This is data listed on Wikipedia, which I have not specifically confirmed, but a quick look supports it’s accuracy. These are the top close to close percentage gains in S&P 500 as well as NASDAQ. Note, S&P goes back to the 1930’s crash and that actually dominates. The second SPX list excludes those 1930’s dates. As you note, the record days aggregate in down markets.
Three bear market periods stand out for having the largest day gains close to close. The Global Financial Crisis (2008-2009) for both indexes, the Dotcom Bubble Crisis (2000-2002) mainly NASDAQ and Covid Com (2020) mainly impacting the SPX. Technically, the Dotcom bubble never regained it’s highs before plunging into the Global Financial Crisis, so most technicians call it one Bear Market, the “Lost Decade.”
Looking at these major record gain days in the NASDAQ in a couple of market bears, starting with 2007-2009 mortgage crisis that carries two of the largest gains (including the 1st and 2nd largest SPX gains according to Wikipedia). Those dates 10-13-2008 and 10-28-2008 are number 3 and 5 for NASDAQ.
The obvious finding is the large red candles that immediately follow the major gains or abbreviated rallies with significant declines undercutting the entire gain. So in regards to not making near as much holding through the bottom; well, you miss a lot of the huge losses also. I’m curious how the calculations are truly made to say you lose huge gains, other than to simply remove those profits. I’ll have to add it to my to-do spreadsheet/python list to run a more detailed check. Probably not until the rains return heavily next fall.
As a sidenote, I haven’t gone through candle by candle, but using TOS Market School charting, there were seven (7) rally days since the market dropped over 20% from October 2007. Six of them failed for long term recovery, mainly by rally low undercuts (one I haven’t confirmed specifically). That bears repeating, SIX rally attempts failed longer term. The successful one on March 10, 2009 had a follow-through day on day 5 and never looked back. Both were strong Belt-hold or true Marubozu candles. The price action is very interesting in that it visually tightened up in volatility and around the 21 sma going into the successful rally day. It would be interesting to take a look at that with actual ATRs but that’s beyond my free time. The Dotcom leg of the Lost Decade also had some record gain-days in the NASDAQ, not surprisingly. Most also followed by significant decline day(s) or limited short rallies. I’ll spare you the graphs as it takes a couple to cover that.
However, the Covid Flash Bear is interesting. The SPX was more dramatic than the NASDAQ. Again, given the general market target for the crash. It had two record gain days, the first being immediately followed by a record gap down day. However, the second was the one instance I found that you would have lost some advantage. It was the one and only Rally Day for this crash and proved very successful. Note the strong candles on the rally as well as the volatility contraction, especially the tightening around the 21 and then 50 sma’s. That is one solid reason to watch the VIX as it reflects volatility on a daily calculation while ATR is an averaged value. Another study to be done………..
I can’t mathematically support your claim just with graphs, but looking back at other previous bear episodes, the “picture” in my opinion fully supports your position. Not surprised as bear action induces and supports volatility which is reflected in price action as movement extremes.
As a side note, I tried to look up your referenced Putman article but couldn’t get past the negative reports on Putnam such as this one:
As they always say, consider your source.
Lakedog