind of an upper valuation band and a lower, where I sell the upper but close out at the lower. Thus I am more oriented in let's say 10-30% corrections than real sell offs leading to fractions of original price. The key problem is twofold: first shorting is expensive since you are restricted by the uptickrule plus the increased comission due to short fees. second the distribution of stock ptice movements works against you since, no matter what model, you cannot win more than 100% but can loose multiples of that.
Basically there are two important sources of information: sectors against each other and sectors in itself. The target is to analyse one sector against others, eg are financials stocks undervalued compared to utilities, and to find within one sector those stocks with the highest upward diversion from the others in this sector. By adding this two pieces of information up one can find candidates for short selling. It is subject to testing which techncial feature works best for triggering a trade.
peace
Most important single variable except price will be most likely short interest. Now, I haven't found historic data on a daily basis and I am not even sure if that data is available somewhere. Does anyone know sources?
The problem with short selling is the timing and the money management. I think it is useless to try to forecast banruptcies. Trying that is tempting but I heard of no one who succeeded in that. i would rather try to define kind of an upper valuation band and a lower, where I sell the upper but close out at the lower. Thus I am more oriented in let's say 10-30% corrections than real sell offs leading to fractions of original price. The key problem is twofold: first shorting is expensive since you are restricted by the uptickrule plus the increased comission due to short fees. second the distribution of stock ptice movements works against you since, no matter what model, you cannot win more than 100% but can loose multiples of that.
Basically there are two important sources of information: sectors against each other and sectors in itself. The target is to analyse one sector against others, eg are financials stocks undervalued compared to utilities, and to find within one sector those stocks with the highest upward diversion from the others in this sector. By adding this two pieces of information up one can find candidates for short selling. It is subject to testing which techncial feature works best for triggering a trade.
peace
