Teams are afraid to pitch to Barry Bonds. And this, also fresh off the newswire: trickle-down economics didn't turn out so well for people beyond trophy wives, home-staff sous chefs and Japanese gardeners.

Bonds drew a whopping 68 intentional walks, and many of his 130 other walks were semi-intentional, plate appearances where he was given nothing to hit in the hopes he might go insane with frustration and give up the plate discipline that's made him the best hitter in baseball. Those 198 walks, by the way, are a new all-time single-season record.

	Barry Bonds Left Field San Francisco Giants Profile
	2002 SEASON STATISTICS AB HR RBI R BB AVG 403 46 110 117 198 .370

Much of this is psychological, managers not wanting to let Bonds beat them, but it's also the product of the Giants' lineup: they're driven about two-thirds by Bonds, a quarter by Kent, and a little by Benito Santiago. Outside of those three, no Giant is a significantly better hitter than their peers at their position. And with Kent batting in front of Bonds, once you get past the greatest hitter ever, it's all people who you can at least scout and pitch to. The difference between Bonds and the guys behind him is larger than the difference between, say, Jeff Kent and Neifi Perez.

Here's who's batted behind Bonds in the 5-6 slots in the playoffs so far: Benito Santiago/Reggie Sanders four times, Santiago/ J.T. Snow four times, and Santiago/Tom Goodwin once. Always directly behind him is Santiago, which is nice for purposes of writing this article and, as we'll see, a less-than-ideal choice for Bonds.

There's a good way to look at game situations, called the run expectation chart. Given an average offensive team, you can look at the situation and see how many runs they should be expected to get. 2002's run expectation table, courtesy of BP's own Michael Wolverton, looks like this:

       Bases               Outs
                     0       1      2
        empty   0.5111  0.2718  0.1006
          1st   0.8960  0.5358  0.2271
          2nd   1.1422  0.6815  0.3222
      1st 2nd   1.5106  0.9365  0.4501
          3rd   1.4051  0.9445  0.3629
      1st 3rd   1.8379  1.1852  0.5243
      2nd 3rd   1.9564  1.3580  0.6327
  1st 2nd 3rd   2.3324  1.5105  0.7756

You'll get more runs if you're facing a bad pitching staff, like the Devil Rays, and conversely you'll score fewer if you've got a terrible offense, like the Devil Rays. But on average, the chart means a leadoff walk leads to .896 runs, or a leadoff double leads to 1.142 runs, and so on.

Barry Bonds hit .370/.582/.799 in the best pitcher's park in the NL. That includes his 68 intentional walks. Take those out, and Barry's still got a whopping .513 OBP which would still be 50 points higher than the second best in the category, Brian Giles. In fact, check this out: if Bonds "expanded his strike zone" as Tony La Russa said he should, and Bonds turned those 68 intentional walks into strikeouts, he would have hit .333/.456/.540, led the league in on-base percentage, and still had a better year than Sammy Sosa.

It's hard to believe at some level that pitching to a guy who hits as well as Bonds can be a winning proposition, but it is; in almost every situation, putting another man on greatly increases the scoring chances for that inning. Pitching to Bonds is dangerous, but it still will get you an out around 50 percent of the time, and you've got to get 27 outs to end the game. To put it another way, you're five times more likely to get an out than to have him hit a home run. You can win that game. Plus, every time you walk Bonds, you're increasing the chance you're going to see him again. Which is bad.

Tom Tippett did some research into this for a classic Rob Neyer column, looking into the same question regarding the 1999 version of Mark McGwire, and found that "in no case did his extra production add up to the expected runs from walking him." Bonds is now even better than McGwire, with fewer strikeouts, even more walks, and when he's pitched to, the same incredible home-run production. That there are any situations where it makes sense to walk Barry to get to a reasonably good hitter like Benito Santiago is a demonstration of how amazingly good Bonds is, and how historic his performances.

To work out the numbers for Bonds, I made a couple of assumptions that slight Bonds: that all runners only advance one base on his singles and only two bases on his doubles. However, Kenny Lofton is swift, Rich Aurilia can run the bases, and Kent doesn't drag his feet either. I estimate that in situations with men on and Bonds up, the Giants can expect to score more, though it's perhaps as little as a tenth of a run in each situation. But as we'll see, that'll make a big difference.

With a man on second and no outs, a team will generally score a run -- the run expectation on that table is 0.926. Pitching to Bonds has a run expectation of .714 using our conservative assumptions. That's a product of the combination of possibilities: he'll single, double or walk (advancing runners), or homer (scoring two runs); those are weighed against the greater chance he'll make an out. Walking Bonds intentionally makes the situation first and second with no outs, which has a run expectation of 1.541. So putting Bonds on first base is a difference in run expectation of .827, or almost a run.

Because of Barry's prowess, there are a couple of situations where it does become a coin flip whether or not to walk him: with two outs and the bases empty, or with a man on first, second, or third. With two outs, any out off of a ball in play ends the inning.

It depends a lot on the pitcher, of course. For a home run-plagued pitcher, like the Anaheim Angels' Ramon Ortiz (who gave up 40 this year), it's much more dangerous to pitch to Bonds, and then it definitely becomes worthwhile to walk him in those situations.

Santiago complicates this. He's a big groundball hitter, about 35th in the majors in the number of groundballs to flyballs. He's also not a speedster. Bonds is a slow fellow himself who was afflicted with a strained hamstring that for a time forced him to stop tossing aside the bat after hits, so he could use it as a cane and hobble around the bases.

This means that extreme groundball pitchers like Derek Lowe or Roy Halladay have an even better incentive to walk Bonds: in the on-deck circle they might get to see a nice, fat double play or an inning-ending forceout. If you're confident in the infield defense behind you, it actually makes sense to walk Bonds with one out and no one on, taking away the solo home run while setting up the double play. It's a gutsy play, of course. If you've got a set of Gold Glovers behind you, you can push that even farther and do it in every two-out situation where the bases aren't loaded, setting up the inning-ending force at more bases. That's it, though, and even then, taking the break-even options on Bonds' at-bat means someone else gets a plate appearance to harm you later; so walking him late in a game makes more sense then walking him early.

Even facing an extreme groundball pitcher with a good defense and an inappropriately chosen groundout in waiting behind him, there are eight situations where the opposing team should start thinking about walking Bonds, considering who's on base, who's coming up, and particularly their pitcher's ability to get a groundball out from Santiago. And remember again that in the man-on situations, we're working with very conservative assumptions about advancing the runner.