I could write an essay on SPI/XJO premium/discount meanings. I have been watching or trading SPI since 1983 when it started.
Probably the best exercise is simply to take the historical files and subtract one from the other and observe against price action.
Yes, ther are times when SPI premiums lead markets up and times when markets go up for a long time with SPI discounts and vice versa.
In brief, this is how I see some aspects.
The best kind of bull runs are when SPI remains at a discount as it did through 1985 into early 1987. This suggests that sentiment is bearish despite the obvious price appreciation and we know about contrarian sentiment.
In the latter stages of bull markets SPI premiums are the norm and increase with every contract change as a rule as the sentiment becomes even more bullish such as early to late '87.
So the greater the overall SPI premium as a trend then the closer a bull is to ending.
All this should work in inverse in a bear.
During this bear market there was bascically no SPI discounts till the Jan lows.
This showed a stubborn bullish denial IMO.
From then till mid March the discount has been mixed to bearish but still not what I would consider normal for a bear market.
Now we have 3 days of high premiums, and yes we are early into a new contract and some may wish to calculate "fair value" into the equation, but I am more than a little alarmed at the lack of real serious discounts on the way down and the eagerness for big premiums after just a week of upside.
It may well be, as I stated today, that the bull market has yet to complete its manic or final stage, and a bit more time will tell.
Apart from the trend aspect, outliers are also something worth watching.
Spikes in the direction of a trend are usually a capitulation indicator.
Similarly, spikes against the trend are usually a gift.
Like any indicator, there are exceptions to general rules, and this indicator like some others requires an artistic(historical perspective) interpretation rather than say momentum indicators.
I hope I gave you a few of my ideas, but really viewing history is the best way to understand.
What amazes me is that on the whole, I don't believe I see behaviour any different today than I did 25 years ago.
(20min delay)