Eric And Stats
I am with Eric this week end and visit his new place around the corner from his old place, which is no longer his. Eric is going through a divorce and this not an easy period. Like Todd, our together about old times .. in fact, just like old times since we fall into humour and patterns enjoyed since the summer of '87 when we met in Providence, RI, and painted houses. He actually painted and I drove around finding the houses to paint. Depending on one's perspective then, not clear who was the better off. I owning responsibility that made me ill - every day my pager brought some new, horrible, calamity. Then as now, we survived and Eric's moment will pass into a new life.
Arriving in Sommerville, we head for a new Japanese and discuss why, in an increasing random sample size, the mean and median return (IRR) converge. For instance, a portfolio of 100 investments should have a tighter spread. I note this reduces volatility and hence risk; Eric explains that it is directional - since the mean captures outliers one should know where such points occur - above or below the mean - since they pull the median up or down (down being bad). For instance, suppose a portfolio of 20 investments has an IRR of 20% but a large spread and the mean below the 20% IRR, one would wish to pull the mean up, closer to the median. This may be done by increasing the portfolio from 20 to, say, 100. A manager who enjoys 20% IRRs where the median and mean tighter is better than the manager who enjoys the same IRR with a greater spread, which suggest greater uncertainty in those returns. In venture capital, the typical fund holds 25 companies regardless of fund size .. so why don't manager own more investments? A reason may be the partnership's ability to find, select and execute deals .. 25 may reflect the best size for post-investment monitoring .. a partner's time valuable and she can only do so much.
Back in London, the kids begin 'half-term' and no school for the next week, lucky devils. Madeleine pre-occupied with her hamster 'Monty' who is a girl. She wanted a male hamster BTW but their balls unseemly. Eitan gears up for ManU vs. Liverpool Sunday, which he will watch with Joe (he: very excited). Now this a huge game. Tomorrow he has a swim gala which I will miss. Bunk.
Arriving in Sommerville, we head for a new Japanese and discuss why, in an increasing random sample size, the mean and median return (IRR) converge. For instance, a portfolio of 100 investments should have a tighter spread. I note this reduces volatility and hence risk; Eric explains that it is directional - since the mean captures outliers one should know where such points occur - above or below the mean - since they pull the median up or down (down being bad). For instance, suppose a portfolio of 20 investments has an IRR of 20% but a large spread and the mean below the 20% IRR, one would wish to pull the mean up, closer to the median. This may be done by increasing the portfolio from 20 to, say, 100. A manager who enjoys 20% IRRs where the median and mean tighter is better than the manager who enjoys the same IRR with a greater spread, which suggest greater uncertainty in those returns. In venture capital, the typical fund holds 25 companies regardless of fund size .. so why don't manager own more investments? A reason may be the partnership's ability to find, select and execute deals .. 25 may reflect the best size for post-investment monitoring .. a partner's time valuable and she can only do so much.
Back in London, the kids begin 'half-term' and no school for the next week, lucky devils. Madeleine pre-occupied with her hamster 'Monty' who is a girl. She wanted a male hamster BTW but their balls unseemly. Eitan gears up for ManU vs. Liverpool Sunday, which he will watch with Joe (he: very excited). Now this a huge game. Tomorrow he has a swim gala which I will miss. Bunk.