Friday, June 01, 2007

trend spotting update

this page put forward a tentative metholodogy for trend definition in run differential back on may 11, identifying both crossover points (ie, convergence) and extrema (ie, divergence) in moving averages as potentially interesting events. on may 14, this moving average convergence-divergence (or macd) method was used to note a convergence point for the chicago cubs which could have portended upcoming difficulties for the team -- and also recognized a divergence extreme for the division-leading milwaukee brewers which could have indicated a similar fate.

what has happened since has been something of a confirmation of the method. the convergence point for the cubs came with game 34, whereupon this page said:

with the trend having turned negative -- in spite of not reaching an upside extreme -- it would seem that run differential data are indicating that the cubs best work may be behind them, at least temporarily. the shorter-term measure of pythagorean record, while still positive, is collapsing; crossover points are normally followed by a move of the 12-game measure into sub-.500 territory. this needn't indicate a disaster for the team -- as the calendar advances, past wins will be falling out of the 12-game sample. it's conceivable (though perhaps unlikely) that the club will play winning baseball even if probability holds and their differential contracts further over the coming week.


that proved too optimistic, as the club has gone 6-11 since, scoring 72 and allowing 84.



outcomes have persistently worsened through the most recent game, which marked a new season low in both 12-game and 26-game pythagorean estimated record. as was noted at the outset of the study, seasons invariably consist of trends in performance, with troughs followed by peaks and again by troughs -- and this ballclub has fallen into one of those troughs.

the picture going forward is, however, somewhat ambiguous. while the averages have again started to converge, putting a low in the macd line back at game 40, the low value was just (-1.22) -- as shown in the history of recent years, a value sometimes not negative enough to indicate that a downside extreme of momentum has been reached. it seems probable to this writer that the macd low will hold and that the next event would be a positive convergence, sending the macd into positive territory -- implying improving prospects -- but it would also seem significantly possible that an extreme low in macd (often less than -1.50) would be arrived at first.

for now, this writer chooses the more optimistic view. one can find sufficient example -- say, in 2004 -- of trend turnarounds at similarly mild macd divergences. indeed, a significant minority of trend changes are forecast from such points. time will tell, but there is cause for hope here.

in examining the milwaukee club, a less ambiguous picture is painted.



after marking a wild upside extreme of 2.46 on game 36 (may 12), this page said:

from a 12-game pythagorean reaching .866, a correction is nearly inevitable -- such runs of success are simply unsustainable. like the cubs, the games falling out of the back of the 12-game average are wins, some considerable -- again, this doesn't necessarily presage a total collapse, and the signal is perhaps not as pessimistic as the crossover which has befallen the cubs. but a moderation of their pace is nonetheless to be expected.


this too proved too mild an expectation, as the brewers have subsequently gone 5-13, scoring 58 and allowing 95! this collapse began with a 3-9 run in which the brewers were outscored 68-42 -- but at that point, game 48 on may 25, the macd line reached a low extreme of (-2.74), from which it has turned higher. 12-day estimated record there stabilized from its precipitous crash, and has now too turned higher -- it would seem here too that the worst has passed for milwaukee. the club has won two of their last three.

as was noted yesterday, the gap between the clubs has not diminished -- this morning it stands at 6.5 games. should milwaukee recover their bearings, one can add to the inability to convert run differential into victories earlier this season the inability to capitalize on a singular collapse by the division leader on the list of the failures of 2007. should the cubs further find a need to seek a truly extreme low in run differential before rebounding, one would imagine few of the remaining optimists would survive uncracked.

No comments: