The long side is essentially no alpha playing. The alpha is played when he turns to the short side, so what matter are the stats of the short trades. The mean of short trades, their variance, win/loss ratio, profit/loss factor, and the drawdown would help assess it.
We also leverage up or down, so our view on the S&P for the following trading day would hold as -2, -1, +1, or +2. But essentially you're right, when we are on +1 mode we can only do the same as the S&P. But one should not look at the S&P as the strategy's benchmark, our objective is to beat it because it is the hardest one to consistently beat. When the markets are down it's when we can really outperform so although we invest purely on the S&P, our "benchmark" awareness is close to zero.
I dont think his leveraged call has anything to do with past losses or winners, but on stronger signal.
Take the example of 2:1, 50/50. The average is 0.25. So a loss is actually 4 times in terms of average. In a short/long context, the average of the outperformance is a lot smaller than 0.25/day. Thus, the loss when projected to average is a lot bigger. I believe that is one main reason why funds underperform.