I custom developed a trade bot that made alot of money during this move, let me explain to you how I look at profit forcasting on trades.

1. when you sell, you are charged a 0.6% fee, and the price is at the high point of your trade cycle.
2. when  you buy your btc back, you are also charged at 0.6% fee, and the price is at the low point of your trade cycle, so your fee from your sale was higher than the fee from your purchase.
3. this means you must cover at least 1.2% in profit JUST to cover fees, if the move in price is less than 1.2%, you lost money. If the move in price is less than 2.4%, you made gox more money than you made for yourself.

I wrote a trade backtester that has all mtgox trades ever made up to this second. It then spends days analyzing every buy/sell price combination testing which two prices would have made the most money in any given timeframe i want to look at. This data shows that longer term, larger moves are win. Moves around 12-18% are winners, they happen often enough and at a high enough margin to make you real profits.

In fact, with my trade strategy i have some fixed incremental trades programmed in, but the bot picks the rest of it's trades based on this backtesting technology i put together. It does a way better job at picking prices than i do.

Hope this helps.